A Revised Definition of Left Bundle Branch Block Using Time to Notch in Lead I

Abstract

Importance: Current left bundle branch block (LBBB) criteria are based on animal experiments or mathematical models of cardiac tissue conduction and may misclassify patients. Improved criteria would impact referral decisions and device type for cardiac resynchronization therapy.

Objective: To develop a simple new criterion for LBBB based on electrophysiological studies of human patients, and then to validate this criterion in an independent population.

Design, setting, and participants: In this diagnostic study, the derivation cohort was from a single-center, prospective study of patients undergoing electrophysiological study from March 2016 through November 2019. The validation cohort was assembled by retrospectively reviewing medical records for patients from the same center who underwent transcatheter aortic valve replacement (TAVR) from October 2015 through May 2022.

Exposures: Patients were classified as having LBBB or intraventricular conduction delay (IVCD) as assessed by intracardiac recording.

Main outcomes and measures: Sensitivity and specificity of the electrocardiography (ECG) criteria assessed in patients with LBBB or IVCD.

Results: A total of 75 patients (median [IQR] age, 63 [53-70.5] years; 21 [28.0%] female) with baseline LBBB on 12-lead ECG underwent intracardiac recording of the left ventricular septum: 48 demonstrated complete conduction block (CCB) and 27 demonstrated intact Purkinje activation (IPA). Analysis of surface ECGs revealed that late notches in the QRS complexes of lateral leads were associated with CCB (40 of 48 patients [83.3%] with CCB vs 13 of 27 patients [48.1%] with IPA had a notch or slur in lead I; P = .003). Receiver operating characteristic curves for all septal and lateral leads were constructed, and lead I displayed the best performance with a time to notch longer than 75 milliseconds. Used in conjunction with the criteria for LBBB from the American College of Cardiology/American Heart Association/Heart Rhythm Society, this criterion had a sensitivity of 71% (95% CI, 56%-83%) and specificity of 74% (95% CI, 54%-89%) in the derivation population, contrasting with a sensitivity of 96% (95% CI, 86%-99%) and specificity of 33% (95% CI, 17%-54%) for the Strauss criteria. In an independent validation cohort of 46 patients (median [IQR] age, 78.5 [70-84] years; 21 [45.7%] female) undergoing TAVR with interval development of new LBBB, the time-to-notch criterion demonstrated a sensitivity of 87% (95% CI, 74%-95%). In the subset of 10 patients with preprocedural IVCD, the criterion correctly distinguished IVCD from LBBB in all cases. Application of the Strauss criteria performed similarly in the validation cohort.

Conclusions and relevance: The findings suggest that time to notch longer than 75 milliseconds in lead I is a simple ECG criterion that, when used in conjunction with standard LBBB criteria, may improve specificity for identifying patients with LBBB from conduction block. This may help inform patient selection for cardiac resynchronization or conduction system pacing.

Figure 1.. Mapping Left Ventricular (LV) Septal Activation Patterns

Surface electrocardiography (ECG) recordings (top) and intracardiac mapping of LV septal activation (bottom) from patients who meet standard American College of Cardiology/American Heart Association/Heart Rhythm Society surface ECG criteria for left bundle branch block. A, Septal mapping in this patient shows His-Purkinje activation (highlighted in blue) with a site of interrupted conduction (red lines) as well as nonphysiologic activation of the septal myocardium from basal to apical. These findings are consistent with complete conduction block. B, This patient demonstrates intact basal to apical His-Purkinje conduction down the length of the LV septum (highlighted in blue) as well as physiologic apical to basal activation of the septal myocardium. These findings indicate intact Purkinje activation. Standard 12-lead ECGs for these 2 patients are shown in eFigure 1 in Supplement 1.

Figure 2.. Measurement of Notch Times

Time to notch is measured as the time from QRS onset to the nadir of the notch or midpoint of a slur. If multiple notches or slurs are present in the QRS complex, the latest one is used. The measurement shown demonstrates a time to notch of approximately 90 milliseconds in lead I.

Figure 3.. Periprocedural Left Bundle Branch Block During Transcatheter Aortic Valve Replacement Implant

Example of preprocedural (A) and postprocedural (B) electrocardiograms (ECGs) for a patient who received a transcatheter aortic valve replacement implant. The postprocedural ECG shows a new marked widening of the QRS complex with new lateral QRS notching compared with the preprocedural ECG. This suggests that the patient developed a new left bundle branch block due to the transcatheter aortic valve replacement implant. This is a known complication of the procedure and is due to complete conduction block rather than intraventricular conduction delay.

Figure 4.. Analyses of Lead I Notch Times in the Derivation and Validation Cohorts

A, Box plots for lead I notch times in patients from the derivation group who were found to have complete conduction block (CCB) on electrophysiological study vs those with intact Purkinje activation (IPA). Notch times from patients in the transcatheter aortic valve replacement (TAVR) validation cohort with new left bundle branch block are also shown. The lower and upper borders of the boxes indicate the 25th and 75th percentiles, respectively; center horizontal line, median notch time; and whiskers, range. For this plot, patients without an identifiable notch or slur in lead I were assigned a notch time of 0 milliseconds. B, Percentage of patients who had an identifiable notch or slur in lead I on surface electrocardiography in the derivation CCB group (40 of 48 patients [83.3%]) compared with the derivation IPA group (13 of 27 patients [48.1%]) and the TAVR validation cohort (42 of 46 patients [91.3%]). The difference between the CCB and IPA groups was statistically significant. Derivation and validation cohorts were not statistically compared with each other as these represent 2 distinct populations.