- Email: [email protected]
Simple electrocardiographic criteria for rapid identification of wide QRS complex tachycardia: The new limb lead algorithm
Journal Pre-proof Simple Electrocardiogaphic Criteria for Rapid Identification of Wide QRS Complex Tachycardia: the new Limb Lead Algorithm Qiong Chen, MD, Jinyi Xu, MD, Carola Gianni, MD PHD, Chintan Trivedi, MD MPH, Domenico G. Della Rocca, MD, Mohammed Bassiouny, MD, Ugur Canpolat, MD, Alfredo Chauca Tapia, MD, J. David Burkhardt, MD, Javier E. Sanchez, MD, Patrick Hranitzky, MD, G. Joseph Gallinghouse, MD, Amin Al-Ahmad, MD, Rodney Horton, MD, Luigi Di Biase, MD PHD, Sanghamitra Mohanty, MD, Andrea Natale, MD FESC FACC FHRS PII:
S1547-5271(19)30854-9
DOI:
https://doi.org/10.1016/j.hrthm.2019.09.021
Reference:
HRTHM 8153
To appear in:
Heart Rhythm
Received Date: 23 April 2019
Please cite this article as: Chen Q, Xu J, Gianni C, Trivedi C, Della Rocca DG, Bassiouny M, Canpolat U, Tapia AC, Burkhardt JD, Sanchez JE, Hranitzky P, Gallinghouse GJ, Al-Ahmad A, Horton R, Di Biase L, Mohanty S, Natale A, Simple Electrocardiogaphic Criteria for Rapid Identification of Wide QRS Complex Tachycardia: the new Limb Lead Algorithm Heart Rhythm (2019), doi: https://doi.org/10.1016/ j.hrthm.2019.09.021. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc. on behalf of Heart Rhythm Society.
Simple Electrocardiogaphic Criteria for Rapid Identification of Wide QRS Complex Tachycardia: the new Limb Lead Algorithm Brief running title: LLA in differential diagnosis of WTC
Qiong Chen, MD1,2, Jinyi Xu, MD1, Carola Gianni, MD PHD2, Chintan Trivedi, MD MPH2, Domenico G. Della Rocca, MD2, Mohammed Bassiouny, MD2, Ugur Canpolat, MD2,3, Alfredo Chauca Tapia, MD2, J. David Burkhardt, MD2, Javier E. Sanchez, MD2, Patrick Hranitzky, MD2, G. Joseph Gallinghouse, MD2, Amin Al-Ahmad, MD2, Rodney Horton, MD2, Luigi Di Biase, MD PHD2,4, Sanghamitra Mohanty, MD2,5，Andrea Natale2,5-8, MD FESC FACC FHRS Affiliations: 1: Department of Cardiopulmonary Function Test, Henan Provincial People’s Hospital, People’s Hospital of Zhengzhou University 2: Texas Cardiac Arrhythmia Institute, St. David’s Medical Center, Austin, TX, USA 3: Hacettepe University Faculty of Medicine, Department of Cardiology, Ankara, Turkey 4: Albert Einstein College of Medicine at Montefiore Hospital, New York, USA 5: Dell Medical School, Austin, TX 6: Interventional Electrophysiology, Scripps Clinic, San Diego, CA, USA 7: Metro Health Medical Center, Case Western Reserve University School of Medicine, Cleveland, OH, USA 8: Division of Cardiology, Stanford University, Stanford, CA, USA Address for Correspondence: Andrea Natale, MD, FACC, FESC, FHRS Executive Medical Director, Texas Cardiac Arrhythmia Institute, St. David's Medical Center, Austin, Texas. Consulting Professor, Division of Cardiology, Stanford University, Palo Alto, California. Clinical Associate Professor of Medicine, Case Western Reserve University, Cleveland, Ohio Senior Medical Director, Interventional Electrophysiology, Scripps Clinic, San Diego, CA, USA Address: 3000 N. I-35, Suite 720; Austin, TX 78705 Email: [email protected] Conflict statement: Dr. Andrea Natale received honoraria from Boston Scientific, Biosense Webster, Janssen, St. Jude Medical, Biotronik and Medtronic
Abstract Background: The electrocardiogram (ECG) is essential for the differential diagnosis of wide QRS complex tachycardia (WCT). Objective: To evaluate the diagnostic value of a novel ECG algorithm based on the morphologic characteristics of the QRS on the limb leads.
Methods: Limb Leads Algorithm (LLA) was evaluated by analyzing 528 monomorphic WCTs with electrophysiology (EP)-confirmed diagnoses. In the LLA, ventricular tachycardia (VT) is diagnosed in the presence of at least one of the following: ⑴ monophasic R wave in lead aVR; ⑵ predominantly negative QRS in leads I, II, III; ⑶ Opposing QRS complex in the Limb leads (OQL): concordant monophasic QRS complex in all three inferior leads; AND concordant monophasic QRS complex in two or three of the remaining limb leads with an opposite polarity to that of the inferior leads. The diagnostic performance of the LLA was compared with that of the Brugada, Vereckei and RWPT algorithm.
Results: Of 528 WCT cases, 397 were VT and 131 supraventricular tachycardia (SVT). The interobserver agreement for LLA was excellent (k = 0.98), better than the other algorithms. The overall accuracy of LLA (88.1%) was similar to Brugada (85.4%) and Vereckei (88.1%), moreover superior than RWPT (70.8%). LLA had a lower sensitivity (87.2%) when compared with Brugada (94.0%) and Vereckei (92.4%), but not RWPT (67.8%). Furthermore, LLA showed a higher specificity (90.8%), when compared than Brugada (59.5%), Vereckei (76.3%) and RWPT (80.2%).
Conclusion: The LLA is a simple, yet accurate method to diagnose VT when approaching WCTs on ECG.
Key words: Wide QRS complex tachycardia, electrocardiography, ventricular tachycardia, algorithm, Opposing QRS complex in Limb leads
Introduction The 12-lead surface electrocardiogram (ECG) is a useful tool to identify VT in the differential diagnosis of wide QRS complex tachycardia (WCT) 1. This is mainly based on the analysis of the cardiac electrical activity in the precordial leads on the horizontal plane (HP) and/or limb leads on the frontal plane (FP) 2. In clinical practice, the most widely used algorithms to diagnose VT are based on the analysis of the electrical vector in the HP, such as those proposed by Wellens et al, Kindwall et al, and Brugada et al
3-5
. Other algorithms, such as the
aVR lead four steps algorithm 6 and the R-wave peak time (RWPK) algorithm 7, are based on FP vectors. A common problem with the existing algorithms is the inability of independent authors 8-10 to reproduce the high sensitivity, specificity and accuracy reported in the original papers. The most likely explanation for this discrepancy is the complexity of these algorithms, secondary to the number of steps involved and/or the difficulty and variability in calculation of necessary measurements. This also presents an important limitation for their use in clinical practice by non-electrophysiology (EP) medical personnel. The aim of the current study is to find a simple, accurate and rapid algorithm for VT diagnosis based on ECG patterns on the limb leads.
Methods Study Population A total of 528 monomorphic WCT with an established electrophysiology (EP) diagnosis were included in this study. WCT was defined as a regular rhythm with a monomorphic wide complex QRS (duration ≥ 0.12 s) and a heart rate of ≥100 bpm. Patients were excluded from the study when the ECG quality was poor.
EP studies were performed from January 2015 to December 2018 at St David’s Medical Center and Henan Provincial People’s Hospital. Patient information was collected in the prospective IRB approved VT database at St. David’s Medical Center, USA and de-identified data were captured from Henan Provincial People’s Hospital in China.
Study design
The 12-lead ECG of the WCT was performed at the time of the EP procedure in every patient. Each ECG was analyzed independently by two qualified electrophysiologists who were blinded to the patient’s general information and their EP results. Each observer was asked to analyze the ECGs based on the four different algorithms: •
4-steps Brugada algorithm 5, where diagnosis of VT is made if any of the following criteria is met: 1. absence of a RS complex in all precordial lead 2. RS interval in one precordial lead > 100 msec 3. atrioventricular dissociation 4. the morphology criteria for VT present in both leads V1 and V6, according to the RBBB or LBBB morphology 5
•
aVR Vereckei algorithm 6, where diagnosis of VT is made if any of the following criteria in aVR is met: 1. presence of an initial R wave 2. width of an initial r or q wave >40 msec 3. notching of the down-stroke of the QS wave
4. ventricular activation–velocity ratio (vi/vt) > 1 • RWPT algorithm 7, where diagnosis of VT is made if the interval between the onset of the QRS to the first visible change in polarity (i.e. peak of the R wave/nadir of the S wave, or any notch on the ascending/descending limb of the R/S wave) ≥ 50 msec • limb leads algorithm (LLA), where diagnosis of VT is made if any of the following criteria in the limb leads is met: 1. monophasic R wave in lead aVR (Figure 1-A) 2. predominantly negative QRS in limb leads I, II, and III
the QRS complex may appear as QS, Qr, rS, Qrs, qrS, rSr’, or even four or more waves, but the main overall voltage must be negative in all these leads (Figure 1-A, B)
3. opposing QRS complex in the Limb leads (OQL): (Figure 1-C)
monophasic QRS complexes (QS/R) in all three inferior leads sharing the same polarity (ALL positive or ALL negative, but including a notched R or QS complex), AND
monophasic QRS complexes (R/QS) in two or three of the remaining limb leads, with an opposite polarity to that of the inferior leads
Presence or absence of VT was detected at each step of four methods (Figure 2). In case of inter-operator disagreement, the final diagnosis was made after reassessment by a third independent electrophysiologist. The EP study result was considered the gold standard for the final diagnosis. The positive predictive value (PPV), negative predictive value (NPV), specificity, sensitivity, and accuracy were compared with those of the Brugada, aVR Vereckei and RWPT algorithms. In addition, given the low specificity of the Brugada algorithm, we separately analyzed its performance for RBBB and LBBB pattern (step 4).
Development of the Limb Lead Algorithm
At first, we developed the “opposing QRS Complex in limb leads (OQL)” algorithm, by retrospectively analyzing 130 WCTs from patients with a proven EP diagnosis at the Henan Provincial People’s Hospital.
11
The OQL algorithm is based on the different depolarization directions of VT and SVT with aberrant ventricular conduction on FP. If the WCT is SVT with aberrancy, the initial wavefront activates one side of the ventricle fast via the His-Purkinje system (normal conduction), and then spreads to the other side with a slower conduction. Thus, the depolarizing vector in the ventricles will be mainly horizontal, pointing from the fast to the slow side. However, in VT this is totally different: the activation wavefront will spread muscle by muscle via gap junctions, thus, the main depolarization vector will have a direction that depends on the origin of the VT. Most commonly, VTs originate either in the superior or inferior aspect of the ventricle. In the former, the main depolarization vector will have a superior to inferior direction, therefore the inferior leads will show monophasic R wave, and the superior leads will show QS wave. In the latter, the main depolarization vector will have an inferior to superior direction, therefore the inferior leads will show monophasic QS wave, whereas the superior leads will show R wave. (Figure 3A)
While the specificity of the OQL was high, its sensitivity was relatively low. This is explained by, 1) VTs originating near the conduction system, where the inferior leads may be discordant; 2) VTs originating in the subendocardial aspect of the inferior left ventricle, an initial “r” wave may present on the inferior leads (Figure 3B).
12
To encompass these situations, we revised the algorithm adding two additional criteria: presence of
monophasic R wave in aVR13 (criterion 1), and predominantly negative voltage in leads I, II, III14 (criterion 2). Together, these criteria can cover most VTs originating the left ventricle, including those don’t show a typical OQL pattern (subendocardial or apical).
Statistical Analysis
Continuous data were presented as mean ± standard deviation and categorical data were presented as counts (percentage). Interobserver agreement for EKG analysis between two reviewers was estimated using kappa (k)-values. The strength of concordance according to k-values was interpreted as follows: k < 0.21, poor; k = 0.21–0.40, fair; k
= 0.41–0.60, moderate; k = 0.61–0.80, good; and k > 0.80, excellent. The accuracy was calculated as the proportion of correctly diagnosed WCTs (VT and SVT) divided by all WCTs. The sensitivity of VT was calculated as the proportion of correctly diagnosed VTs divided by all VTs. The specificity of VT was calculated as the proportion of correctly diagnosed SVTs divided by all SVTs. The PPV was calculated as the proportion correctly diagnosed VTs divided by all WCTs diagnosed as VT by each ECG algorithm. The NPV was calculated as the proportion of correctly diagnosed SVTs divided by all WCTs diagnosed as SVT by each ECG algorithm. Comparison analysis of sensitivity and specificity was performed with the McNemar's test. Comparison analysis of PPV and NPV value was performed using the method described by Moskowitz and Pepe 15. To evaluate the diagnostic performance of each method, a receiver operating characteristic (ROC) curve was created and area under the curve (AUC) of ROC was computed. AUC of ROCs were compared by DeLong's test. A two-tailed p-value < 0.05 was considered statistically significant. Statistical tests were performed by statistical software R 3.5.0 and IBM SPSS Statistics 23.0 (IBM SPSS Inc, Chicago, IL, USA).
Results Mean age was 54.1 ± 17.3 years and 70.4% patients were male, % with baseline wide QRS. According to the EP study, 397/528 (75.2%) WTCs were VT and 131 (24.8%) had SVT. There was excellent interobserver agreement for all three methods (k > 0.80), of which the LLA (k = 0.98) was better than Brugada (k = 0.89), Vereckei (k = 0.90) and RWPT (k = 0.83). Overall accuracy, sensitivity, specificity, PPV and NPV of the four algorithms are shown in Table 1, whereas results of the comparison analysis are shown in Table 2. The LLA correctly identified 465/528 cases (overall test accuracy 88.1% (95% CI 85.0-90.7)), which was similar to Brugada (450/528, 85.4% (82.1 - 88.3) and Vereckei (467/528, 88.5 (85.4 - 91.15)). All of the aforementioned outperformed the RWPT algorithm (374/528, 70.8 (66.8 - 74.7)). The sensitivity of the LLA [87.2% (95% CI 83.5 - 90.3)] was lower than to that of Brugada [93.95% (91.1 96.1)] and Vereckei [92.4% (89.4 – 94.8)], but higher than the sensitivity of the RWPT algorithm [67.8% (62.9 – 72.3)]. Importantly, the LLA demonstrated the highest specificity [90.8% (84.6 – 95.2)] among the four methods, excelling that of Brugada [59.5% (50.6 - 68.0)], Vereckei [76.3% (68.1 - 83.3)] and RWPT [80.2% (72.3-86.6)]. The PPV of the LLA [96.7% (94.4-98.0)] was better when compared to Brugada [87.6 % (85.1-89.7)], Vereckei [92.2% (89.7-94.2)] and RWPT [91.2% (87.9–93.6)]. Of note, the PPV was lower for Brugada, when compared to the Vereckei algorithm. The NPV of LLA [70.0 % (64.2-75.2)] showed no difference when compared with Brugada [76.5 % (68.3-83.1)] and Vereckei [76.9 % (70.0-82.7)] but was significantly higher than that of the RWPT algorithm [45.1% (41.0-49.2)]. In the ROC curve analysis (Figure 4, supplemental Table 1), AUC for the LLA [0.89 (0.86-0.92)] was better than the AUC calculated for the Brugada algorithm [0.77 (0.71-0.82)] and RWPT [0.74% (0.69 – 0.79)], but comparable to that of the Vereckei algorithm [0.84 (0.80-0.89)]. In addition, the Brugada algorithm was analyzed separately for RBBB and LBBB patterns. RBBB was present in 314 patients and LBBB in 214 patients. Accuracy was comparable for RBBB and LBBB [268/314 (85.4%)
vs 183/214 (85.5 %), p-value: 0.96). The sensitivity was higher for RBBB when compared to LBBB [97.5% (94.6-99.1) vs 88.8% (82.8-93.2), p-value < 0.001]; however, RBBB had worse specificity [48.1% (36.5-59.7) vs. 75.9% (62.4-86.5), p-value: 0.001]. In the ROC curve analysis, AUC was lower for RBBB, when compared to LBBB type [0.73 (0.65-0.80) vs 0.82 (0.75-0.90), p-value: 0.027].
Discussion Our study results showed that, when compared to the Brugada, Vereckei, and RWPT algorithms, the LLA has the highest specificity and greatest AUC in diagnosis of VT, with high reproducibility among different observers. An important advantage of LLA is its high specificity. The high specificity mainly depends on the OQL pattern, which is the core component of the LLA. Indeed, in our previous study, which incorporated only OQL in the differential diagnosis of VT, its specificity reached as high as 92.1%11. This novel concept is based on the different depolarization directions of VT and SVT with aberrant ventricular conduction on FP. The depolarization direction of VT mainly determined by its origin. If the VT wavefront spreads from the superior ventricle, the QRS complex will appear as a tall, monophasic R wave in the inferior leads 16-18. Conversely, if the initial VT wavefront spreads from the inferior ventricle to the superior ventricle, the QRS complex will be positive in the superior leads and negative in the inferior leads.
19
. The reason behind the high specificity of
OQL pattern is how the ventricles are activated during SVT with aberrancy. During SVT, the vector points from the fast to the slow side, and it is mainly horizontal. In RBBB aberrancy, the left ventricle activates first and the initial vector points left and inferior, and to complete the depolarization of the right ventricle, the vector will then turn to the right-superior. In this scenario, the QRS complex in the left lateral leads (I and aVL), will usually mirror that of the inferior leads, i.e. a +/- biphasic wave, resulting in less than 5 leads with a monophasic QRS waves. In LBBB aberrancy, the initial depolarization vector will point from right anterior superior to the left posterior inferior, and, as the left ventricle is dominant in adults, the vector representing the activation of the right ventricle might be canceled. Thus, an R or r wave will be present in the left lateral leads (I and aVL), whereas the QRS complex will have a r or R wave, but its morphology can be anything from R to Rs, rS, rs, rSr, rsr depending on the axis deviation. In this scenario, it is impossible to have more than 2 leads show a QS pattern. Therefore, OQL pattern can help us excludes SVT in WCT differential diagnosis. Of note, the OQL pattern does not occur in all the inferior or superior originating VTs, which limits the sensitivity of OQL. First, lead DI, aVL, aVR, and inferior leads are not completely symmetric, therefore even if all the inferior leads show a concordant monophasic pattern, it is possible for the remaining limb leads to show
variable, non-monophasic patterns. Second, when VTs originate from the endocardium of the inferior wall, the depolarization vector will create an initial “r” wave in the inferior leads 12, determining a rS pattern instead of QS. For these reasons, in order to improve the sensitivity of our algorithm, the presence of a monophasic R wave in aVR (Criterion 1)
13
and predominantly negative voltage value in I, II, III (Criterion 2)
14
was
incorporated. Those two criteria both indicate an extreme axis deviation, which suggests VT originating from the lower left ventricle, and have been already used in other algorithms 20, 21. We’ve also shown a lower than expected specificity when testing the Brugada algorithm (59.2%). This is consistent with what other independent authors have reported,
2, 6, 8, 10, 22-24
none of which were able to
corroborate the high specificity (96.5%) of the Brugada’s original paper 5. When testing the performance of the Brugada algorithm step-by-step, we’ve found a high specificity in the first three steps (98%, 93.9%, 93.9%, respectively), and the major drawback was found in the last step. To understand this discrepancy, we’ve analyzed separately LBBB-like and RBBB-like pattern WCTs: misdiagnosis of SVT with aberrant conduction as VT are more likely to happen with RBBB-like pattern WCTs (specificity, 46.6% vs 77.5% for LBBB). One possible explanation to this is that RBBB is more common than LBBB in aberrant SVTs. Another, more important explanation might be that, in cases of RBBB and left anterior fascicular block (LAFB) aberrancy, the ECG features R waves on lead V1 and R/S < 1 on lead V6, which is easily classified as VT when applying the RBBB morphology Brugada criteria (Figure 5). The other important advantage of LLA lies in its simplicity and higher reproducibility among different observers, with a kappa index of 1.0. Compared to Brugada algorithm, LLA is defined clearly with no ambiguity and has less steps. There are two main difficulties associated with the Brugada criteria, especially when used by non-EP doctors. First, it can be challenging to identify the AV dissociation on the standard 12lead ECG; second, the complexity of the last step (criterion 4), which is hard to remember and leads to confusion. Even in our study, where the ECGs were analyzed by EPs, Brugada has shown the lowest interobserver agreement, when compared with LLA and Vereckei. Compared with Vereckei algorithm, LLA requires no measurement. While Vereckei is relatively easier to remember than the Brugada algorithm, measuring vi and vt accurately is hard, since identification of the exact start of the QRS requires experience and
can be arbitrary. This is especially problematic when the aVR lead has a very low amplitude (≤ ±0.01 mV) or the initial q wave duration at aVR lead is close to 40 msec (Figure 6). Unfortunately, this last “measuring” step is required for the differential diagnosis in many cases, up to 40% in this study. In addition, the aVR was almost isoelectric in 15 patients, with 32 patients having an initial q wave duration close to 40 msec. These can lead to discrepancy among different observers, explaning the inferior kappa index of the Veckerei algorithm compared with that of the LLA. With only one step, the RWPT criterion is the most straightforward algorithm in diagnosing WCT, however, its limitation is apparent: the identification of an exact start and visible change point of the QRS in lead II can be tricky in some patients and thus subsides its feasibility and reproducibility. Also, it ranked the lowest in overall accuracy and sensitivity among all four algorithms.
Clinical applications The LLA can be used for a quick differential diagnosis of WCT by EP and non EP providers. Of note, the LLA is suitable for the Holter and telemetry recordings, which usually incorporate all limb leads, with limited information on the precordial leads. Incorporating the LLA in the automatic arrhythmia detection criteria of these devices can improve their accuracy for diagnosis of WCT.
Limitation: The main limitation of the LLA is its inability to identify VTs originating from the conduction system (para hisbundle, fascicular, or the moderator band), as well as from intracavitary structures (such as papillary muscle VTs 25): in these cases, the inferior leads will mostly be discordant, thus limiting the value of LLA (Figure 5B). Also, all the abnormal morphology of LLA can occur in patients with abnormal baseline QRS such as the congenital heart disease, we may need more cases and studies in these special population in future. Finally, we did not assess the effect of limb lead electrodes positioning on the LAA, as part of the ECGs were performed by placing the limb electrodes on the shoulders and hips, which might affect the ECG morphology
26
. However,
this means that the LLA retains a high specificity, irrespective of the limb leads electrodes placement, and can be used even in non-standard settings (like telemetry or intra-procedural recordings).
Conclusion: The LLA is a novel simple ECG algorithm to quickly diagnose VT with good specificity and inter-observer agreement, thus complementing the diagnostic algorithms previously described in the literature.
Reference:
1.
2.
3. 4. 5. 6. 7. 8. 9.
10. 11.
12. 13. 14. 15. 16.
17.
18. 19.
Al-Khatib SM, Stevenson WG, Ackerman MJ, et al. 2017 AHA/ACC/HRS guideline for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death: Executive summary: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. Heart Rhythm Oct 2018;15:e190-e252. Kaiser E, Darrieux FC, Barbosa SA, Grinberg R, Assis-Carmo A, Sousa JC, Hachul D, Pisani CF, Kosa E, Pastore CA, Scanavacca MI. Differential diagnosis of wide QRS tachycardias: comparison of two electrocardiographic algorithms. Europace Sep 2015;17:1422-1427. Wellens HJ, Bar FW, Lie KI. The value of the electrocardiogram in the differential diagnosis of a tachycardia with a widened QRS complex. Am J Med Jan 1978;64:27-33. Kindwall KE, Brown J, Josephson ME. Electrocardiographic criteria for ventricular tachycardia in wide complex left bundle branch block morphology tachycardias. Am J Cardiol Jun 1 1988;61:1279-1283. Brugada P, Brugada J, Mont L, Smeets J, Andries EW. A new approach to the differential diagnosis of a regular tachycardia with a wide QRS complex. Circulation May 1991;83:1649-1659. Vereckei A, Duray G, Szenasi G, Altemose GT, Miller JM. New algorithm using only lead aVR for differential diagnosis of wide QRS complex tachycardia. Heart Rhythm Jan 2008;5:89-98. Pava LF, Perafan P, Badiel M, Arango JJ, Mont L, Morillo CA, Brugada J. R-wave peak time at DII: a new criterion for differentiating between wide complex QRS tachycardias. Heart Rhythm Jul 2010;7:922-926. Jastrzebski M, Kukla P, Czarnecka D, Kawecka-Jaszcz K. Comparison of five electrocardiographic methods for differentiation of wide QRS-complex tachycardias. Europace Aug 2012;14:1165-1171. Szelenyi Z, Duray G, Katona G, Frituz G, Szego E, Kovacs E, Szenasi G, Vereckei A. Comparison of the "real-life" diagnostic value of two recently published electrocardiogram methods for the differential diagnosis of wide QRS complex tachycardias. Acad Emerg Med Nov 2013;20:1121-1130. Isenhour JL, Craig S, Gibbs M, Littmann L, Rose G, Risch R. Wide-complex tachycardia: continued evaluation of diagnostic criteria. Acad Emerg Med Jul 2000;7:769-773. Chen Q, Mohanty S, Rocca DD, Trivedi C, Gianni C, Tapia A, Gedikli O, Natale A, Xu J. A New, Simple ECG Algorithm for the Rapid Identification of Ventricular Tachycardia Based on Opposing QRS Complexes on Limb Leads. Journal of the American College of Cardiology 2019;73:319. Yamada T. Transthoracic epicardial catheter ablation: indications, techniques, and complications. Circ J 2013;77:1672-1680. Vereckei A, Duray G, Szenasi G, Altemose GT, Miller JM. Application of a new algorithm in the differential diagnosis of wide QRS complex tachycardia. Eur Heart J Mar 2007;28:589-600. Kashou AH, Kashou HE. Electrical Axis (Normal, Right Axis Deviation, and Left Axis Deviation). StatPearls. Treasure Island (FL)2018. Moskowitz CS, Pepe MS. Comparing the predictive values of diagnostic tests: sample size and analysis for paired study designs. Clin Trials 2006;3:272-279. Kamakura S, Shimizu W, Matsuo K, Taguchi A, Suyama K, Kurita T, Aihara N, Ohe T, Shimomura K. Localization of optimal ablation site of idiopathic ventricular tachycardia from right and left ventricular outflow tract by body surface ECG. Circulation Oct 13 1998;98:1525-1533. Tada H, Tadokoro K, Ito S, et al. Idiopathic ventricular arrhythmias originating from the tricuspid annulus: Prevalence, electrocardiographic characteristics, and results of radiofrequency catheter ablation. Heart Rhythm Jan 2007;4:7-16. Tada H, Ito S, Naito S, et al. Idiopathic ventricular arrhythmia arising from the mitral annulus: a distinct subgroup of idiopathic ventricular arrhythmias. J Am Coll Cardiol Mar 15 2005;45:877-886. Della Rocca DG, Gianni C, Mohanty S, Trivedi C, Di Biase L, Natale A. Localization of Ventricular Arrhythmias for Catheter Ablation: The Role of Surface Electrocardiogram. Card Electrophysiol Clin Jun 2018;10:333-354.
20.
21. 22.
23. 24.
25. 26.
Surawicz B, Childers R, Deal BJ, et al. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: part III: intraventricular conduction disturbances: a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society. Endorsed by the International Society for Computerized Electrocardiology. J Am Coll Cardiol Mar 17 2009;53:976-981. Vereckei A. Current algorithms for the diagnosis of wide QRS complex tachycardias. Curr Cardiol Rev Aug 2014;10:262-276. Ceresnak SR, Liberman L, Avasarala K, Tanel R, Motonaga KS, Dubin AM. Are wide complex tachycardia algorithms applicable in children and patients with congenital heart disease? J Electrocardiol Nov-Dec 2010;43:694-700. Lau EW, Ng GA. Comparison of the performance of three diagnostic algorithms for regular broad complex tachycardia in practical application. Pacing Clin Electrophysiol May 2002;25:822-827. Jastrzebski M, Kukla P, Czarnecka D, Kawecka-Jaszcz K. Specificity of the wide QRS complex tachycardia algorithms in recipients of cardiac resynchronization therapy. J Electrocardiol May-Jun 2012;45:319326. Enriquez A, Pathak RK, Santangeli P, et al. Inferior lead discordance in ventricular arrhythmias: A specific marker for certain arrhythmia locations. J Cardiovasc Electrophysiol Oct 2017;28:1179-1186. Kligfield P, Gettes LS, Bailey JJ, et al. Recommendations for the standardization and interpretation of the electrocardiogram: part I: the electrocardiogram and its technology a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society endorsed by the International Society for Computerized Electrocardiology. J Am Coll Cardiol Mar 13 2007;49:11091127.
Table 1. Overall accuracy sensitivity, specificity, PPV, and NPV of the four algorithms LLA
Brugada
Vereckei
RWPT
Accuracy, % (95% CI)
88.1 (85.0 - 90.7)
85.4 (82.1 - 88.3)
88.5 (85.4 - 91.15)
70.8 (66.8 - 74.7)
Sensitivity, % (95% CI)
87.2 (83.5 - 90.3)
93.95 (91.1 - 96.1)
92.4 (89.4 - 94.8)
67.8 (62.9 - 72.3)
Specificity, % (95% CI)
90.8 (84.6 - 95.2)
59.5 (50.6 - 68.0)
76.3 (68.1 - 83.3)
80.2 (72.3 - 86.6)
PPV, % (95% CI)
96.7 (94.4 - 98.0)
87.6 (85.1 - 89.7)
92.2 (89.7 - 94.2)
91.2 (87.9 - 93.6)
NPV, % (95% CI)
70.0 (64.2 - 75.2)
76.5 (68.3 - 83.1)
76.9 (70.0 - 82..7)
45.1 (41.0 - 49.2)
LLA, limb lead algorithm; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; RWPT, R wave peak time
Table 2. Results of comparison analysis (p-values) of accuracy, sensitivity, specificity, PPV and NPV among the four algorithms.
Accuracy
Sensitivity
Specificity
PPV
NPV
LLA vs. Brugada
0.20
<0.001***
<0.001***
<0.001***
0.13
LLA vs. Vereckei
0.84
0.01*
0.001**
0.002**
0.09
Brugada vs. Vereckei
0.14
0.44
0.003**
0.003**
0.92
RWPT vs. Brugada
<0.001***
<0.001***
<0.001***
0.06
<0.001***
RWPT vs. Vereckei
<0.001***
<0.001***
0.47
0.51
<0.001***
RWPT vs. LLA
<0.001***
<0.001***
0.02*
0.002**
<0.001***
LLA, limb lead algorithm; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; RWPT, R wave peak time
Figure Legends: Figure 1: Examples of VT correctly identified by the LLA. A: monophasic R wave in aVR (highlighted yellow); B: predominantly negative QRS in leads I, II, III all (highlighted purple). C: OQL pattern: monophasic QRS complexes in all three inferior leads sharing the same polarity (C1-4, ALL positive, C5 ALL negative), and monophasic QRS complexes in two or three of the remaining limb leads, with an opposite polarity to that in the inferior leads (all cases). LLA, limb leads algorithm; OQL, opposing QRS complex in Limb leads, VT, ventricular tachycardia Figure 2: Results of the analysis of each ECG by the LLA, Brugada, and Vereckei algorithms. LLA, limb lead algorithm; RWPT, R wave peak time. Figure 3. Rationale for the OQL and LLA. A, resulting depolarization vectors when VTs originate either in the superior (A-1) or inferior (A-2) aspect of the ventricle. B, resulting depolarization vectors when VTs originate in the subendocardial aspect of the inferior left ventricle (B-1) when compared to the epicardium (B-2).
Figure 4. ROC curve for the four algorithms. LLA, limb lead algorithm; RWPT, R wave peak time algorithm. Figure 5: ECG example comparing LLA and the Brugada algorithm A). 12-leads ECG of a 32 y.o. male patient with orthodromic AVRT. While LLA suggest SVT (all three criteria negative), this is VT according to the Brugada algorithm (criterion 4 for RBBB). B). 12-leads ECG of a 17 y.o. male patient with VT originated from left posterior fascicle. According to OQL, it is SVT; according to Brugada algorithm (criterion 4 for RBBB), it is VT. LLA, limb leads algorithm; VT, ventricular tachycardia; RBBB, right bundle branch block; SVT, supraventricular tachycardia; EP, electrophysiology; AVRT, atrioventricular reentrant tachycardia. Figure 6: ECG examples comparing LLA and the Vereckei algorithm. A) orthodromic AVRT: according to LLA (all criteria are negative), this is SVT; according to Vereckei, since the initial r wave in aVR is easily missed, hence vi/vt >1, this is VT
B) VT: according to LLA (criterion 2), this is VT; applying Vereckei is difficult, the initial q wave is around 40 msec, and the upstroke of the R wave is very steep making it difficult to measure vi on a standard ECG. LLA, limb leads algorithm; VT, ventricular tachycardia; SVT, supraventricular tachycardia; EP, electrophysiology; AVRT, atrioventricular reentrant tachycardia.
S1547-5271(19)30854-9
DOI:
https://doi.org/10.1016/j.hrthm.2019.09.021
Reference:
HRTHM 8153
To appear in:
Heart Rhythm
Received Date: 23 April 2019
Please cite this article as: Chen Q, Xu J, Gianni C, Trivedi C, Della Rocca DG, Bassiouny M, Canpolat U, Tapia AC, Burkhardt JD, Sanchez JE, Hranitzky P, Gallinghouse GJ, Al-Ahmad A, Horton R, Di Biase L, Mohanty S, Natale A, Simple Electrocardiogaphic Criteria for Rapid Identification of Wide QRS Complex Tachycardia: the new Limb Lead Algorithm Heart Rhythm (2019), doi: https://doi.org/10.1016/ j.hrthm.2019.09.021. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc. on behalf of Heart Rhythm Society.
Simple Electrocardiogaphic Criteria for Rapid Identification of Wide QRS Complex Tachycardia: the new Limb Lead Algorithm Brief running title: LLA in differential diagnosis of WTC
Qiong Chen, MD1,2, Jinyi Xu, MD1, Carola Gianni, MD PHD2, Chintan Trivedi, MD MPH2, Domenico G. Della Rocca, MD2, Mohammed Bassiouny, MD2, Ugur Canpolat, MD2,3, Alfredo Chauca Tapia, MD2, J. David Burkhardt, MD2, Javier E. Sanchez, MD2, Patrick Hranitzky, MD2, G. Joseph Gallinghouse, MD2, Amin Al-Ahmad, MD2, Rodney Horton, MD2, Luigi Di Biase, MD PHD2,4, Sanghamitra Mohanty, MD2,5，Andrea Natale2,5-8, MD FESC FACC FHRS Affiliations: 1: Department of Cardiopulmonary Function Test, Henan Provincial People’s Hospital, People’s Hospital of Zhengzhou University 2: Texas Cardiac Arrhythmia Institute, St. David’s Medical Center, Austin, TX, USA 3: Hacettepe University Faculty of Medicine, Department of Cardiology, Ankara, Turkey 4: Albert Einstein College of Medicine at Montefiore Hospital, New York, USA 5: Dell Medical School, Austin, TX 6: Interventional Electrophysiology, Scripps Clinic, San Diego, CA, USA 7: Metro Health Medical Center, Case Western Reserve University School of Medicine, Cleveland, OH, USA 8: Division of Cardiology, Stanford University, Stanford, CA, USA Address for Correspondence: Andrea Natale, MD, FACC, FESC, FHRS Executive Medical Director, Texas Cardiac Arrhythmia Institute, St. David's Medical Center, Austin, Texas. Consulting Professor, Division of Cardiology, Stanford University, Palo Alto, California. Clinical Associate Professor of Medicine, Case Western Reserve University, Cleveland, Ohio Senior Medical Director, Interventional Electrophysiology, Scripps Clinic, San Diego, CA, USA Address: 3000 N. I-35, Suite 720; Austin, TX 78705 Email: [email protected] Conflict statement: Dr. Andrea Natale received honoraria from Boston Scientific, Biosense Webster, Janssen, St. Jude Medical, Biotronik and Medtronic
Abstract Background: The electrocardiogram (ECG) is essential for the differential diagnosis of wide QRS complex tachycardia (WCT). Objective: To evaluate the diagnostic value of a novel ECG algorithm based on the morphologic characteristics of the QRS on the limb leads.
Methods: Limb Leads Algorithm (LLA) was evaluated by analyzing 528 monomorphic WCTs with electrophysiology (EP)-confirmed diagnoses. In the LLA, ventricular tachycardia (VT) is diagnosed in the presence of at least one of the following: ⑴ monophasic R wave in lead aVR; ⑵ predominantly negative QRS in leads I, II, III; ⑶ Opposing QRS complex in the Limb leads (OQL): concordant monophasic QRS complex in all three inferior leads; AND concordant monophasic QRS complex in two or three of the remaining limb leads with an opposite polarity to that of the inferior leads. The diagnostic performance of the LLA was compared with that of the Brugada, Vereckei and RWPT algorithm.
Results: Of 528 WCT cases, 397 were VT and 131 supraventricular tachycardia (SVT). The interobserver agreement for LLA was excellent (k = 0.98), better than the other algorithms. The overall accuracy of LLA (88.1%) was similar to Brugada (85.4%) and Vereckei (88.1%), moreover superior than RWPT (70.8%). LLA had a lower sensitivity (87.2%) when compared with Brugada (94.0%) and Vereckei (92.4%), but not RWPT (67.8%). Furthermore, LLA showed a higher specificity (90.8%), when compared than Brugada (59.5%), Vereckei (76.3%) and RWPT (80.2%).
Conclusion: The LLA is a simple, yet accurate method to diagnose VT when approaching WCTs on ECG.
Key words: Wide QRS complex tachycardia, electrocardiography, ventricular tachycardia, algorithm, Opposing QRS complex in Limb leads
Introduction The 12-lead surface electrocardiogram (ECG) is a useful tool to identify VT in the differential diagnosis of wide QRS complex tachycardia (WCT) 1. This is mainly based on the analysis of the cardiac electrical activity in the precordial leads on the horizontal plane (HP) and/or limb leads on the frontal plane (FP) 2. In clinical practice, the most widely used algorithms to diagnose VT are based on the analysis of the electrical vector in the HP, such as those proposed by Wellens et al, Kindwall et al, and Brugada et al
3-5
. Other algorithms, such as the
aVR lead four steps algorithm 6 and the R-wave peak time (RWPK) algorithm 7, are based on FP vectors. A common problem with the existing algorithms is the inability of independent authors 8-10 to reproduce the high sensitivity, specificity and accuracy reported in the original papers. The most likely explanation for this discrepancy is the complexity of these algorithms, secondary to the number of steps involved and/or the difficulty and variability in calculation of necessary measurements. This also presents an important limitation for their use in clinical practice by non-electrophysiology (EP) medical personnel. The aim of the current study is to find a simple, accurate and rapid algorithm for VT diagnosis based on ECG patterns on the limb leads.
Methods Study Population A total of 528 monomorphic WCT with an established electrophysiology (EP) diagnosis were included in this study. WCT was defined as a regular rhythm with a monomorphic wide complex QRS (duration ≥ 0.12 s) and a heart rate of ≥100 bpm. Patients were excluded from the study when the ECG quality was poor.
EP studies were performed from January 2015 to December 2018 at St David’s Medical Center and Henan Provincial People’s Hospital. Patient information was collected in the prospective IRB approved VT database at St. David’s Medical Center, USA and de-identified data were captured from Henan Provincial People’s Hospital in China.
Study design
The 12-lead ECG of the WCT was performed at the time of the EP procedure in every patient. Each ECG was analyzed independently by two qualified electrophysiologists who were blinded to the patient’s general information and their EP results. Each observer was asked to analyze the ECGs based on the four different algorithms: •
4-steps Brugada algorithm 5, where diagnosis of VT is made if any of the following criteria is met: 1. absence of a RS complex in all precordial lead 2. RS interval in one precordial lead > 100 msec 3. atrioventricular dissociation 4. the morphology criteria for VT present in both leads V1 and V6, according to the RBBB or LBBB morphology 5
•
aVR Vereckei algorithm 6, where diagnosis of VT is made if any of the following criteria in aVR is met: 1. presence of an initial R wave 2. width of an initial r or q wave >40 msec 3. notching of the down-stroke of the QS wave
4. ventricular activation–velocity ratio (vi/vt) > 1 • RWPT algorithm 7, where diagnosis of VT is made if the interval between the onset of the QRS to the first visible change in polarity (i.e. peak of the R wave/nadir of the S wave, or any notch on the ascending/descending limb of the R/S wave) ≥ 50 msec • limb leads algorithm (LLA), where diagnosis of VT is made if any of the following criteria in the limb leads is met: 1. monophasic R wave in lead aVR (Figure 1-A) 2. predominantly negative QRS in limb leads I, II, and III
the QRS complex may appear as QS, Qr, rS, Qrs, qrS, rSr’, or even four or more waves, but the main overall voltage must be negative in all these leads (Figure 1-A, B)
3. opposing QRS complex in the Limb leads (OQL): (Figure 1-C)
monophasic QRS complexes (QS/R) in all three inferior leads sharing the same polarity (ALL positive or ALL negative, but including a notched R or QS complex), AND
monophasic QRS complexes (R/QS) in two or three of the remaining limb leads, with an opposite polarity to that of the inferior leads
Presence or absence of VT was detected at each step of four methods (Figure 2). In case of inter-operator disagreement, the final diagnosis was made after reassessment by a third independent electrophysiologist. The EP study result was considered the gold standard for the final diagnosis. The positive predictive value (PPV), negative predictive value (NPV), specificity, sensitivity, and accuracy were compared with those of the Brugada, aVR Vereckei and RWPT algorithms. In addition, given the low specificity of the Brugada algorithm, we separately analyzed its performance for RBBB and LBBB pattern (step 4).
Development of the Limb Lead Algorithm
At first, we developed the “opposing QRS Complex in limb leads (OQL)” algorithm, by retrospectively analyzing 130 WCTs from patients with a proven EP diagnosis at the Henan Provincial People’s Hospital.
11
The OQL algorithm is based on the different depolarization directions of VT and SVT with aberrant ventricular conduction on FP. If the WCT is SVT with aberrancy, the initial wavefront activates one side of the ventricle fast via the His-Purkinje system (normal conduction), and then spreads to the other side with a slower conduction. Thus, the depolarizing vector in the ventricles will be mainly horizontal, pointing from the fast to the slow side. However, in VT this is totally different: the activation wavefront will spread muscle by muscle via gap junctions, thus, the main depolarization vector will have a direction that depends on the origin of the VT. Most commonly, VTs originate either in the superior or inferior aspect of the ventricle. In the former, the main depolarization vector will have a superior to inferior direction, therefore the inferior leads will show monophasic R wave, and the superior leads will show QS wave. In the latter, the main depolarization vector will have an inferior to superior direction, therefore the inferior leads will show monophasic QS wave, whereas the superior leads will show R wave. (Figure 3A)
While the specificity of the OQL was high, its sensitivity was relatively low. This is explained by, 1) VTs originating near the conduction system, where the inferior leads may be discordant; 2) VTs originating in the subendocardial aspect of the inferior left ventricle, an initial “r” wave may present on the inferior leads (Figure 3B).
12
To encompass these situations, we revised the algorithm adding two additional criteria: presence of
monophasic R wave in aVR13 (criterion 1), and predominantly negative voltage in leads I, II, III14 (criterion 2). Together, these criteria can cover most VTs originating the left ventricle, including those don’t show a typical OQL pattern (subendocardial or apical).
Statistical Analysis
Continuous data were presented as mean ± standard deviation and categorical data were presented as counts (percentage). Interobserver agreement for EKG analysis between two reviewers was estimated using kappa (k)-values. The strength of concordance according to k-values was interpreted as follows: k < 0.21, poor; k = 0.21–0.40, fair; k
= 0.41–0.60, moderate; k = 0.61–0.80, good; and k > 0.80, excellent. The accuracy was calculated as the proportion of correctly diagnosed WCTs (VT and SVT) divided by all WCTs. The sensitivity of VT was calculated as the proportion of correctly diagnosed VTs divided by all VTs. The specificity of VT was calculated as the proportion of correctly diagnosed SVTs divided by all SVTs. The PPV was calculated as the proportion correctly diagnosed VTs divided by all WCTs diagnosed as VT by each ECG algorithm. The NPV was calculated as the proportion of correctly diagnosed SVTs divided by all WCTs diagnosed as SVT by each ECG algorithm. Comparison analysis of sensitivity and specificity was performed with the McNemar's test. Comparison analysis of PPV and NPV value was performed using the method described by Moskowitz and Pepe 15. To evaluate the diagnostic performance of each method, a receiver operating characteristic (ROC) curve was created and area under the curve (AUC) of ROC was computed. AUC of ROCs were compared by DeLong's test. A two-tailed p-value < 0.05 was considered statistically significant. Statistical tests were performed by statistical software R 3.5.0 and IBM SPSS Statistics 23.0 (IBM SPSS Inc, Chicago, IL, USA).
Results Mean age was 54.1 ± 17.3 years and 70.4% patients were male, % with baseline wide QRS. According to the EP study, 397/528 (75.2%) WTCs were VT and 131 (24.8%) had SVT. There was excellent interobserver agreement for all three methods (k > 0.80), of which the LLA (k = 0.98) was better than Brugada (k = 0.89), Vereckei (k = 0.90) and RWPT (k = 0.83). Overall accuracy, sensitivity, specificity, PPV and NPV of the four algorithms are shown in Table 1, whereas results of the comparison analysis are shown in Table 2. The LLA correctly identified 465/528 cases (overall test accuracy 88.1% (95% CI 85.0-90.7)), which was similar to Brugada (450/528, 85.4% (82.1 - 88.3) and Vereckei (467/528, 88.5 (85.4 - 91.15)). All of the aforementioned outperformed the RWPT algorithm (374/528, 70.8 (66.8 - 74.7)). The sensitivity of the LLA [87.2% (95% CI 83.5 - 90.3)] was lower than to that of Brugada [93.95% (91.1 96.1)] and Vereckei [92.4% (89.4 – 94.8)], but higher than the sensitivity of the RWPT algorithm [67.8% (62.9 – 72.3)]. Importantly, the LLA demonstrated the highest specificity [90.8% (84.6 – 95.2)] among the four methods, excelling that of Brugada [59.5% (50.6 - 68.0)], Vereckei [76.3% (68.1 - 83.3)] and RWPT [80.2% (72.3-86.6)]. The PPV of the LLA [96.7% (94.4-98.0)] was better when compared to Brugada [87.6 % (85.1-89.7)], Vereckei [92.2% (89.7-94.2)] and RWPT [91.2% (87.9–93.6)]. Of note, the PPV was lower for Brugada, when compared to the Vereckei algorithm. The NPV of LLA [70.0 % (64.2-75.2)] showed no difference when compared with Brugada [76.5 % (68.3-83.1)] and Vereckei [76.9 % (70.0-82.7)] but was significantly higher than that of the RWPT algorithm [45.1% (41.0-49.2)]. In the ROC curve analysis (Figure 4, supplemental Table 1), AUC for the LLA [0.89 (0.86-0.92)] was better than the AUC calculated for the Brugada algorithm [0.77 (0.71-0.82)] and RWPT [0.74% (0.69 – 0.79)], but comparable to that of the Vereckei algorithm [0.84 (0.80-0.89)]. In addition, the Brugada algorithm was analyzed separately for RBBB and LBBB patterns. RBBB was present in 314 patients and LBBB in 214 patients. Accuracy was comparable for RBBB and LBBB [268/314 (85.4%)
vs 183/214 (85.5 %), p-value: 0.96). The sensitivity was higher for RBBB when compared to LBBB [97.5% (94.6-99.1) vs 88.8% (82.8-93.2), p-value < 0.001]; however, RBBB had worse specificity [48.1% (36.5-59.7) vs. 75.9% (62.4-86.5), p-value: 0.001]. In the ROC curve analysis, AUC was lower for RBBB, when compared to LBBB type [0.73 (0.65-0.80) vs 0.82 (0.75-0.90), p-value: 0.027].
Discussion Our study results showed that, when compared to the Brugada, Vereckei, and RWPT algorithms, the LLA has the highest specificity and greatest AUC in diagnosis of VT, with high reproducibility among different observers. An important advantage of LLA is its high specificity. The high specificity mainly depends on the OQL pattern, which is the core component of the LLA. Indeed, in our previous study, which incorporated only OQL in the differential diagnosis of VT, its specificity reached as high as 92.1%11. This novel concept is based on the different depolarization directions of VT and SVT with aberrant ventricular conduction on FP. The depolarization direction of VT mainly determined by its origin. If the VT wavefront spreads from the superior ventricle, the QRS complex will appear as a tall, monophasic R wave in the inferior leads 16-18. Conversely, if the initial VT wavefront spreads from the inferior ventricle to the superior ventricle, the QRS complex will be positive in the superior leads and negative in the inferior leads.
19
. The reason behind the high specificity of
OQL pattern is how the ventricles are activated during SVT with aberrancy. During SVT, the vector points from the fast to the slow side, and it is mainly horizontal. In RBBB aberrancy, the left ventricle activates first and the initial vector points left and inferior, and to complete the depolarization of the right ventricle, the vector will then turn to the right-superior. In this scenario, the QRS complex in the left lateral leads (I and aVL), will usually mirror that of the inferior leads, i.e. a +/- biphasic wave, resulting in less than 5 leads with a monophasic QRS waves. In LBBB aberrancy, the initial depolarization vector will point from right anterior superior to the left posterior inferior, and, as the left ventricle is dominant in adults, the vector representing the activation of the right ventricle might be canceled. Thus, an R or r wave will be present in the left lateral leads (I and aVL), whereas the QRS complex will have a r or R wave, but its morphology can be anything from R to Rs, rS, rs, rSr, rsr depending on the axis deviation. In this scenario, it is impossible to have more than 2 leads show a QS pattern. Therefore, OQL pattern can help us excludes SVT in WCT differential diagnosis. Of note, the OQL pattern does not occur in all the inferior or superior originating VTs, which limits the sensitivity of OQL. First, lead DI, aVL, aVR, and inferior leads are not completely symmetric, therefore even if all the inferior leads show a concordant monophasic pattern, it is possible for the remaining limb leads to show
variable, non-monophasic patterns. Second, when VTs originate from the endocardium of the inferior wall, the depolarization vector will create an initial “r” wave in the inferior leads 12, determining a rS pattern instead of QS. For these reasons, in order to improve the sensitivity of our algorithm, the presence of a monophasic R wave in aVR (Criterion 1)
13
and predominantly negative voltage value in I, II, III (Criterion 2)
14
was
incorporated. Those two criteria both indicate an extreme axis deviation, which suggests VT originating from the lower left ventricle, and have been already used in other algorithms 20, 21. We’ve also shown a lower than expected specificity when testing the Brugada algorithm (59.2%). This is consistent with what other independent authors have reported,
2, 6, 8, 10, 22-24
none of which were able to
corroborate the high specificity (96.5%) of the Brugada’s original paper 5. When testing the performance of the Brugada algorithm step-by-step, we’ve found a high specificity in the first three steps (98%, 93.9%, 93.9%, respectively), and the major drawback was found in the last step. To understand this discrepancy, we’ve analyzed separately LBBB-like and RBBB-like pattern WCTs: misdiagnosis of SVT with aberrant conduction as VT are more likely to happen with RBBB-like pattern WCTs (specificity, 46.6% vs 77.5% for LBBB). One possible explanation to this is that RBBB is more common than LBBB in aberrant SVTs. Another, more important explanation might be that, in cases of RBBB and left anterior fascicular block (LAFB) aberrancy, the ECG features R waves on lead V1 and R/S < 1 on lead V6, which is easily classified as VT when applying the RBBB morphology Brugada criteria (Figure 5). The other important advantage of LLA lies in its simplicity and higher reproducibility among different observers, with a kappa index of 1.0. Compared to Brugada algorithm, LLA is defined clearly with no ambiguity and has less steps. There are two main difficulties associated with the Brugada criteria, especially when used by non-EP doctors. First, it can be challenging to identify the AV dissociation on the standard 12lead ECG; second, the complexity of the last step (criterion 4), which is hard to remember and leads to confusion. Even in our study, where the ECGs were analyzed by EPs, Brugada has shown the lowest interobserver agreement, when compared with LLA and Vereckei. Compared with Vereckei algorithm, LLA requires no measurement. While Vereckei is relatively easier to remember than the Brugada algorithm, measuring vi and vt accurately is hard, since identification of the exact start of the QRS requires experience and
can be arbitrary. This is especially problematic when the aVR lead has a very low amplitude (≤ ±0.01 mV) or the initial q wave duration at aVR lead is close to 40 msec (Figure 6). Unfortunately, this last “measuring” step is required for the differential diagnosis in many cases, up to 40% in this study. In addition, the aVR was almost isoelectric in 15 patients, with 32 patients having an initial q wave duration close to 40 msec. These can lead to discrepancy among different observers, explaning the inferior kappa index of the Veckerei algorithm compared with that of the LLA. With only one step, the RWPT criterion is the most straightforward algorithm in diagnosing WCT, however, its limitation is apparent: the identification of an exact start and visible change point of the QRS in lead II can be tricky in some patients and thus subsides its feasibility and reproducibility. Also, it ranked the lowest in overall accuracy and sensitivity among all four algorithms.
Clinical applications The LLA can be used for a quick differential diagnosis of WCT by EP and non EP providers. Of note, the LLA is suitable for the Holter and telemetry recordings, which usually incorporate all limb leads, with limited information on the precordial leads. Incorporating the LLA in the automatic arrhythmia detection criteria of these devices can improve their accuracy for diagnosis of WCT.
Limitation: The main limitation of the LLA is its inability to identify VTs originating from the conduction system (para hisbundle, fascicular, or the moderator band), as well as from intracavitary structures (such as papillary muscle VTs 25): in these cases, the inferior leads will mostly be discordant, thus limiting the value of LLA (Figure 5B). Also, all the abnormal morphology of LLA can occur in patients with abnormal baseline QRS such as the congenital heart disease, we may need more cases and studies in these special population in future. Finally, we did not assess the effect of limb lead electrodes positioning on the LAA, as part of the ECGs were performed by placing the limb electrodes on the shoulders and hips, which might affect the ECG morphology
26
. However,
this means that the LLA retains a high specificity, irrespective of the limb leads electrodes placement, and can be used even in non-standard settings (like telemetry or intra-procedural recordings).
Conclusion: The LLA is a novel simple ECG algorithm to quickly diagnose VT with good specificity and inter-observer agreement, thus complementing the diagnostic algorithms previously described in the literature.
Reference:
1.
2.
3. 4. 5. 6. 7. 8. 9.
10. 11.
12. 13. 14. 15. 16.
17.
18. 19.
Al-Khatib SM, Stevenson WG, Ackerman MJ, et al. 2017 AHA/ACC/HRS guideline for management of patients with ventricular arrhythmias and the prevention of sudden cardiac death: Executive summary: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society. Heart Rhythm Oct 2018;15:e190-e252. Kaiser E, Darrieux FC, Barbosa SA, Grinberg R, Assis-Carmo A, Sousa JC, Hachul D, Pisani CF, Kosa E, Pastore CA, Scanavacca MI. Differential diagnosis of wide QRS tachycardias: comparison of two electrocardiographic algorithms. Europace Sep 2015;17:1422-1427. Wellens HJ, Bar FW, Lie KI. The value of the electrocardiogram in the differential diagnosis of a tachycardia with a widened QRS complex. Am J Med Jan 1978;64:27-33. Kindwall KE, Brown J, Josephson ME. Electrocardiographic criteria for ventricular tachycardia in wide complex left bundle branch block morphology tachycardias. Am J Cardiol Jun 1 1988;61:1279-1283. Brugada P, Brugada J, Mont L, Smeets J, Andries EW. A new approach to the differential diagnosis of a regular tachycardia with a wide QRS complex. Circulation May 1991;83:1649-1659. Vereckei A, Duray G, Szenasi G, Altemose GT, Miller JM. New algorithm using only lead aVR for differential diagnosis of wide QRS complex tachycardia. Heart Rhythm Jan 2008;5:89-98. Pava LF, Perafan P, Badiel M, Arango JJ, Mont L, Morillo CA, Brugada J. R-wave peak time at DII: a new criterion for differentiating between wide complex QRS tachycardias. Heart Rhythm Jul 2010;7:922-926. Jastrzebski M, Kukla P, Czarnecka D, Kawecka-Jaszcz K. Comparison of five electrocardiographic methods for differentiation of wide QRS-complex tachycardias. Europace Aug 2012;14:1165-1171. Szelenyi Z, Duray G, Katona G, Frituz G, Szego E, Kovacs E, Szenasi G, Vereckei A. Comparison of the "real-life" diagnostic value of two recently published electrocardiogram methods for the differential diagnosis of wide QRS complex tachycardias. Acad Emerg Med Nov 2013;20:1121-1130. Isenhour JL, Craig S, Gibbs M, Littmann L, Rose G, Risch R. Wide-complex tachycardia: continued evaluation of diagnostic criteria. Acad Emerg Med Jul 2000;7:769-773. Chen Q, Mohanty S, Rocca DD, Trivedi C, Gianni C, Tapia A, Gedikli O, Natale A, Xu J. A New, Simple ECG Algorithm for the Rapid Identification of Ventricular Tachycardia Based on Opposing QRS Complexes on Limb Leads. Journal of the American College of Cardiology 2019;73:319. Yamada T. Transthoracic epicardial catheter ablation: indications, techniques, and complications. Circ J 2013;77:1672-1680. Vereckei A, Duray G, Szenasi G, Altemose GT, Miller JM. Application of a new algorithm in the differential diagnosis of wide QRS complex tachycardia. Eur Heart J Mar 2007;28:589-600. Kashou AH, Kashou HE. Electrical Axis (Normal, Right Axis Deviation, and Left Axis Deviation). StatPearls. Treasure Island (FL)2018. Moskowitz CS, Pepe MS. Comparing the predictive values of diagnostic tests: sample size and analysis for paired study designs. Clin Trials 2006;3:272-279. Kamakura S, Shimizu W, Matsuo K, Taguchi A, Suyama K, Kurita T, Aihara N, Ohe T, Shimomura K. Localization of optimal ablation site of idiopathic ventricular tachycardia from right and left ventricular outflow tract by body surface ECG. Circulation Oct 13 1998;98:1525-1533. Tada H, Tadokoro K, Ito S, et al. Idiopathic ventricular arrhythmias originating from the tricuspid annulus: Prevalence, electrocardiographic characteristics, and results of radiofrequency catheter ablation. Heart Rhythm Jan 2007;4:7-16. Tada H, Ito S, Naito S, et al. Idiopathic ventricular arrhythmia arising from the mitral annulus: a distinct subgroup of idiopathic ventricular arrhythmias. J Am Coll Cardiol Mar 15 2005;45:877-886. Della Rocca DG, Gianni C, Mohanty S, Trivedi C, Di Biase L, Natale A. Localization of Ventricular Arrhythmias for Catheter Ablation: The Role of Surface Electrocardiogram. Card Electrophysiol Clin Jun 2018;10:333-354.
20.
21. 22.
23. 24.
25. 26.
Surawicz B, Childers R, Deal BJ, et al. AHA/ACCF/HRS recommendations for the standardization and interpretation of the electrocardiogram: part III: intraventricular conduction disturbances: a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society. Endorsed by the International Society for Computerized Electrocardiology. J Am Coll Cardiol Mar 17 2009;53:976-981. Vereckei A. Current algorithms for the diagnosis of wide QRS complex tachycardias. Curr Cardiol Rev Aug 2014;10:262-276. Ceresnak SR, Liberman L, Avasarala K, Tanel R, Motonaga KS, Dubin AM. Are wide complex tachycardia algorithms applicable in children and patients with congenital heart disease? J Electrocardiol Nov-Dec 2010;43:694-700. Lau EW, Ng GA. Comparison of the performance of three diagnostic algorithms for regular broad complex tachycardia in practical application. Pacing Clin Electrophysiol May 2002;25:822-827. Jastrzebski M, Kukla P, Czarnecka D, Kawecka-Jaszcz K. Specificity of the wide QRS complex tachycardia algorithms in recipients of cardiac resynchronization therapy. J Electrocardiol May-Jun 2012;45:319326. Enriquez A, Pathak RK, Santangeli P, et al. Inferior lead discordance in ventricular arrhythmias: A specific marker for certain arrhythmia locations. J Cardiovasc Electrophysiol Oct 2017;28:1179-1186. Kligfield P, Gettes LS, Bailey JJ, et al. Recommendations for the standardization and interpretation of the electrocardiogram: part I: the electrocardiogram and its technology a scientific statement from the American Heart Association Electrocardiography and Arrhythmias Committee, Council on Clinical Cardiology; the American College of Cardiology Foundation; and the Heart Rhythm Society endorsed by the International Society for Computerized Electrocardiology. J Am Coll Cardiol Mar 13 2007;49:11091127.
Table 1. Overall accuracy sensitivity, specificity, PPV, and NPV of the four algorithms LLA
Brugada
Vereckei
RWPT
Accuracy, % (95% CI)
88.1 (85.0 - 90.7)
85.4 (82.1 - 88.3)
88.5 (85.4 - 91.15)
70.8 (66.8 - 74.7)
Sensitivity, % (95% CI)
87.2 (83.5 - 90.3)
93.95 (91.1 - 96.1)
92.4 (89.4 - 94.8)
67.8 (62.9 - 72.3)
Specificity, % (95% CI)
90.8 (84.6 - 95.2)
59.5 (50.6 - 68.0)
76.3 (68.1 - 83.3)
80.2 (72.3 - 86.6)
PPV, % (95% CI)
96.7 (94.4 - 98.0)
87.6 (85.1 - 89.7)
92.2 (89.7 - 94.2)
91.2 (87.9 - 93.6)
NPV, % (95% CI)
70.0 (64.2 - 75.2)
76.5 (68.3 - 83.1)
76.9 (70.0 - 82..7)
45.1 (41.0 - 49.2)
LLA, limb lead algorithm; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; RWPT, R wave peak time
Table 2. Results of comparison analysis (p-values) of accuracy, sensitivity, specificity, PPV and NPV among the four algorithms.
Accuracy
Sensitivity
Specificity
PPV
NPV
LLA vs. Brugada
0.20
<0.001***
<0.001***
<0.001***
0.13
LLA vs. Vereckei
0.84
0.01*
0.001**
0.002**
0.09
Brugada vs. Vereckei
0.14
0.44
0.003**
0.003**
0.92
RWPT vs. Brugada
<0.001***
<0.001***
<0.001***
0.06
<0.001***
RWPT vs. Vereckei
<0.001***
<0.001***
0.47
0.51
<0.001***
RWPT vs. LLA
<0.001***
<0.001***
0.02*
0.002**
<0.001***
LLA, limb lead algorithm; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; RWPT, R wave peak time
Figure Legends: Figure 1: Examples of VT correctly identified by the LLA. A: monophasic R wave in aVR (highlighted yellow); B: predominantly negative QRS in leads I, II, III all (highlighted purple). C: OQL pattern: monophasic QRS complexes in all three inferior leads sharing the same polarity (C1-4, ALL positive, C5 ALL negative), and monophasic QRS complexes in two or three of the remaining limb leads, with an opposite polarity to that in the inferior leads (all cases). LLA, limb leads algorithm; OQL, opposing QRS complex in Limb leads, VT, ventricular tachycardia Figure 2: Results of the analysis of each ECG by the LLA, Brugada, and Vereckei algorithms. LLA, limb lead algorithm; RWPT, R wave peak time. Figure 3. Rationale for the OQL and LLA. A, resulting depolarization vectors when VTs originate either in the superior (A-1) or inferior (A-2) aspect of the ventricle. B, resulting depolarization vectors when VTs originate in the subendocardial aspect of the inferior left ventricle (B-1) when compared to the epicardium (B-2).
Figure 4. ROC curve for the four algorithms. LLA, limb lead algorithm; RWPT, R wave peak time algorithm. Figure 5: ECG example comparing LLA and the Brugada algorithm A). 12-leads ECG of a 32 y.o. male patient with orthodromic AVRT. While LLA suggest SVT (all three criteria negative), this is VT according to the Brugada algorithm (criterion 4 for RBBB). B). 12-leads ECG of a 17 y.o. male patient with VT originated from left posterior fascicle. According to OQL, it is SVT; according to Brugada algorithm (criterion 4 for RBBB), it is VT. LLA, limb leads algorithm; VT, ventricular tachycardia; RBBB, right bundle branch block; SVT, supraventricular tachycardia; EP, electrophysiology; AVRT, atrioventricular reentrant tachycardia. Figure 6: ECG examples comparing LLA and the Vereckei algorithm. A) orthodromic AVRT: according to LLA (all criteria are negative), this is SVT; according to Vereckei, since the initial r wave in aVR is easily missed, hence vi/vt >1, this is VT
B) VT: according to LLA (criterion 2), this is VT; applying Vereckei is difficult, the initial q wave is around 40 msec, and the upstroke of the R wave is very steep making it difficult to measure vi on a standard ECG. LLA, limb leads algorithm; VT, ventricular tachycardia; SVT, supraventricular tachycardia; EP, electrophysiology; AVRT, atrioventricular reentrant tachycardia.