Between March and June 2018, we prospectively enrolled consecutive patients with AVB or SSS who were scheduled for pacemaker implantation. The inclusion criteria were: (1) no history of pacemaker implantation, (2) no pregnancy, (3) at least 18 years of age, (4) New York Heart Association (NYHA) classification I or II. Patients were excluded for the following conditions: (1) severe valvular regurgitation, (2) recent acute myocardial infarction, (3) a history of cardiac surgery or atrioventricular node ablation, (4) poor acoustic window condition, (5) confirmed infra-His bundle block, or (6) the presence of severe chronic diseases. The study conformed with the 1975 Declaration of Helsinki, and the protocol was approved by the Zhongshan Hospital Ethics Committee. All patients provided their written informed consent to participate in the study.

The pacing procedures were performed in a cardiac catheterization laboratory. Twelve-lead electrocardiogram (ECG) and intracardiac electrograms were simultaneously displayed and continuously recorded during all pacing interventions on a multichannel Bard Electrophysiology Lab System recorder (Bard, Haverhill, MA, USA). A catheter with a 6-Fr quadripolar electrode was inserted via the right external jugular vein; the electrodes were positioned within the right ventricular (RV) apex (RVA) and RV outflow tract (RVOT) (Figures 1A,B). For HBP, a preformed sheath (C315 HIS, Medtronic, Minneapolis, MN USA) was inserted via the right external jugular vein and placed in the region near the tricuspid valve septal leaflet. A Select Secure pacing lead (Model 3830, 69 cm, Medtronic) was delivered along the sheath with its distal part beyond the tip of the sheath for HBP recording (Figure 1C). For LBBP, the lead was twisted deeply through the ventricular septum from the RV septum to the endocardium of the LV septum to activate the LBB region (Figure 1D). According to the intracardiac electrograms, the LBB potential gradually appeared and increased as the electrode was screwed in, and the QRS morphology was gradually transformed from LBBB to RBBB. The interval between the LBB potential and ventricular activation was shorter than between the His bundle potential and ventricular activation. The imaging characteristics of LBBP showed that the pacing site was in the ventricular septum. The left or right anterior oblique projection was used to assist in identifying catheter positions; endocardial ECG was utilized to confirm these positions.

Each pacing mode was separated by a 10-min washing-out interval. In all patients, the pacing sequence ended with LBBP, and the lead was left in place after LBBP. During each procedure, the atrial lead was implanted in the right atrium appendage. Both the atrial and ventricular leads at the four pacing sites were connected to the programmer (Medtronic 2290) in DDD mode, with an AV delay of 150 ms and a pacing output of 3.5 V/0.5 ms during unipolar configuration.

ECGs and echocardiography were performed at baseline and during each pacing modality. During each session, patients were kept in the left lateral decubitus position with the ECG connected. Two-dimensional echocardiography was performed, according to current guidelines, using a Vivid E95 scanner (GE Vingmed Ultrasound, Horten, Norway) equipped with an M5S probe (4.0-MHz transducer) having frame rates higher than 40 fps (10). LV end-diastolic volume (LVEDV), LV end-systolic volume (LVESV), and tricuspid annulus plane systolic excursion (TAPSE) were derived from M-mode images. The LV ejection fraction (LVEF) was measured using the biplane Simpson's method, per guideline recommendations (11). To evaluate interventricular dyssynchrony, we measured the interventricular mechanical delay (IVMD) as the time interval between the beginning of QRS and the beginning of the systolic waves of aortic and pulmonary ejections, using conventional Doppler (12). Intraventricular dyssynchrony was assessed using the Yu index, defined as the standard deviation of the time between the onset of QRS and the peak systolic velocity of tissue Doppler for 12 LV segments (six basal and six middle) in apical triplane-mode (4-V probe) (13, 14).

The apical triplane-mode data were analyzed offline using an EchoPAC 203 workstation (GE Vingmed Ultrasound). The best cardiac cycle with good quality or clear endocardial boundaries was chosen, and the endocardial borders were automatically identified and tracked throughout the cardiac cycle. If the images were not optimal, manual adjustments were made. The LV wall of each apical view was divided into six segments. The global longitudinal peak strain (GLPS) (Figures 2A,C,E,G) and the standard deviation of the time to longitudinal peak strain of 17 segments (PSD) were automatically calculated. The longitudinal strain of the ventricular endocardium, mid-myocardium, and epicardium (Figures 2B,D,F,H) and the time to longitudinal peak strain (Ts) of the basal, middle, and apical segments of the lateral ventricular wall were simultaneously obtained. All echocardiograms were analyzed by an independent echocardiologist, blinded to the pacing modalities.

Continuous variables are described as means ± standard deviations; categorical variables are described as counts or percentages. When the data were or approximated normal distributions, comparisons among three or more conditions were evaluated using repeated measures one-way analysis of variance tests followed by the Tukey post-hoc analysis. Otherwise, the Friedman test was performed, and Dunn's post-hoc test was used to adjust the P-value. Statistical significance was defined as a two-sided P < 0.05. All statistical analyses were performed using GraphPad Prism 7.0 (GraphPad Software, San Diego, CA, USA).

A total of 21 patients (15 men and 6 women) were enrolled in the study, with a mean age of 66.1 ± 13.0 years. All procedures were successfully performed in these patients. Of these, 19 patients were diagnosed with AVB, including one with first-degree AVB, 10 with second-degree AVB, and eight with third-degree AVB, and two with SSS. The mean heart rate was 53.8 ± 16.2 beats/min and the mean QRS duration was 118.8 ± 24.6 ms at the baseline ECG (Table 1).

To compare the acute changes in cardiac contraction between the different pacing sites, we measured the LVEDV, LVESV, LVEF, and TAPSE. We also evaluated the GLPS and the longitudinal layer-specific myocardial strains of the LV and RV [endocardium (endo), mid-myocardium (mid), and epicardium (epi): LVendo, LVmid, LVepi, RVendo, RVmid, and RVepi] in the apical four-chamber view using EchoPAC 203 (Table 2). LVEDV, LVESV, LVEF, GLPS, and LV strains were used to evaluate left ventricle systolic function, whereas TAPSE and RV strains for right ventricle systolic function.

We analyzed heart rate (HR) and QRS duration using ECG; IVMD, PSD, and the Yu index were analyzed by echocardiography. Moreover, we also measured the Ts of the apical, middle, and basal segments of the lateral wall of left ventricle (Lat_ap, Lat_mid, Lat_bas, respectively) or right ventricle (RV_ap, RV_mid, RV_bas, respectively) in the four-chamber view, as well as the difference in Ts between the basal, middle, and apical segments of left and right ventricle lateral walls (LV-RV_bas, LV-RV_mid, LV-RV_ap, respectively; Table 3).

Many studies have investigated the feasibility, safety, and clinical outcomes of HBP. Sharma et al. (8) attempted HBP in 94 patients, and succeeded in 75(80%). They found that the HBP group required longer implantation times and a higher pacing threshold than the RVP group (98 patients). Heart failure hospitalization was significantly reduced in patients with >40% ventricular pacing in the HBP group than in the RVP group. For patients with no response to CRT or failure of LV electrode implantation, HBP corrected basal conduction disturbances and improved echocardiographic measurements as an alternative treatment for CRT (19). HBP was also employed to control atrial fibrillation in patients with heart failure who underwent atrioventricular node ablation, significantly improving their LVEF and NYHA classification (20). Our results provide complementary information with previous findings that showed that LVEF and TAPSE deteriorated during RVAP and RVOP, but had little influence on HBP and LBBP; LBBP evidently improved LVEDV. We measured GLPS and longitudinal layer-specific myocardial strains to accurately evaluate the regional mechanical motion of the ventricular myocardium. Although the longitudinal layer-specific strains of LV and RV and GLPS were significantly decreased at all pacing sites as the HR was corrected to within the normal range, LBBP showed the least impact. These results indicate that HBP and LBBP best maintained cardiac contraction. However, HBP has several limitations (21), including the requirement for skilled operation due to the difficulty in locating the His bundle and having a high pacing threshold and low sense. It is also not applicable to blocks below His bundle or to diffuse ventricular blocks caused by myocardial disease. Moreover, HBP cannot provide protection when cardiac conduction system lesions deteriorate. Hence, the investigation of new LV pacing (LVP) sites is required. In 2003, Peschar et al. (22) first conducted LVP in anesthetized, open-chest dogs with normal ventricular conduction. The immediate results demonstrated better maintained LV pump functioning associated with the LVP sites than with RVP sites. Mills et al. (23) carried out LVP in dogs after atrioventricular nodal ablation and further verified that LVP was superior to RVP in chronically maintaining LV contractile coordination and pump function. LV septal pacing was first clinically applied in 2016 and showed better hemodynamic effects than RVP (24). In 2017, Huang et al. (4) successfully implemented the first LBBP in a heart failure patient with LBBB, and the cardiac function of the patient improved during 1 year of follow-up. Previous studies have reported that RBBB can be corrected during LBBP. Li et al. observed a narrowing of the complete RBBB morphology using unipolar LBBP at a high output (25). Sometimes bipolar pacing (26) or adjusting atrioventricular delay (4) can also correct incomplete RBBB. In this study, LBBP did not induce RBBB and did not influence cardiac hemodynamics. However, further studies with long-term follow-up are necessary.

Cardiac synchrony is essential for cardiac structure and function, and cardiac resynchronization can reverse LV remodeling and reduce the risk of heart failure events (27, 28). Pastore et al. (29) performed permanent HBP in 37 patients with normal cardiac function and added an RVA backup lead in each patient. Compared with HBP, RVAP resulted in a wider QRS duration, significantly longer LV isovolumetric contraction and relaxation times, and higher pulmonary arterial systolic pressure. In this study, we evaluated electrical and mechanical synchrony at four pacing sites. The QRS duration is the main index used to evaluate electrical synchrony; its normal value is <120 ms. RVAP and RVOP significantly prolonged the QRS duration, which did not change and remained within the normal range during HBP and LBBP. Therefore, RVAP and RVOP caused electrical dyssynchrony; HBP and LBBP preserved physiological electrical synchrony.

Currently, multiple methods are utilized to evaluate mechanical synchrony. For example, Zhang et al. (30) adopted gated, single-photon emission computed tomography (SPECT) myocardial perfusion imaging to study the LV mechanical synchrony associated with HBP and found that HBP resulted in better LV mechanical synchrony parameters. However, SPECT is a procedure that involves radiation and cannot be performed in the catheterization laboratory or at the bedside. Echocardiography is accurate and convenient for measuring cardiac contraction and hemodynamics in real-time without radiation exposure. So far, little is known about the acute outcomes on cardiac mechanical synchrony during LBBP, and comparisons among four different pacing sites are rare. In this study, the IVMD, PSD, and Yu index were used to evaluate inter- and intraventricular synchrony. We found that RVAP distinctly extended IVMD and PSD, and the Yu index tended to deteriorate during RVAP, suggesting that RVAP causes inter- and intraventricular dyssynchrony. LBBP significantly improved the IVMD and Yu index compared with RVAP. Among the four pacing modalities, RVAP resulted in the most unfavorable acute impact on mechanical synchrony, whereas HBP and LBBP had little influence on mechanical synchrony. LBBP represented the best physiological pacing mode and maintained ventricular synchrony.

This study was limited by its small sample size. Further, we only measured the immediate changes in myocardial mechanics after implantation of pacemaker electrodes. The long-term effects of pacing at each site require further investigation.