Wide complex tachycardia: differentiating ventricular tachycardia from supraventricular tachycardia

Authors, Journal, Affiliations, Type, DOI

• Authors: Wern Yew Ding, Saagar Mahida
• Journal: Heart 2021;107:1995–2003 (Education in Heart)
• Affiliations: Liverpool Centre for Cardiovascular Science, University of Liverpool and Liverpool Heart & Chest Hospital, Liverpool, UK
• Type: Narrative educational review
• DOI: 10.1136/heartjnl-2020-316874

Overview

This Education in Heart review explains how the activation pattern of a wide QRS complex tachycardia (WCT, QRS >120 ms) generates the ECG features used to separate ventricular tachycardia (VT) from supraventricular tachycardia (SVT) with aberrancy, SVT conducted over an accessory pathway, ventricular pacing and artefact. It organises the evidence into AV-relationship criteria, QRS morphology, activation delay, axis, concordance and duration, then traces the evolution of the Wellens, Kindwall, Brugada, Vereckei, Pava, limb-lead and Pachón algorithms. The review is explicit that most individual criteria are not diagnostic in isolation, that the published algorithm performance figures have often not been reproduced, and that roughly one in ten WCTs defies reliable ECG differentiation. It closes with the VT score, automated WCT formula and Bayesian approaches and with practical pitfalls, including the limited value of haemodynamic status and the importance of a baseline ECG.

Keywords

Wide complex tachycardia; ventricular tachycardia; supraventricular tachycardia; AV dissociation; ECG algorithm; capture and fusion beats

Key Takeaways

Causes and activation principles

• The differential diagnosis is VT, SVT with bundle branch block aberrancy, SVT with antegrade accessory-pathway conduction, ventricular pacing and ECG artefact.
• AV dissociation — more P waves than QRS complexes — is strongly suggestive of VT with specificity approaching 100%, but is visible in only about one fifth of VTs (sensitivity 22%). A 1:1 P–QRS relationship does not exclude VT because retrograde VA conduction is common.
• Capture beats (sudden narrowing to the baseline QRS) and fusion beats (a hybrid QRS from collision of atrial and ventricular activation) are specific for AV dissociation but require a baseline ECG for comparison.
• In SVT with aberrancy, the ventricle is activated through the His–Purkinje system, producing a typical RBBB or LBBB pattern. In VT, slow initial myocyte-to-myocyte conduction produces atypical morphology and a delayed initial deflection.

Morphology, intervals and axis

• LBBB-like tachycardia suggesting VT: rS in V1/V2 with an r wave >30 ms, a notched downsloping S wave, onset-to-nadir S >60 ms, or a q wave in V6.
• RBBB-like tachycardia suggesting VT: monophasic or biphasic V1, a notched V1 with the taller peak on the left ("bad rabbit ears"), R/S amplitude ratio <1 in V6, or any Q/QR/QS in V6. A triphasic V1 pattern favours SVT with aberrancy.
• Initial versus terminal delay: VT delays the initial QRS, whereas aberrancy delays the mid-to-terminal QRS. Vi/Vt ≤1 (Vereckei) indicates slower initial conduction and therefore VT.
• Axis: northwest axis (−90° to ±180°) is strongly suggestive of VT; left-axis deviation with an RBBB-like pattern or right-axis deviation with an LBBB-like pattern also supports VT. An initial R wave in aVR is a Vereckei criterion. A QRS-axis change ≥40° from baseline is strongly predictive of VT (Griffith).
• Concordance: positive or negative concordance across V1–V6 favours VT, but positive concordance can also occur with antidromic AVRT over a left-sided accessory pathway.
• QRS duration: >140 ms for RBBB-like and >160 ms for LBBB-like tachycardias suggests VT in the absence of pre-existing bundle branch block; QRS narrowing during tachycardia compared with sinus rhythm also favours VT.

Likelihood ratios for individual features

Adapted from Lau et al.:

Algorithms

• Wellens (1987): stepwise high-specificity criteria; SVT is a diagnosis of exclusion. Not independently validated.
• Kindwall (1988): four LBBB-like criteria; with all four present, sensitivity 100%, specificity 88.9%, accuracy 97.4%. Left-axis deviation was not useful in LBBB-like tachycardia.
• Brugada (1991): absence of an RS complex in all precordial leads, RS interval >100 ms, AV dissociation, then morphology. Originally reported sensitivity 98.7% and specificity 96.5%.
• Vereckei (2007): AV dissociation, initial R in aVR, Vi/Vt ≤1; reported sensitivity 95.7%, specificity 72.4%. The 2008 aVR-only version had similar accuracy (96.5%/75.0%).
• Pava (2010): R-wave peak time ≥50 ms in lead II; originally sensitivity 93.2%, specificity 99.3%, but subsequent studies questioned the standalone performance.
• Chen limb-lead algorithm (2019): monophasic R in aVR, or negative QRS in I/II/III, or opposing limb-lead concordance; no measurements required; sensitivity 87.2%, specificity 90.8%.
• Pachón scoring (2019): seven high-PPV criteria, four of which compare the WCT with the baseline ECG; identical morphology to baseline strongly favours SVT (PPV 98%).

The review warns that these figures come from the original derivation studies and have not been reproduced by independent authors.

Alternative approaches

• Jastrzebski VT score: seven features; score ≥3 has specificity 99.6% but sensitivity only 56.9% — designed to rule in VT.
• WCT formula (2019): QRS duration and frontal/horizontal amplitude change versus baseline; sensitivity 89.7%, specificity 92.9%.
• WCT Formula II (2020): percent time-voltage area change; automated and clinician-independent.
• VT prediction model (2020): uses readily available measurements and emphasises the baseline ECG; can run on mobile or web platforms.
• Bayesian approach (Lau 2000): multiplies pretest odds by individual likelihood ratios, retaining all features rather than stopping at the first positive criterion; outperformed three cardiologists in the original study but requires serial calculation.

Clinical pitfalls

• Haemodynamic compromise does not reliably separate VT from SVT: many patients with VT remain stable, and some with SVT decompensate.
• A history of structural heart disease raises the pretest probability of VT, but its absence has poor negative predictive value.
• The absence of VT features does not exclude VT.
• A baseline ECG is essential for capture/fusion beats, pre-existing bundle branch block, axis change, morphology change and prior infarction.
• Criteria are unreliable in the presence of class I/III antiarrhythmic drugs, severe electrolyte disturbance, congenital heart disease, prior cardiac surgery or ablation, and in the immediate post-arrest or post-cardioversion state.

Limitations of the document

• Narrative review, no new data. The performance tables reproduce the original derivation studies; the authors themselves note that independent validation has often failed.
• Derivation bias. Most algorithms were developed in electrophysiology-laboratory populations with a high VT prevalence and expert interpreters, so accuracy in emergency-department or community practice is uncertain.
• Incomplete coverage of special populations. Fascicular VT, bundle-branch re-entrant VT, pre-excited tachycardia and drug-induced widening are acknowledged but not given algorithm-specific solutions.
• No prospective comparison. The review does not provide a head-to-head prospective trial of algorithms, and the 2021 publication predates newer machine-learning ECG classifiers.
• "One in ten defies differentiation" is an expert estimate rather than a measured pooled statistic.

Key Concepts Mentioned

• concepts/Wide-Complex-Tachycardia — the parent topic
• concepts/Fascicular-Ventricular-Tachycardia — the main exception to morphological algorithms
• concepts/AVRT-Accessory-Pathway — antidromic AVRT and pre-excited AF
• concepts/SVT-Management — treatment once SVT is diagnosed
• concepts/VT-Management — treatment once VT is diagnosed

Key Entities Mentioned

• None

Wiki Pages Updated

• Created: wiki/sources/wct-bmj-2021.md
• Updated: wiki/concepts/Wide-Complex-Tachycardia.md
• Updated: wiki/sourceindex.md
• Updated: log.md