Evaluation of the Specificity of Morphological Electrocardiographic Criteria for the Differential Diagnosis of Wide QRS Complex Tachycardia in Patients With Intraventricular Conduction Defects
Background Although several ECG criteria have been described for the differential diagnosis of tachycardias with a wide QRS complex, their applicability in patients with preexisting intraventricular conduction defects (IVCDs) has been questioned. The specificity of previously described criteria in this context is unknown.
Methods and Results We analyzed prospectively the specificity of the QRS morphological criteria previously described in ECGs during sinus rhythm of 232 patients with IVCD. Only 5 of 12 analyzed criteria had a specificity ≥0.90 among our patients: (1) a triphasic configuration (Rsr′or Rr′) QRS complex in V1 in the presence of a right bundle-branch block morphology (BBBM); (2) a QS, QR, or R QRS pattern in V6 in the presence of a right BBBM; (3) any Q in V6 in the presence of a left BBBM; (4) a concordant pattern in all precordial leads; and (5) the absence of an RS complex in all precordial leads (particularly useful for left BBBM). The following criteria—QRS duration >140 ms; a left axis with right BBBM, right superior axis with right BBBM, monophasic or biphasic R wave in V1 with right BBBM, and a relation R/S <1 with right BBBM; an R >30 ms in lead V1 or V2 with left BBBM, >60 ms from QRS onset to S nadir with left BBBM, a notched downstroke S wave with left BBBM, and an R-to-S interval >100 ms in one precordial lead—had a specificity of 0.43, 0.54, 0.87, 0.80, 0.85, 0.78, 0.66, 0.69, and 0.63 (0.84 in right BBBM), respectively.
Conclusions Most of the previously described morphological criteria favoring ventricular tachycardia are present in a substantial percentage of patients with IVCD during sinus rhythm. These findings suggest a limited applicability of these criteria in this subset of patients.
Several ECG criteria have been described for the differentiation between (VT) and SVT with a wide QRS complex.1 2 3 Some of these criteria involve an analysis of the QRS morphology and were developed by comparing tracings of VT to a “control” group of SVT with BBB. In fact, in a recent study involving a step-by-step ECG analysis, the first two criteria were morphological.3 One of the limitations of those studies is that the control group involved mostly patients in whom BBB was functional; patients with organic BBB were either excluded1 2 or unlikely, because the population was selected on the basis of an indication for an electrophysiological study and such patients infrequently have organic BBB. The only study analyzing the proposed morphological criteria in patients with organic BBB involved a small number of patients and was published before the two new morphological criteria were described.4 Thus, the available information about the applicability of the described morphological criteria to patients with organic BBB is limited.
To study the specificity of the previously published morphological criteria in patients with organic BBB, we have prospectively analyzed the tracings of 232 unselected consecutive patients admitted to our institution who had IVCDs during sinus rhythm.
The 12-lead ECGs performed at our institution from February 1992 to December 1994 were prospectively screened for the following inclusion criteria: (1) sinus rhythm, (2) a wide QRS ≥120-ms QRS complex, (3) a constant PR interval of a duration ≥140 ms, and (4) admission to the hospital at the time. Once an ECG meeting the inclusion criteria was identified, one of the investigators obtained clinical information about the patient by direct questioning and reviewing the medical records. The presence of structural heart disease was determined from the history and objective evidence on the basis of noninvasive and/or invasive diagnostic techniques, including two-dimensional echocardiography and cardiac catheterization when clinically indicated. Particular attention was given to the use of antiarrhythmic drugs (specially class IA, IC, and III).
Two hundred thirty-two consecutive patients met the inclusion criteria. Their clinical characteristics are described in Table 1⇓ (156 men and 76 women). Their ages ranged from 26 to 91 years (mean±SD, 68±11 years). Clinical and/or laboratory findings of structural heart disease were present in 151 patients (65%). An estimation of the left ventricular ejection fraction (by echocardiography and/or contrast ventriculography) was obtained in 130 patients (56%). The left ventricular ejection fraction was significantly depressed (<40%) in 60% of those patients.
Among the morphological ECG criteria previously described for the differential diagnosis of wide QRS complex tachycardia, we selected for analysis those with ≥0.90 specificity and electrophysiological documentation of the tachycardias studied in the original publication. Criteria were also included if only applicable to a subset of ECG (ie, those with left or right BBBM QRS complex or those with RS morphology in any of the precordial leads3 ). Because these criteria were meant to be used for the recognition of VT, right and left BBBM was defined, as it is usually, during VT4 5 6 7 8 : a QRS complex was considered to have a right BBBM if it had a dominant R wave in V1; if the dominant wave in V1 was an S wave, the QRS complex was considered to have a left BBBM. The criteria selected and analyzed as suggestive of VT were the following: (1) QRS duration >140 ms1 2 ; (2) left axis deviation (QRS axis in the frontal plane between −30° and −90°) in the presence of a right BBBM1 2 ; (3) right superior axis (QRS axis in the frontal plane between −90° and −180°) in the presence of a right BBBM4 6 ; (4) monophasic (R) or biphasic (qR, QR, Rs, or RS) or a rabbit-ear triphasic configuration (Rsr′ or Rr′) QRS complex in V1 in the presence of a right BBBM1 2 4 ; (5) rS, QS, QR, or R configuration in lead V6 in the presence of a right BBBM1 2 ; (6) R-wave duration >30 ms in leads V1 or V2 in the presence of a left BBBM7 ; (7) >60 ms from QRS onset to the nadir of the S wave in lead V1 or V2 in the presence of a left BBBM7 ; (8) notched downstroke S wave in lead V1 or V2 in the presence of a left BBBM7 ; (9) any q wave in lead V6 in the presence of a left BBBM7 ; (10) the presence of a concordant pattern (entirely positive or negative QRS complexes) in all precordial leads1 2 4 ; (11) absence of an RS patterns in all precordial leads3 ; and (12) an interval >100 ms from the onset of the R wave to the nadir of the S wave in precordial leads with an RS morphology.3 According to the method of Wellens et al,1 2 we considered “triphasic” the patterns rR′ and Rr′ (types 3 and 5 of Reference 22 ).
Only one tracing was analyzed per patient. If more than one was obtained, the tracing with the best technical quality was selected for analysis.
The mean±SD is reported. To calculate specificity of each criterion, we considered all of our patients to belong to the “nondiseased” population, because they did not have actual VT. However, all the criteria analyzed in the study have been described for the diagnosis of VT. Thus, cases that were positive for each criterion were false positives, and those cases that were negative were true negatives. The specificity for each criterion was calculated by dividing true negatives (or those negative for that criterion) by all nondiseased cases (the entire population or the particular subset to which each criterion can be applied). Categorical variables were compared with the χ2 test. Statistical significance was at the level of P<.05.
Specificity of the Proposed Criteria
Ninety-nine petients (43%) had a right BBBM, and 133 (57%) had a left BBBM.
Table 2⇓ reflects the specificity of each of the proposed criteria. It can be noted that only 5 of the 12 criteria analyzed had specificity >0.90 (see Figs 1⇓ and 2⇓). Thus, only 5 of the 12 criteria had a specificity in patients with IVCD similar to that in patients without IVCD. Three other criteria, all among those evaluated in patients with right BBBM, had >0.80 specificity (Figs 3⇓ and 4⇓). When criteria were combined, the specificity decreased, except if only the criteria with excellent specificity were included. For example, the specificity of the combined criteria with a specificity ≥0.80 in the presence of right BBBM (ie, how many ECGs had either a rabbit-ear triphasic configuration (Rsr′or Rr′) QRS complex in V1 or a QS, QR, or R configuration in lead V6; a right superior axis; a monophasic (R) or biphasic (qR, QR, RS) in lead V1, or an rS in lead V6) was 0.58 (57 of 99). However, if only the two criteria with a specificity ≥0.90 were considered, the combined specificity was high, 0.96 (95 of 99). The specificity of the four combined criteria in left BBBM was only 0.55 (59 of 132) (Figs 5⇓ and 6⇓). The consecutive specificity of the first and second criteria described by Brugada et al3 was also only 0.57 (Fig 7⇓).
The criteria that could be applied in both BBBM were also analyzed separately in patients with right BBBM and with left BBBM (Table 3⇓). In the presence of right BBBM, the specificities of the absence of an RS complex in all precordial leads and an R-to-S interval >100 ms in one precordial lead were 0.81 and 0.84, respectively (consecutive, 0.68) (Figs 2⇑ and 7⇓). In cases with left BBBM, the specificities were 0.98 and 0.50, respectively (consecutive, 0.50). Figs 5⇓ and 6⇑ illustrate examples of an R-to-S interval >100 ms in ECGs with left BBBM. There was a significant relationship between the presence of right BBBM and the absence of an RS complex in all precordial leads (P<.0001). There was also a significant relationship between the presence of left BBBM and an R-to-S interval >100 ms in one precordial lead (P<.0001) (Table 4⇓).
Finally, we analyzed the criteria in the patients ECGs who were taking class I or III antiarrythmic drugs (Table 5⇓). Seven patients had this characteristic: 4 were taking amiodarone, 1 was on propafenone, 1 was taking procainamide, and 1 was on quinidine. Six had a QRS duration >140 ms. The absence of an RS complex in all precordial leads was not a feature of any of the patients taking antiarrythmic drugs, but 6 had a R-to-S interval >100 ms in one precordial lead. Five patients had left BBBM, and the remaining 2 had right BBBM. Interestingly, 3 of the 4 patients with left BBBM in whom the downstroke of the S wave in V1 or V2 could be evaluated had a notched pattern.
The main observation of the present study is that the specificity of most morphological criteria for the differential diagnosis of wide QRS tachycardias decreases in the presence of IVCD. Only five criteria were found to have and specificity >0.90. It has been suggested that the value of the classic morphological criteria for the diagnosis of VT may decrease if preexistent BBB is present.2 4 However, because most of the studies on the differential diagnosis of wide QRS tachycardia were performed in patients with BBB developed during tachycardia in the electrophysiology laboratory (presumably functional in most of them), the information regarding the above suggestion has been limited until now.
In the presence of fixed (ie, nonintermittent) complete BBB and in the absence of accessory pathways, all beats of supraventricular (including sinus) origin preferentially conduct over the contralateral bundle. If all supraventricular impulses use the same fixed pathway for activating the ventricle, the QRS morphological pattern is expected to remain unaltered regardless of rate. Therefore, the QRS morphological configuration during SVT in patients with preexisting BBB would be identical to that observed during sinus rhythm.6 8 This contention supports our approach of an analysis of QRS morphology during sinus rhythm to assess the specificity of morphological criteria to be used during tachycardia. Had our patients developed an SVT, they would have been expected to display a similar QRS morphology.
We analyzed the 12 previously described morphological criteria of VT that had a specificity >0.90 in different series. However, in our ECG, only 5 of them had a specificity >0.90. The sensitivity of these criteria was quite low for most of them in different studies: 0.11,2 0.31,2 0.55,7 0.03,2 and 0.21,3 respectively. Thus, even these criteria with high specificities have limited clinical value in this patient population.
After having stated the former, it is also fair to recognize that if a specificity ≥0.80 is considered acceptable for a medical test, other more sensitive criteria, such as a right superior axis in the presence of right BBBM, a monophasic R or biphasic (qR, QR, or RS) in V1, and a relation R/S <1 in V6 in the presence of right BBBM, could be considered “reasonably” useful in these patients.
However, some criteria that have been noted to have limited value in relatively recent studies (not necessarily including patients with organic BBB) were also found to have a low specificity in the present series. A QRS duration >140 ms, which has been found to be of limited value,3 6 9 had a low specificity (0.43) in our study. A left axis deviation in the presence of right BBBM was not found to be useful by Wellens et al,2 Akhtar et al,6 Dongas et al,8 and Brugada et al3 ; it had a low specificity in the present series.
In right BBBM, the specificity of combined criteria with a specificity ≥0.80 is also low (0.58) and cannot be used to accurately diagnose the origin of wide complex tachycardia with this morphology. However, only when the two criteria with >0.90 specificity are combined does the specificity remains high.
The only series that specifically addressed the value of morphological criteria in patients with organic BBB was that of Kremers et al.4 This study compared 106 ECGs of VTs in 70 patients with a sinus rhythm ECG; of this group, 18 patients had preexisting BBB. They showed a high specificity for three criteria: a monophasic R wave in V1 in presence of right BBBM (specificity=1.0), an R wave >30 ms in V1 or V2 with left BBBM (specificity=1.0), and a right superior axis (specificity=1.0).
The presence of a monophasic R wave in V1 is included in our study as part of a single criterion, along with a biphasic pattern in the same lead, and had an 0.80 specificity. It was analyzed in exactly the same way as its original description. However, if only the presence of a monophasic R wave was considered, the specificity would have increased to 0.95, in accordance with the finding of Kremers et al4 of a high specificity for this criterion. In contrast, the finding of an R wave >30 ms in patients with a left BBBM had a lower specificity in our series. However, in the study by Kremers et al,4 there were only 8 patients with left BBBM compared with 133 in our series. The QRS axis, as such, was considered in our analysis only for ECGs with a right BBBM, because its sensitivity is unknown in ECGs with a left BBBM in the absence of organic BBB during sinus rhythm. For patients with right BBBM, its specificity was close to 0.90 (0.87).
A different and simpler approach to address the same problem was undertaken by Dongas et al.8 They demonstrated that when the sinus QRS morphology of patients with preexisting BBB is compared with the wide QRS complex tachycardia morphology of the same patient, any change in QRS morphology is indicative of a ventricular origin. However, this criterion is not applicable when a patient is first seen during tachycardia, justifying our efforts to recognize the specificity of different patterns.
While the specificity of the combined criteria described by Kindwall et al7 was 0.89, in our series, it is only 0.55. However, only 15 of ECGs (13%) in the series of Kindwall et al had left BBB in sinus rhythm.
In 1991, two new criteria were included as part of a step-by-step diagnostic algorithm for the ECG differential diagnosis of a wide QRS complex tachycardia. In this study, the specificity of the first criterion was 1.00, and when the second criterion was applied, the specificity was 0.98. However, the control group of that study (inducible SVTs) probably included many patients in whom BBB was functional. Our tracings showed a high specificity (0.91) for VT if an absence of RS in all precordial leads was present. However, the low specificity (0.63) for the second criterion (R-to-S interval >100 ms) and the low consecutive specificity (0.57) could potentially limit the diagnostic efficiency of the proposed algorithm in a patient with IVCD should he or she develop SVT. When the criteria that could be analyzed in both right and left BBBM were analyzed separately in these two groups of ECGs, some differences in specificity were noted. An absence of RS complex in all precordial leads could be particularly useful for left BBBM (seen in only 2 of 133 ECGs), whereas the converse could be true for an RS >100 ms in any precordial lead (0.84 specificity in right BBBM). Likewise, the QRS duration was less specific in cases with left BBBM. This is consistent with the well-known association between left BBB and structural heart disease, with its propensity to alter intraventricular conduction.
Analysis of ECGs in patients taking antiarrhythmic drugs showed that the QRS duration was >140 ms in most patients, as could be expected from the pharmacological effect. The second criterion described by Brugada et al3 tended to be present in most of these patients. This observation was described by Le Davay et al.10 They reported the case of a patient treated with flecainide for atrial fibrillation who was hospitalized because of a wide QRS tachycardia with an R-to-S interval >100 ms and with the diagnosis of SVT demonstrated by subsequent endocavitary electrophysiological studies.
In conclusion, our results highlight the limitations in the specificity of some of the morphological criteria for the differential diagnosis of wide QRS tachycardia as derived from the analysis of tracing with IVCD in sinus rhythm. In contrast, five previously described criteria have >0.90 specificity in this setting. In view of the relatively high probability of some of them having a BBB in sinus rhythm (18 of 70 in the Kremers et al series4 ), this information may be of particular clinical importance in patients who are first seen in wide QRS tachycardia and for whom information about sinus rhythm ECG is lacking.
Selected Abbreviations and Acronyms
|BBB||=||bundle branch block|
|IVCD||=||intraventricular conduction defect|
Reprint requests to Dr Jesús Almendral, Cardiología (planta 5), Hospital General Gregorio Marañon. Calle Doctor Esquerdo 46, 28007 Madrid, Spain.
- Received October 21, 1996.
- Revision received June 18, 1997.
- Accepted June 26, 1997.
- Copyright © 1997 by American Heart Association
Wellens HJJ, Bar FW, Vanagt EJ, Brugada P, Farré J. The differentiation between ventricular tachycardia and supraventricular tachycardia with aberrant conduction: the value of the 12-lead electrocardiogram. In: Wellens HJJ, Kulbertus HE, eds. What’s New in Electrocardiography? Hague, Netherlands: Martinus Nijhoff Publishing; 1981:184-199.
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. 1991;83:1649-1659.
Miller JM, Marchlinski FE, Buxton AE, Josephson ME. Relationship between the 12-lead electrocardiogram during ventricular tachycardia and endocardial site of origin in patients with coronary artery disease. Circulation. 1988;77:759-766.
Akhtar M, Shenasa M, Jazayeri M, Caceres J, Tchou P. Wide QRS complex tachycardia: reappraisal of a common clinical problem. Ann Intern Med. 1988;109:905-912.
Le Davay M, Victor J, Dupuis JM, Merheb M, Tran P, Tadeï A. Tachycardies à QRS larges: Limites des critères électrocardiographiques chez les patients sous antiarythmiques. Ann Cardiol Angéiol. 1994;43:253-255.