Statistics from Altmetric.com
- ventricular tachycardia
- supraventricular arrhythmias
- cardiac arrhythmias and resuscitation science
Review the underlying principles of physiological cardiac activation.
Assess the differences in cardiac activation between ventricular and supraventricular tachycardias, and how these may relate to electrocardiographic changes.
Discuss the electrocardiographic algorithms for diagnosis of wide complex tachycardia.
A wide QRS complex tachycardia (WCT) is defined as a tachycardia with QRS duration of >120 ms. Distinguishing between the different potential causes of a WCT can have important implications, particularly in terms of determining the urgency of treatment, deciding between different antiarrhythmic drugs and risk stratification for sudden cardiac death. Potential differential diagnoses for a WCT include ventricular tachycardia (VT), supraventricular tachycardia (SVT) with aberrant conduction, SVT with antegrade conduction via an accessory pathway (AP), ventricular pacing and ECG artefact (box 1). In most cases, paced ventricular activation and ECG artefact can be excluded with relative ease based on the ECG features and background clinical history. On the other hand, distinguishing VT from SVT may represent a challenge. In this regard, sufficient knowledge of the background history and detailed analysis of the ECG are central to making an accurate diagnosis.1 2 This article will outline the approaches and algorithms for distinguishing between different forms of WCT with a particular focus on ECG characteristics to distinguishing VT from SVT.
Differential diagnosis of wide complex tachycardia
Causes of wide complex tachycardia
Supraventricular tachycardia with aberrancy
Supraventricular tachycardia with antegrade accessory pathway conduction
Cardiac activation during VT and SVT
In general, there are two major differences in the activation patterns that can be exploited when distinguishing VT from SVT: (1) the relationship between atrial and ventricular activation, and (2) the sequence of ventricular activation.
Relationship between atrium and ventricle
During SVT, the tachycardia originates from the atria or involves the atria in the tachycardia circuit. During VT, cardiac activation originates from the ventricle and atrial activation may or may not be linked to ventricular activation. In the event of VT with no VA conduction, the ventricle and atrium are dissociated. The presence of a less than 1:1 ratio of P wave to QRS complex (more ‘vs’ than ‘As’) is strongly suggestive of VT, with a specificity approaching 100%.3 4 In contrast, an equal P–QRS relationship (1:1 ratio) does not reliably exclude VT as this pattern may occur due to retrograde VA conduction during VT. Of note, identification of AV dissociation on 12-lead ECG may be the only non-invasive means to distinguish VT and SVT due to antegrade AP conduction.
In addition to the relationship between P waves and QRS complexes, ‘capture’ and ‘fusion’ beats may also be used as markers of AV dissociation. Capture beats are characterised by a sudden narrowing of the QRS complex (resembling the patient’s baseline rhythm) during WCT caused by ventricular activation entirely from the atria and using the specialised conduction system (figure 1). Fusion beats are characterised by a ‘hybrid’ QRS complex caused by collision of simultaneous ventricular activation from both ventricular and atrial sources. Importantly, differentiation between capture and fusion beats rely on a comparison with baseline ECG.
Patterns of ventricular activation
SVT with aberrant conduction activates the ventricle via the AV node with subsequent recruitment of all or part of the specialised His-Purkinje system. The rapid initial His-Purkinje–medicated activation is typically followed by a late phase of slower myocyte-to-myocyte activation, resulting in a typical RBBB or LBBB morphology (figure 2). VT on the other hand may originate from anywhere within the ventricle and is commonly associated with slow initial myocyte-to-myocyte conduction that may eventually engage the specialised His-Purkinje system. The differences in ventricular activation between SVT with aberrant conduction and VT result in ‘typical’ and ‘atypical’ QRS morphologies (table 1; figure 3). There are a number of specific features that differentiate between VT and SVT, including the QRS duration, QRS axis and QRS morphology. The specific features are discussed in more detail in the next section.
An atypical QRS morphology, often assessed using leads V1/2 and V6, favours a diagnosis of VT (figure 3).5 In LBBB-like tachycardia, atypical features include an rS complex with a prolonged r wave (>30 ms), a notched downsloping S wave, slow initial conduction (onset to nadir S wave of >60 ms), all in lead V1 or V2, or a q wave in lead V6.6 In RBBB-like tachycardia, atypical features include a monophasic or biphasic QRS complex in lead V1, notched QRS with taller R (over R′) peak in lead V17 or R/S amplitude ratio <1 in lead V6.3 A Q wave, either QR or QS pattern, in lead V6 favours the diagnosis of VT in both LBBB-like and RBBB-like tachycardias. However, a triphasic QRS morphology in lead V1 indicates SVT with aberrancy.7 8
QRS activation delays
A significant delay during the initial part of the QRS complex is more compatible with VT (figure 4), whereas in SVT with aberrancy the QRS delay more frequently occurs in the mid-to-terminal portions of the QRS complex. Vereckei et al proposed that an index of slower conduction (assumed to be directly proportional to amplitude) at the initial versus terminal 40 ms portion of the QRS as evidenced by Vi/Vt ≤1 is suggestive of VT (figure 5).9
Vertical QRS axis
Extreme northwest axis (−90° to ±180°) is strongly suggestive of VT (figure 6).10 11 Left-axis deviation in the context of a RBBB-like tachycardia or right-axis deviation in the context of a LBBB-like tachycardia also supports VT.3 12 Vereckei et al reported that an initial R wave in aVR, indicative of inferior-to-superior ventricular activation, is sufficient to support the diagnosis of VT.9 Of note, this feature was not observed in patients with pre-excited tachycardias. In the presence of a baseline SR ECG, Griffith et al found that a QRS axis change of equal or more than 40° during WCT is strongly predictive of VT.13
Horizontal QRS axis
Positive concordance is defined as a positive QRS complex in all precordial leads. Negative concordance is defined as a negative QRS complex in all precordial leads. Strictly speaking, concordance is only present when all the precordial leads have monomorphic QRS complexes on the same side relative to the baseline. VT from the basal anterior left ventricle would be predicted to result in positive concordance while VT originating from the apical left or right ventricle result in negative concordance (figure 7).7 Ventricular activation via the His-Purkinje system during SVT tends to produce an overall ‘balanced’ QRS complex from leads V1–6 as ventricular activation starts from the mid-inferior interventricular septum.
In the absence of pre-existing bundle branch block, QRS duration thresholds that may indicate VT are >140 ms for RBBB-like tachycardias and >160 ms for LBBB-like tachycardias.3 The distinction between RBBB-like and LBBB-like tachycardias is based on the polarity of the QRS complex in lead V1 (RBBB-like has a positive polarity and LBBB-like has a negative polarity). Rarely, a narrower QRS is seen during tachycardia compared with SR. QRS narrowing in this context would support a diagnosis of VT.5 This phenomenon is more commonly observed in patients with pre-existing bundle branch block who have VT originating from the left ventricular septum.
QRS morphology relative to baseline
The baseline ECG may identify pre-existing bundle branch block or pre-excitation. WCT with unchanged QRS configuration in leads I, II, (±III) and V1 is virtually synonymous with SVT, whereas a different QRS configuration in these leads is suggestive of VT.14 Alternatively, an identical QRS vector during the initial 20 ms of WCT compared with SR favours SVT with aberrancy.8 As discussed previously, differentiation between capture and fusion beats also rely on a comparison with baseline ECG.
Limitations of ECG criteria
While the criteria outlined previously are of value for distinguishing between VT and SVT, it is important to note that they are associated with limitations. Apart from AV dissociation, the majority of criteria are of limited value for distinguishing between VT and SVT with antegrade conduction via an accessory pathway. Individual criteria are also associated with specific limitations. While AV dissociation is associated with a high specificity in terms of diagnosing VT, definitive AV dissociation may only be appreciated in a minority of cases (overall sensitivity of 22%).15 Furthermore, there are very rare exceptions of SVT which may demonstrate AV dissociation. In terms of analysis of ventricular activation, VT originating close to (or within) the proximal His-Purkinje system may result in early recruitment of the specialised conduction system producing ventricular activation patterns similar to SVT. Although positive or negative concordance in the precordial leads is a reliable ECG feature to distinguish between VT and SVT with aberrancy, an important caveat is that positive concordance may also be observed in SVT with antegrade AP conduction.5 In isolation, the overall QRS duration is of limited value in distinguishing between VT and SVT, particularly if a patient has pre-existing bundle branch block.
Evolution of electrocardiographic algorithms
As discussed previously, the majority of criteria to distinguish between VT and SVT have limitations, which may result in diagnostic errors.1 2 16 A number of algorithms have therefore been developed by incorporating several of the aforementioned as well as some additional criteria to increase the accuracy for diagnosing WCT (figure 8). Data on the accuracy of these algorithms (where available) are included in table 2. A brief summary of the algorithms is included below.
Wellens stepwise algorithm (1987)
The Wellens algorithm uses a series of high specificity but low sensitivity criteria in a stepwise fashion.17 If a VT criterion is fulfilled at any stage, a diagnosis of VT can be made. Therefore, based on the Wellens criteria, SVT is a diagnosis of exclusion.
Kindwall combined algorithm (1988)
Using electrophysiological evaluation of LBBB-like tachycardia, Kindwall et al described four additional ECG features suggestive of VT.6 Importantly, left-axis deviation in the presence of a LBBB-like tachycardia, which is a criterion in the Wellens algorithm, was not a useful marker for VT. The Kindwall algorithm focuses on acquisition of simultaneous recordings in multiple leads and the importance of analysing both leads V1 and V2 to avoid misclassification, as the onset of QRS complex in lead V1 is often isoelectric. If all four criteria were present, the algorithm is diagnostic of VT with a sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of 100%, 88.9%, 96.8%, 100% and 97.4%, respectively.
Brugada algorithm (1991)
Brugada et al refined the earlier algorithms in an attempt to create a simpler tool.15 The authors introduced a novel criterion of absent RS complex in all precordial leads to be suggestive of VT (ie, positive or negative concordance). They also reported that an RS interval of >100 ms was not observed in any SVT and therefore was highly specific for VT. This finding was independent of the QRS duration and morphology. The authors reported that the algorithm is diagnostic of VT with a sensitivity, specificity, PPV, NPV and accuracy of 98.7%, 96.5%, 98.4%, 97.0% and 98.0%, respectively.15
Vereckei stepwise algorithm (2007)
Vereckei et al devised a simplified stepwise algorithm without the morphology criteria.9 They also incorporated two new ECG features for VT: (1) initial R wave in aVR and (2) Vi/Vt ≤1. The algorithm was diagnostic of VT with a sensitivity, specificity, PPV, NPV and accuracy of 95.7%, 72.4%, 92.0%, 83.5% and 90.3%, respectively.
Vereckei aVR stepwise algorithm (2008)
The aVR algorithm is based on the concept of classifying VTs into two main categories based on the origin of ventricular activation and velocity of initial conduction.18 In the Vereckei aVR algorithm, the criteria of AV dissociation was removed as it did not influence the overall accuracy. Nevertheless, the authors acknowledged that given the highly specific nature of this criteria, it may still be used preceding application of their algorithm.
‘R’-wave peak time criteria (2010)
Pava et al analysed R-wave peak time (RWPT) in lead II, defined as QRS duration from the start of depolarisation until the first change of polarity, independent of whether the QRS deflection was positive or negative.19 They reported a significantly longer RWPT in VT compared with SVT, with a cut-off of 50 ms being optimal for differentiating VT from SVT. This RWPT criterion was first reported to be diagnostic of VT with a sensitivity, specificity, PPV and NPV of 93.2%, 99.3%, 98.2% and 93.3%, respectively. However, subsequent studies have put into question the sensitivity and accuracy of this single, stand-alone criterion.13 20
Limb lead combined algorithm (2019)
It is important to emphasise that a major limitation of the described algorithms is the inability of independent authors to reproduce the reported sensitivity, specificity and accuracy in the original studies.9 21–24 Chen et al proposed that this was related to the complexity of these algorithms secondary to the number of steps involved, and/or the difficulty and variability in the calculation of necessary measurements.25 The limb lead algorithm was introduced as an alternative means for simple, accurate and rapid diagnosis of VT. The algorithm is based on the concept that unlike SVT with aberrancy which relies on horizontal conduction, most VTs propagate from the superior-to-inferior portions of the ventricle and may therefore be best detected in the frontal plane as assessed with the limb leads.25 An advantage of this algorithm was that no measurements were required and hence it may be less prone to errors.
Scoring algorithm (2019)
Pachón et al devised a scoring algorithm for VT based on 7 of 16 criteria, each considered to have a high PPV.26 Using this algorithm, the presence of a QRS morphology during WCT that is identical to the QRS on baseline ECG was strongly indicative of SVT (PPV 98%). In contrast, an abnormal Q wave on the baseline ECG, AV dissociation, presence of a Q wave or initial q in lead V6 with LBBB-like tachycardia, sudden normalisation and morphology changes in patients with AF on baseline ECG, WCT with complete or high-grade AV block, and contralateral BBB morphology in patients with organic BBB were each strongly associated with VT.
Limitations of electrocardiographic algorithms
There are practical limitations to the application of conventional algorithms in clinical practice, especially while clinicians are under high-pressured situations. Moreover, the studies developing the algorithms have commonly involved invasive electrophysiolological studies and/or expert interpreters who were removed from the clinical care setting, and thus generalisability of the findings may be questionable.
In terms of limitations of specific algorithms, while the Wellens stepwise algorithm provides a systematic approach in the assessment of WCT, it has not been validated in independent studies and the overall accuracy remains questionable.17 The Kindwall algorithm may be associated with difficulties in determining the interval from onset of QRS complex to nadir of the S wave.6 Furthermore, the algorithm has not been tested in patients with an AP. The Brugada algorithm introduced several changes to the aforementioned algorithms while maintaining the morphological criteria. This contributed to a significant disadvantage in the Brugada algorithm as more than one-third of patients had discordance in their morphological criteria.15 An important limitation of the Vereckei stepwise algorithm is that the third and fourth criteria were unable to reliably exclude SVT with antegrade AP conduction.9 The Vereckei aVR stepwise algorithm was tested in a relatively small proportion of patients without structural heart disease and may therefore not apply to this cohort.18 There may be difficulties in applying the RWPT criteria as the initiation and peak of ventricular complexes can be hard to define.19 Furthermore, the Brugada and Vereckei aVR algorithms were demonstrated by independent authors to substantially underperform compared with the original studies.21 23 27 28 Finally, the limb lead combined algorithm may misclassify VTs originating from the conduction system or intracavitary structures (eg, papillary muscles),25 while the scoring algorithm has not been prospectively tested and used induced tachycardias in some patients.26
Alternative diagnostic methods
Despite improvements in ECG algorithms for the evaluation of WCT, there remain significant limitations with traditional models that rely on subjective interpretation and provide insufficient information regarding the probability of VT. Hence, alternative strategies have been developed.
Jastrzebski et al described a VT scoring system based on seven ECG features: initial R wave in lead V1; initial r >40 ms in lead V1/2; notched S wave in lead V1; initial R wave in aVR; R wave peak time ≥50 ms in lead II; no RS complex in leads V1–6; AV dissociation.29 The model was highly specific (99.6%) for VT with a score of ≥3, despite the lack of sensitivity (56.9%). An advantage of this approach is that it considers a comprehensive list of ECG features of VT in each patient before proposing a diagnosis. Unlike previous methods, the VT score was designed to ‘rule in’ VT.
The WCT formula evaluates the WCT QRS duration and changes in amplitude (frontal and horizontal) during tachycardia compared with the baseline ECG.30 By using a VT probability partition of 50%, this method produces a sensitivity, specificity, PPV, NPV and accuracy of 89.7%, 92.9%, 89.7%, 92.9% and 91.5%, respectively. An advantage of this approach was the ability to incorporate the formula into automated analysis algorithms, which may reduce inter-observer variability. Furthermore, it provides an estimation of the likelihood of VT. Kashou et al developed a new automated means to differentiate WCTs (WCT Formula II).31 This was a system that was independent of the clinicians’ ECG interpretation competency and, like the original WCT formula, could be integrated into contemporary ECG software. Of note, the WCT formula does however rely on pairing of WCT and baseline ECGs (with QRS duration <120 ms and heart rate <100 bpm), which may not always be possible.
VT prediction model
May et al also recently described an automated VT prediction model that used readily available ECG measurements to distinguish VT from SVT.32 The model places more emphasis on the baseline ECG to arrive at the correct rhythm diagnosis. This system delivers an estimated probability of VT that may assist with the decision-making process. A benefit of the VT prediction model is that it does not have to be integrated into existing ECG interpretation softwares (unlike the WCT formula) but rather its use could be supported by mobile device applications or web-based platforms. However, further studies are still needed to refine this tool.
Independent studies of the various algorithms for differentiating VT and SVT have failed to reproduce results from the original studies. Overall, each algorithm has only been found to be modestly accurate for classification purposes.24 Drew and Scheinman found that even in the best of circumstances, 1 in 10 WCT ECGs defy differentiation.33 When using a Bayesian approach, an assumption of the likelihood of an event occurring is first made based on prior knowledge and subsequently modified as additional information is acquired.34 Advantages of applying the Bayesian inference are that all relevant ECG features may be given due consideration (in contrast to the hierarchical nature of most algorithms which discards features lower in the diagnostic tree once a diagnosis is made) and it allows the problem of imperfect ascertainment of a particular ECG feature to be circumvented.35 Adoption of the Bayesian approach in a study by Lau et al demonstrated significant improvement compared with the clinical assessment from three separate, independent cardiac specialists using the various algorithms.35 However, the practical limitations of this approach include the use of multiple criteria and serial calculations. Furthermore, as each criterion may not be entirely independent, the final tabulated results may not accurately reflect the true likelihood of VT.
Important considerations and pitfalls in a clinical situation
There are a number of important factors to consider when assessing a WCT in a clinical situation. First, clinical features are of limited value in distinguishing VT from SVT. While haemodynamic status may be a useful distinguishing marker,16 it should be emphasised that a significant proportion of patients with VT may have no significant haemodynamic compromise.36 In addition, some patients with SVT may present with haemodynamic compromise. Second, a pre-existing history of structural cardiac disease increases the pre-test probability of VT.37 Of note however, the absence of these conditions has a poor negative predictive value.37 Third, the presence of clinical or electrocardiographic features of VT are indicative of the condition, but their absence does not reliably exclude it. Fourth, the value of a baseline ECG should not be underestimated,38 as it can be useful to determine capture and fusion beats, pre-existing bundle branch block, QRS axis change, QRS morphology change and underlying ischaemic heart disease. Finally, the features and algorithms described previously should be interpreted with caution in patients with anti-arrhythmic medications, severe electrolyte disturbance, medication-induced arrhythmia, known congenital heart disease or prior cardiac instrumentation (surgery or ablation).39 Furthermore, its applicability in the immediate post-cardiac arrest or cardioversion stage remains limited.39
Accurate diagnosis of VT and SVT from an ECG requires a good understanding of the underlying principles of physiological cardiac activation and how this differs in specific pathologies. In this regard, numerous algorithms are available for assistance although each has its own inherent limitations and low reproducibility.
In the context of a wide complex tachycardia:
Based on the differences in ventricular activation patterns, a range of different electrocardiographic (ECG) criteria can be used to distinguish between ventricular tachycardia (VT) and supraventricular tachycardia with aberrant conduction.
While some ECG criteria are strongly predictive of a diagnosis of VT, the majority of criteria are not diagnostic of VT in isolation.
Various algorithms that combine multiple ECG criteria have been developed to improve the diagnostic value.
While the existing diagnostic algorithms have been reported to enhance ECG-based diagnosis, reproducibility of diagnostic algorithms between studies has been limited and some algorithms do not lend themselves to rapid ECG diagnosis.
CME credits for Education in Heart
Education in Heart articles are accredited for CME by various providers. To answer the accompanying multiple choice questions (MCQs) and obtain your credits, click on the 'Take the Test' link on the online version of the article. The MCQs are hosted on BMJ Learning. All users must complete a one-time registration on BMJ Learning and subsequently log in on every visit using their username and password to access modules and their CME record. Accreditation is only valid for 2 years from the date of publication. Printable CME certificates are available to users that achieve the minimum pass mark.
Patient consent for publication
Contributors WYD performed the literature search and drafted the manuscript. SM provided critical revisions. All authors approve the final version of the manuscript for publication.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Commissioned; externally peer reviewed.
Author note References which include a * are considered to be key references.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.