Author + information
- Received July 25, 2016
- Revision received December 9, 2016
- Accepted January 12, 2017
- Published online July 17, 2017.
- Anil V. Yadav, MDa,
- Babak Nazer, MDb,
- Barbara J. Drew, PhDb,
- John M. Miller, MDa,
- Hicham El Masry, MDc,
- William J. Groh, MDa,
- Andrea Natale, MDd,
- Nassir Marrouche, MDe,
- Nitish Badhwar, MDb,
- Yanfei Yang, MDb and
- Melvin M. Scheinman, MDb,∗ ()
- aKrannert Institute of Cardiology, Indiana University School of Medicine, Indianapolis, Indiana
- bDivision of Cardiology, University of California San Francisco School of Medicine, San Francisco, California
- cCardiology Division, Saint Vincent Heart Center, Indianapolis, Indiana
- dTexas Cardiac Arrhythmia Institute, Austin, Texas
- eDivision of Cardiovascular Medicine, University of Utah School of Medicine, Salt Lake City, Utah
- ↵∗Address for correspondence:
Dr. Melvin M. Scheinman, University of California-San Francisco, 500 Parnassus Avenue, San Francisco, California 94118.
Objectives This study sought to determine the ability of conventional electrocardiographic (ECG) criteria to correctly differentiate idiopathic ventricular tachycardia (VT) from supraventricular tachycardia (SVT) with aberrancy.
Background Previously reported VT ECG criteria were developed from cohorts of patients with structural heart disease and have not been applied to patients with idiopathic VT.
Methods ECGs of 115 idiopathic VTs, 101 post-myocardial infarction (MI) VTs, and 111 wide QRS SVTs were analyzed using standard criteria. VT was diagnosed in patients when at least 1 criterion was met, SVT when no criteria were met, and indeterminate when there were conflicting criteria.
Results Standard ECG criteria more frequently diagnosed VT in the post-MI group than the idiopathic group (95% vs. 82%, respectively; p < 0.01). Diagnosis in only 12 of the 111 SVT patients (11%) met the criteria for VT. All patients in the idiopathic VT group with right branch bundle block morphology who did not meet VT criteria demonstrated an rsR′ pattern in V1 (consistent with SVT). Among idiopathic VT patients, Purkinje-associated VT had the lowest sensitivity for correct VT diagnosis in 13 of 23 patients (57%), septal sites of origin were correctly diagnosed in only 56 of 76 patients (74%), whereas nonseptal sites had a high sensitivity in 35 of 35 patients (100%; p < 0.005).
Conclusions Conventional ECG criteria have reduced sensitivity to distinguish VT from SVT with aberrancy in patients with idiopathic VT. This is most pronounced in VT originating from septal sites, particularly Purkinje sites and the septal outflow tract regions. Clinicians should be aware that application of conventional ECG criteria in idiopathic VT may underdiagnose VT.
Accurate electrocardiographic (ECG) diagnosis of wide complex tachycardia (WCT) as either supraventricular tachycardia (SVT) with aberrancy or ventricular tachycardia (VT) is essential in the management of cardiac arrhythmias. In this regard, numerous ECG criteria have been developed to distinguish WCTs as either SVT with aberrancy or VT with a high degree of sensitivity and specificity in the correct diagnosis of VT (1–4). However, these prior studies did not differentiate between VT associated with structural heart disease (SHD) and idiopathic (no structural disease) VT. These studies instead either pooled patients with a variety of cardiac conditions or did not precisely specify the patient cohort. The purpose of our study was to assess the ability of established 12-lead ECG VT criteria to correctly distinguish idiopathic VT from SVT with aberrancy.
We studied a series of 115 patients who presented at 3 centers with sustained or nonsustained WCT or frequent monomorphic premature ventricular complex (PVC; idiopathic group). All patients underwent, at minimum, an examination consisting of a complete history, a physical examination, a 12-lead ECG, and an echocardiogram that demonstrated no evidence of SHD prior to electrophysiology study (EPS). Where appropriate, coronary angiography, right ventricular (RV) angiography, and/or cardiac magnetic resonance was performed, all of which were interpreted as demonstrating normal results.
In addition, 42 consecutive patients with a known history of myocardial infarction (MI) and 101 distinct WCTs served as a comparison (post-MI group). Finally, a control group of 111 ECGs from a cohort of patients with WCT known to show SVT with aberrancy based on results of EPS was used to assess the specificity of ECG criteria (SVT group).
The research protocol was approved by the institutional review boards at all 3 centers. All patients’ related data were de-identified and stored in a password-protected database that was accessible only to the investigators.
All patients gave written informed consent. The patients underwent EPS in the fasting state after antiarrhythmic drugs were withdrawn (for at least 4 half-lives). Catheters were placed through the femoral veins, and EPS was performed. WCT occurred spontaneously or was induced by burst pacing, programmed electrical stimulation with up to 3 extrastimuli and/or isoproterenol infusion. Ventricular tachycardia was diagnosed using standard EPS maneuvers, and electroanatomic mapping and RF catheter ablation were then performed based on the mechanism of VT. For idiopathic VTs, site or origin was reported by the physician as the site of earliest local activation, best pace map, or successful RF termination site. For post-MI VTs, site of origin was reported as presumed VT exit site based on mapping or successful RF termination site. Success of ablation was confirmed by ventricular pacing with up to 3 extrastimuli both with and without isoproterenol (consistent with what was required for initiation prior to ablation).
Twelve-lead ECGs of all WCTs representing each patient’s clinical VT or SVT were interpreted by 2 independent cardiologists blinded to the clinical data. Each patient’s tachycardia was categorized as either left bundle branch block (LBBB) or right bundle branch (RBBB) morphology. Criteria by Brugada et al. (1), Drew and Scheinman (2), Kindwall et al. (3), Wellens et al. (4), and Vereckei et al. (5) (hereafter referred to as “conventional criteria”) were applied to categorize each ECG as follows: VT if 1 or more criteria for VT were met and no SVT criteria were met; SVT if no VT criteria were met and SVT morphological criteria were met; and indeterminate if 1 or more of both VT and SVT morphological criteria were met.
Criteria favoring VT
1. Atrioventricular (AV) dissociation or VA block;
2. QRS width of >0.16 s;
3. Right superior axis (−90° to ±180°);
4. Absence of precordial RS complexes or, if present, an RS >100 ms (QRS onset to S nadir measurement);
5. QRS morphology in lead V1 in RBBB-type tachycardia:
a. Monophasic R;
b. Notched waveform with a taller left peak (Rr′);
c. Biphasic RS or QR pattern (Figure 1);
6. QRS morphology in LBBB-type tachycardia equal to any 1 of the following in lead V1 or V2:
a. R-wave >30 ms;
b. Slurred or notched S wave downstroke;
c. Delayed S nadir (QRS onset to S nadir >60 ms);
7. QRS morphology in lead V6:
a. QS pattern;
b. rS pattern (R-to-S <1) in RBBB-type tachycardias;
c. QRS onset to predominant peak or nadir in lead V6 of ≥70 ms (1,3);
8. Presence of an initial R-wave in AVR and width of an initial r or q wave >40 ms (5).
Criteria favoring SVT
1. QRS morphology in lead V1 in RBBB-type tachycardia equal to rR′ or rsR′;
2. QRS morphology in lead V1 in LBBB-type tachycardia:
a. r-Wave width <30 ms;
b. Smooth S-wave down stroke;
c. QRS onset to S nadir <60 ms and no Q in lead V6;
3. QRS morphology in lead V6:
a. Intrinsicoid deflection ≤50 ms;
b. Triphasic qRs pattern (R-to-S >1) in RBBB-type tachycardias.
Sensitivity was defined as the proportion of WCTs which were correctly diagnosed as VT by ECG criteria compared to the gold standard of EPS diagnosis. Comparison of sensitivities was performed using chi-square tests, with p value of <0.05 considered statistically significant.
In the idiopathic group, 79 patients were male and 36 were female. Each patient demonstrated only 1 VT morphology. Sustained monomorphic VT was present in 73 patients, and nonsustained monomorphic VT and/or frequent monomorphic ventricular ectopic beats were present in 42 patients. LBBB morphology was present in 76 patients, whereas 39 had RBBB morphology (Table 1).
In the post-MI group, 37 patients were male and 5 were female. All patients had prior MI as shown by echocardiographic segmental wall motion abnormalities and/or significant coronary arterial disease. Sustained monomorphic VT was present in 99 of 101 morphologies. A LBBB pattern was present in 38, whereas 63 had RBBB morphology (Table 1).
Idiopathic VT ECG analysis
The mean QRS duration of VT (widest measured in any lead) was 139 ± 20 ms. The overall sensitivity to correctly diagnose VT using the “conventional criteria” in patients with idiopathic VT was 82% (Table 2). Sensitivity was greater with LBBB than RBBB VTs when the indeterminate classification was included (Table 2).
In the 76 patients with idiopathic VT and LBBB morphology, 63 of 76 patients (83%) VT was correctly diagnosed (Figure 2). In the remaining 13 of 76 patients (17%), SVT was incorrectly diagnosed in 12 patients (Figure 3). One patient’s condition was classified as indeterminate. In the 39 patients with idiopathic VT and a RBBB morphology, 31 of 39 patients (79%) received correct diagnoses of VT. The remaining 8 patients (21%) received indeterminate diagnoses (Figure 4).
LBBB idiopathic VTs were incorrectly diagnosed as SVT because of an absence of any of the conventional VT criteria (1–4). Most of the correctly diagnosed LBBB idiopathic VTs met morphologic criteria in lead V1 (VT criteria 6 above), usually an initial R-wave >30 ms or an onset of R to nadir of the S-wave >60 ms (Figure 3). Thus, the 12 LBBB idiopathic VTs that were incorrectly diagnosed as SVT lacked this otherwise common VT criteria.
In the 8 patients classified as indeterminate with RBBB morphology, the reason was due to a conflict in morphologic criteria between leads V1 and V6 (VT criteria 5 and 7 described above): rsR′ pattern in V1 (consistent with SVT) was associated with a lead V6 R-to-S ratio <1 (consistent with VT [Figure 4]). These 8 patients were subsequently found to have Purkinje-associated VT.
Post-MI VT ECG analysis
The mean QRS duration was 158 ± 20 ms and significantly wider than that in the idiopathic group (p < 0.001). The overall sensitivity to correctly diagnose VT in this group was 95% (Table 2). When ECGs classified as indeterminate were excluded from the analysis, the sensitivity to diagnose VT was 99% in the remaining 97 VT morphologies. The sensitivity to diagnose VT did not differ in the RBBB group (61 of 63 patients [97%]) compared with the LBBB group (35 of 38 patients [92%]; p = NS), even when the indeterminate tracings were excluded (RBBB: 61 of 61 patients [100%] vs. LBBB: 35 of 36 patients [97%]; p = NS]). Of the 4 VTs classified as indeterminate, 2 demonstrated each type of morphology (LBBB or RBBB). The 2 VTs diagnosed as indeterminate with an RBBB morphology had an rsR′ pattern in V1 with right axis deviation. In the remaining 2 VTs diagnosed as indeterminate with an LBBB pattern, some of the morphological characteristics of the QRS complex for VT were present (such as a qR or qS pattern in V4 and/or V5), but the Kindwall criteria were not met. A diagnosis of indeterminate was assigned with the possibility that SVT with prior MI could not be ruled out in these 2 patients with LBBB patterns.
Analysis of SVT ECG
Among 111 ECGs diagnostic for SVT (based on EPS), 12 were classified incorrectly as VT (11%), whereas 4, all with RBBB morphology, were classified as indeterminate (3%). The remaining 95 of these ECGs failed to satisfy any of the criteria for VT and were classified as SVT (specificity: 86%).
Sensitivity of VT criteria
The sensitivity of ECG criteria for VT was lower in the idiopathic group (82%) than in the post-MI group (95%; p <0.01) (Table 3). These results remained consistent when the indeterminate ECGs from the analysis were excluded (88% vs. 99%, respectively; p < 0.01) (Table 2).
Because some of the patients in the idiopathic VT group had only frequent PVC (n = 8), a source of statistical bias might have been AV dissociation criteria, which could not be used in this subset. In a sensitivity analysis, these patients were excluded, as were the ECG criteria regarding AV dissociation, and sensitivity in the idiopathic group only decreased from 82% to 80% (86 of 107 patients). Although all ECGs in the post-MI group demonstrated VT and not PVC, exclusion of AV dissociation criteria was also similarly applied, minimally changing sensitivity from 95% to 96% (89 of 93 patients).
Origin of VT site
In the idiopathic group, the site of origin was available for 111 of the 115 patients. These sites were divided into 6 groups: right ventricular outflow tract (RVOT), left ventricular outflow tract (LVOT), Purkinje-associated sites (fascicular; mid-to-distal left ventricular [LV] septum), aortic cusp, or other RV and LV sites. A majority of the idiopathic VTs were ablated in the RVOT region (58 of 111 patients [52%]), most of which were in the anteroseptal RVOT (36 of 58 patients [62%]), followed by the posteroseptal RVOT (14 of 58 patients [24%]). The second most common site of idiopathic VT was associated with the Purkinje fibers in the fascicles or mid-distal septal region of the LV (23 of 111 patients [21%]). LVOT sites consisted of 9 of 111 patients (8%), and aortic cusp VT contributed 7 of 111 patients (6%). The remaining patients (14 of 111 [13%]) represented other sites within the LV and RV.
There was site-specific heterogeneity in sensitivity of VT ECG criteria: VT was correctly diagnosed in 48 of 58 RVOT sites (83%), and only 13 of 23 Purkinje-associated VT sites (57%; p < 0.03). Within the RVOT group, VT was correctly diagnosed in 29 of 36 anteroseptal sites (78%) and in 20 of 22 other RVOT sites (91%; p = 0.46). Idiopathic VT of any nonseptal origin (i.e., RVOT of free wall origin, LVOT, aortic cusps, and other LV and RV sites) was correctly diagnosed in 35 of 35 patients (100%). In contrast, idiopathic VT from all septal sites (i.e., LV septum, anteroseptal, midseptal, or posteroseptal RVOT) was correctly diagnosed in only 56 of 76 patients (74%; p < 0.005: compared with nonseptal sites) and all RVOT septal sites in only 43 of 53 patients (81%; p < 0.03: compared with nonseptal RVOT sites).
QRS notching, an ECG sign associated with myocardial activation delay and scar tissue, was more significant in post-MI VT than in outflow tract VT (89% vs. 65%, respectively; p = 0.001). Among the idiopathic group, notching was most frequently seen in Purkinje-associated VT (87%).
Although idiopathic VT is usually associated with a benign prognosis, the delay in establishing the correct diagnosis may lead to inappropriate therapy, as idiopathic VT can be readily treated with either appropriate medical therapy or RF catheter ablation, and implantable cardioverter-defibrillators are usually not warranted in this patient population (6). Our findings indicate that nearly 20% of idiopathic VTs may not be recognized as ventricular in origin, using conventional criteria.
Criteria for the differentiation of SVT with aberrancy from VT have been associated with a high degree of sensitivity and specificity (1). These criteria were developed from a cohort of patients largely with ischemic heart disease. Thus these criteria would be expected to be more accurate in the diagnosis of VT associated with SHD as opposed to idiopathic VT.
The conventional criteria demonstrate an inherent bias toward the diagnosis of VT associated with SHD. Specifically, the absence of any R wave in any precordial lead to assess for negative concordance in the precordial leads would favor anterior or apical exit sites from the LV. Idiopathic VTs are unlikely to have negative concordance because they typically originate from the LV septum or right or left outflow tracts. Although this criterion is very specific for VT, it occurs so rarely in idiopathic VT that it is of little diagnostic value (7). In addition, the criteria regarding the onset of R to nadir of S >100 ms also tends to favor VT associated with SHD, as VT with SHD tends to have a wider QRS than idiopathic VT as demonstrated in previous studies (8). This may be due to the fact that VT due to SHD is associated with myocardial scarring and fibrosis that contributes to delayed activation of the ventricles. In contrast, patients with idiopathic VT have no SHD and are generally younger. Hence, idiopathic VT would be expected to have more rapid global ventricular activation and a narrower QRS morphology than in VT with SHD (9,10). Accordingly, our data also demonstrate a shorter QRS duration for idiopathic VT than for post-MI VT. This narrower QRS duration may lead to a reduced sensitivity to diagnose VT by conventional QRS width criteria.
Although AV dissociation is 1 of the most specific ECG features for diagnosing VT, its sensitivity is poor as AV dissociation has been shown to occur in only approximately 20% of patients with VT (11). Conversely, up to 30% of VTs have 1:1 retrograde VA conduction (11). Our series demonstrated AV dissociation in only 8 of 115 patients (7%) of our idiopathic group, possibly due to greater likelihood of VA conduction among younger patients without SHD.
Bundle branch block morphology criteria in leads V1 and V6 have been established to differentiate VT from SVT (3,7). It has been demonstrated that, in RBBB pattern VT, an Rsr′ morphology in V1 favors VT (7). In our analysis, all patients with idiopathic VT and an RBBB pattern who were classified as indeterminate had an rsR′ pattern in V1, consistent with SVT, whereas lead V6 demonstrated an R-to-S ratio of <1.0, consistent with VT. Ventricular activation in an SVT with an RBBB-type aberration first proceeds down the left bundle branch with delayed activation of the RV, providing an rsR′ pattern in V1. In SHD VT with an RBBB pattern the exit site in the LV often produces an initial vector pointing toward the RV, producing a large initial R-wave. Subsequent activation of the remaining right and left ventricles would produce a smaller ′r′ wave, leading to the Rsr′ pattern often noted in V1 with SHD-related VT. This may not apply to idiopathic VT, where initial activation is unlikely to come from LV myocardium and may not produce a large initial R-wave in V1.
Our findings showed that, in idiopathic VT with an RBBB pattern classified as indeterminate, all patients had fascicular VT. In these patients there is early activation of the His-Purkinje system, initial activation of the posteromedial LV, and late activation of the anterolateral LV, which explains the markedly superior axis and the r/S pattern in V6. The pattern of WCT with a relatively narrow QRS with an rsR′ pattern in V1 and an r/S in V6 should immediately suggest this type of idiopathic VT.
Idiopathic patients with LBBB VT who were misclassified as SVT by ECG criteria most often had septal RVOT VT. Foci from this region of the RV would also tend toward simultaneous activation of the ventricles producing a relatively narrow QRS complex. In addition, our findings are in agreement with those of Krebs et al. (12), who reported that some VTs with a focus in the RVOT lack an initial r-wave in V1. These factors may have contributed to the reduced sensitivity to diagnose VT by conventional morphological criteria.
Prior studies pooling patients with and without prior MI have demonstrated a wide variation in the sensitivity to diagnose VT (12,13). Brugada et al. (1) noted a sensitivity of 98.7% and a specificity of 96.5% for the diagnosis of VT in 554 patients when similar conventional criteria were used. In that study, there was a preponderance of patients with SHD. Similarly, our 101 ECGs in patients with SHD demonstrated a sensitivity of 95% to correctly diagnose VT. Only the previous study by Vereckei et al. (14) systematically assessed sensitivity of VT criteria for idiopathic VT: a VT cohort of 348 VTs included 38 idiopathic VTs in whom sensitivity of their proposed ECG criteria for VT was 86.5% (similar to our finding of 82%). Although that study did not directly compare this with sensitivity in SHD VT patients, the criteria’s sensitivity among their entire VT cohort was 95.7% (similar to our finding of 95% among SHD patients).
One limitation of our study is the lack of a prospective application of the ECG criteria to patients presenting with previously undiagnosed WCTs. As a historical cohort was used, many of the baseline characteristics of the SVT patients are not known, limiting a more thorough comparison with this group. There were no patients in our cohort with pre-excited tachycardias, epicardial VTs (which may be idiopathic or due to structural heart disease), and wide complex tachycardia due to class I antiarrhythmics, reducing generalizability of our data to these populations. Our blinded ECG analysis was performed by 2 cardiologists in a rigorous manner, using the conventional criteria and the algorithm described above, but it is possible that sensitivity and specificity for VT may differ if analyzed by clinicians with less experience in ECG analysis. As our data from 3 centers were analyzed centrally and in an anonymous manner, we were not able to perform statistical analysis for among-center variance. Another limitation of the study was the inability to use AV dissociation criteria fully in the idiopathic VT group as 8 of 115 patients (7%) had only isolated PVC, although a sensitivity analysis was performed. Finally, a Bayesian approach, as demonstrated by Lau et al. (15), using likelihood ratios instead of our algorithmic approach to VT versus SVT diagnosis might have affected sensitivity and specificity, as they noted an increased sensitivity for fascicular VT when Bayes′ theorem was used compared with clinician judgment. Future research using this or other cohorts may identify novel ECG criteria that may specifically be used to differentiate idiopathic VT from SVT.
In patients with idiopathic VT, conventional criteria to distinguish VT from SVT have a lower sensitivity (82%) for the diagnosis of VT. In contrast, in our cohort of patients with post-MI VT, the ability to correctly diagnose VT was 95%. This reduced sensitivity to diagnose idiopathic VT was noted for Purkinje-associated VTs and VT with septal sites of origin and was similar in patients regardless of bundle branch block morphology. In particular, the reduced sensitivity to diagnose VT with RBBB morphology is predominantly due to conflicting criteria, with an rsR′ pattern in V1 which favors SVT and an R-to-S ratio of <1 in V6, favoring VT. In contrast, in patients with LBBB morphology VT, the conventional criteria are occasionally not met, typically because of a lack of an initial R of >30 ms or rS of >60 ms in V1 or the absence of a Q-wave in V6. This leads to an erroneous diagnosis of SVT.
COMPETENCY IN MEDICAL KNOWLEDGE: Conventional criteria to distinguish VT from SVT with aberrancy were developed from cohorts of patients with ischemic and structural heart disease. Sensitivity of these criteria to correctly diagnose VT are lower for idiopathic VT (82%) than for post-MI VT (95%).
COMPETENCY IN PATIENT CARE: When evaluating wide complex tachycardias among patients without underlying structural heart disease, conventional criteria should be used with caution, particularly RBBB V1 rSR′ (conventionally associated with SVT but present in 21% of RBBB idiopathic VTs) and LBBB V1 or V2 R of >30 ms or QRS onset to S nadir of >60 ms (conventionally associated with VT but absent in 17% of our idiopathic VT cohort).
TRANSLATIONAL OUTLOOK: Future criteria specific to idiopathic VT may be developed by excluding criteria that rely on QRS duration (a marker of slow ventricular conduction due to myocardial fibrosis from structural heart disease) and RBBB V1 morphology (Rsr′ vs. rsR′).
Dr. Natale has received speaker honoraria from Boston Scientific, Biosense Webster, Medtronic, and St. Jude Medical. Dr. Marrouche has ownership interest in Marrek, Inc., and Cardiac Designs; has received contracted research funding from Biosense-Webster, Medtronic, St. Jude and Boston Scientific; and has received consulting fees from Preventice and Biotronik. Dr. Scheinman has received speakers honoraria from Medtronic; and has financial relationships with St. Jude Medical, Biotronik, and Biosense Webster; and serves on Amgen data monitoring committee. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Drs. Yadav and Nazer contributed equally to this work.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- electrophysiology study
- left bundle branch block
- left ventricle
- left ventricular outflow tract
- myocardial infarction
- premature ventricular complex
- right bundle branch block
- right ventricle
- right ventricular outflow tract
- structural heart disease
- supraventricular tachycardia
- ventricular tachycardia
- wide complex tachycardia
- Received July 25, 2016.
- Revision received December 9, 2016.
- Accepted January 12, 2017.
- 2017 American College of Cardiology Foundation
- Brugada P.,
- Brugada J.,
- Mont L.,
- Smeets J.,
- Andries E.W.
- Sandler I.A.,
- Marriott H.J.
- ↵Lerman BB, Stein KM, Markowitz SM, Mittal S, Iwai S. Ventricular tachycardia in structurally normal hearts. In: Zipes DP, Jalife J, editors. Cardiac Electrophysiology from Cell to Bedside, 4th edition. WB Saunders: Philadelphia, PA: 2004:668–82.
- ↵Miller JM, Das MK, Arora R, Alberte-Lista C, Wu J. Differential diagnosis of wide QRS complex tachycardia. In: Zipes DP, Jalife J, editors. Cardiac Electrophysiology from Cell to Bedside. 4th edition. WB Saunders: Philadelphia, PA: 2004:747–57.