Assessment of ventricular mechanical synchronization after left bundle branch pacing using 2‐D speckle tracking echocardiography

Abstract Background The left bundle branch pacing (LBBP) makes the ventricular depolarization closer to the physiological state and shortens QRS duration. The purpose of this study is to explore the ventricular systolic mechanical synchronization after LBBP in comparison with traditional right ventricular pacing (RVP) using two‐dimensional strain echocardiography (2D‐STE). Methods Thirty‐two patients who received LBBP (n = 16) or RVP (n = 16) from October 2018 to October 2019 and met the inclusion criteria were included in this retrospective study. Electrocardiogram (ECG) characteristics, pacing parameters, pacing sites, and safety events were assessed before and after implantation. Acquisition and analysis of ventricular systolic synchronization were implemented using 2D‐STE. Results In RVP group, ECG showed left bundle branch block patterns. At LBBP, QRS morphology was in the form of right bundle branch block, and QRS durations were significantly shorter than that of the RVP QRS (109.38 ± 12.89 vs 149.38 \± 19.40 ms, P < .001). Both the maximum time differences (TD) and SDs of the 18‐segments systolic time to peak systolic strain were significantly shorter under LBBP than under RVP (TD, 66.62 ± 37.2 vs 148.62 ± 43.67 ms, P < .01; SD, 21.80 ± 12.13 vs 52.70 ± 17.72 ms, P < .01), indicating that LBBP could provide better left ventricular mechanical synchronization. Left and right ventricular pre‐ejection period difference was significantly longer in RVP group than in LBBP group (10.23 ± 3.07 vs 39.94 ± 14.81 ms, P < .05), indicating left and right ventricular contraction synchronization in LBBP group being better than in RVP group. Conclusion LBBP is able to provide a physiologic ventricular activation pattern, which results in ventricular mechanical contraction synchronization.


| INTRODUCTION
The traditional pacing site at the right ventricular apex or the right side of ventricular septum was demonstrated to have detrimental impact on clinical outcomes due to ventricular mechanical dyssynchrony secondary to electrical dyssynchrony. 1,2 Studies have demonstrated the feasibility and clinical benefits of permanent His-bundle pacing (HBP). 3 Former researchers have found that permanent HBP led to significant narrowing of QRS duration and improvement in the left ventricle (LV) function in patients with reduced LV ejection fraction (LVEF). 4 But His bundle pacing still have several disadvantages, including difficulty of lead implantation, lower R-wave amplitudes, high and unstable pacing threshold, especially in patients who have conduction block distal to the His bundle. 3,5 Upadhyay et al demonstrated that the site of block usually was located within the His or proximal left bundle. 6 Thus, alternative pacing sites have been sought. After penetrating through the membranous atrioventricular septum, the conductive fibers of the left bundle branches spread beneath the endocardium of ventricular septum in a relatively large dimension, 7 which offers an opportunity for pacing the left bundle branch (LBB) in an easier manner. Weijian Huang et al developed a technique for the left bundle branch pacing (LBBP) using a transseptal approach. 8 LBBP has been reported to offer low pacing thresholds and large R waves, and has a lower theoretical risk for the development of distal conduction block due to that the distal conduction system is targeted. 9 Keping Chen et al reported that LBBP had a lower pacing threshold and produced narrower electrocardiogram (ECG) QRS duration compared with the right ventricular pacing (RVP). 10 The purpose of this study is to explore the electrical and mechanical synchrony of the ventricle using two-dimensional strain echocardiography (2D-STE) after LBBP in comparison with the traditional RVP.

| Implantation procedures
Before the pacemaker implantation, the ventricular septal thickness was assessed by Echo. Twelve-lead ECG and intracardiac electrograms were simultaneously displayed and recorded on a multichannel recorder.
LBBP: The delivery sheath (C315 HIS, Medtronic Inc., Minneapolis, MN) was inserted via the left subclavian vein or axillary vein and placed on the right side of the septum inferior to the septal leaflet of the tricuspid valves about 1 to 1.5 cm from HBP site toward right ventricular apex (RAO 30 ). Then, the Select Secure pacing lead (Model 3830 69 cm, Medtronic Inc., Minneapolis, MN) was advanced through the sheath with its tip just beyond the distal part of the sheath for unipolar pacing and local activation potential recording. The sheath and the pacing lead touched the septum and pacing with an output of 5.0 V/0.4 ms was applied, which created ECG QRS morphology of "W" pattern with the notch closer to nadir in lead V1. The pacing lead was then screwed perpendicularly to LV septum (LAO 30 -45 , Figure 1C). During the lead advancement procedure, the changes in the notch in V1 lead, the fulcrum sign, and impedance changes were observed, and sometimes the sheath angiography was used to determine LBBP lead depth into ventricular septum ( Figure 1D). Once ECG QRS morphology during pacing resulted in a pattern of the right bundle branch block (RBBB) or a very narrow QRS complex (<120 ms), the lead had been at or near the left bundle branch and the lead advancement was stopped. Then the test with different output was used to confirm LBB capture. Evidences for direct LBB capture were as follows: (a) pacing morphology of RBBB pattern; (b) recording LBB potential; (c) stimulus-peak R V5 or R V6 shortening abruptly with increasing output or remaining shortest and constant (<90 ms) at low and high outputs; (d) selective LBBP and non-selective LBBP; or (e) recording retrograde His potential or anterograde LBB potential during pacing (not routine in clinical practice). 11 The pacing parameters were measured to confirm stable capture threshold and consistent pacing impedance, and then the sheath was removed.

| Programming
Conventional lead-related parameters including stimulation threshold, impedance, and sensing amplitude were recorded right after implantation and 7 days later. Ventricular pacing rates of all patients were more than 70%. The atrioventricular delay was selected, which corrected bundle branch block and yielded narrower QRS duration. Lead-related adverse events were not observed during follow-up.

| Echocardiographic parameters
Echocardiography was performed using VVIq ultrasound (GE Company, USA) with S-5 transducers 7 days after surgery.
Acquisition and analysis of synchronization using 2D-STE were performed by senior sonographers. Standard apex four-chamber view, apex three-chamber view, and apex two-chamber view were clearly exposed and recorded as well as ECG, and four consecutive cardiac cycles with constant heart rate were collected. The 18-segment systolic times to peak 2-D longitudinal systolic strain were recorded for every patient. Then the maximum time difference of systolic times to peak 2-D strain among the 18 LV segments (2D-TD max ), and the SDs of systolic times to peak systolic strain (PSS) of the 18 LV segments were calculated. The 6-segment time differences (TDs) to PSS of the apex fourchamber view, apex three-chamber view, and apex two-chamber view were also calculated, respectively. Both the LV function and synchronization status using 2D-STE were assessed using the off-  Figure S2).

| Statistical analysis
Data were presented as mean ± SD for continuous variables.
Quantitative data with normal distribution compared between two groups were evaluated using the Student t test. Quantitative data inconsistent with normal distribution compared between two groups were evaluated using the Wilcoxon rank sum test. Categorical data compared between two groups were evaluated using Fisher exact probabilities method. All P values were two-tailed, and P values <.05 were considered to indicate statistical significance. All statistical analyses were performed using IBM SPSS Statistics 26.
3 | RESULTS There was no significant difference in sensing amplitude, pacing impedance, and capture threshold among LBBP and RVP (Table 1).

| Baseline characteristics
Pacing thresholds and sensing amplitude measured at time of hospital discharge remained stable compared with that measured at implantation. In

| LV synchronization status
The LV wall was divided into 18 segments, each of which showed Similar results were obtained when comparing SD of 18-segment systolic times to peak 2-D strain between the two groups at baseline and after operation (before pacing 17.57 ± 10.04 vs 13.30 ± 8.14 ms, P = .187; after pacing 21.80 ± 12.13 vs 52.70 ± 17.72 ms, P < .01).
There was no statistical difference for SD in the LBBP group when comparing with baseline (21.80 ± 12.13 vs 17.57 ± 10.04 ms, P > .05).
The 6-segment maximum TDs to peak 2-D longitudinal strain of the apex four-chamber view, apex three-chamber view, and apex two-chamber view were longer in RVP group than those in the LBBP group (Table 2, Figure 3). The TD between the anteroseptal wall and posterior wall was statistically longer in the RVP group than in the LBBP group after pacing (22.94 ± 29.2 vs 96.63 ± 41.24 ms, P < .01).
Comparing the TD between the anterior wall and inferior wall and the  Figure S1). IVMD was significantly longer in the RVP group than in LBBP group (10.23 ± 3.07 vs 39.94 ± 14.81 ms, P < .05), and the result indicated that interventricular contractile synchronization was better in the LBBP group than in the RVP group (Supplementary Figure S2).

| Limitations in LBBP therapy
During our study, we found that patients with intraventricular block including left anterior and posterior fascicular block cannot achieve LV contraction synchronization after pacing the left bundle branch ( Figure 5).

| DISCUSSION
Conventional pacing therapy places the pacemaker lead in the right ventricular apex or septum, with ventricular pacing forming a wide QRS wave and unsynchronized contraction. The His bundle-Purkinje conduction system ensures rapid activation in both ventricles and synchronized contraction. 12 Figure 5A). 2-D speckle-tracking echocardiography was used to get the time-systolic strain curve of the 18 segments after operation ( Figure 5B); 2D-TDmax of the 18-segment systolic time to peak systolic strain was 152 ms ( Figure 5B) It is precisely because LBBP can bring about synchronization of ventricular contraction that more and more studies report that this technology is currently being used in the treatment of heart failure combined with left bundle branch block. 14,15 LBBP was also used in patients of atrioventricular node ablation with persistent atrial fibrillation and implantable cardioverter-defibrillator therapy in order to improve LV function. 16 In this study, we encountered an interesting case with atrial fibrillation, III AVB and symptoms of heart failure including dyspnea and decreased activity tolerance. After being upgraded to LBBP using a 3830 pacing lead, the patient's symptoms of heart failure alleviated, and the NYHA heart function class improved from class III to class II about 1 month after operation (Supplementary Figure S3). For those right ventricular apex pacing patients with heart failure, upgrading ventricular lead to LBBP may improve LV systolic mechanical synchronization and alleviate heart failure.
In our study, LBBP is clinically feasible in patients, and pacing parameters were normal during pacemaker implantation and followup. Former researches also got similar results. 17 There were also limitations in LBBP therapy. Because LBBP cannot correct the intraventricular block which was distal to the left bundle branch. Thus, if the widened QRS wave and left ventricular systolic dyssynchrony was caused by intraventricular block, it cannot be corrected by LBBP.

| CONCLUSIONS
LBBP was able to provide a physiologic ventricular activation pattern almost identical to the intrinsic conduction pattern, which resulted in ventricular mechanical contraction synchronization. Ventricular mechanical contraction synchronization after permanent LBBP helps to maintain good heart function and prevent detrimental impact of RVP on clinical outcomes due to ventricular mechanical dyssynchrony.

CONFLICT OF INTEREST
We declare that there is no conflict of interest in this work.

AUTHOR CONTRIBUTIONS
Beibing Di and Zhijun Sun contributed equally to this study.

DATA AVAILABILITY STATEMENT
The data used to support the findings of this study are available from Beibing Di (beibingyouxiang@163.com) upon reasonable request.