Heart failure (HF) is a major public health issue with high morbidity and mortality, resulting in considerable financial and service burdens to the health system [1]. Left bundle branch block (LBBB) causes dyssynchronous electrical activation of the heart and creates discoordinate contraction of the left ventricle (LV), which leads to or aggravates HF [2]. Cardiac resynchronization therapy (CRT) using biventricular pacing (BIVP) has been recommended to improve cardiac functionality and enhance prognosis of patients with advanced HF when the optimal drug treatment still fails to improve the symptoms of HF [3]. However, the procedure for implanting the LV pacing lead of BIVP is quite complex, particularly in patients with venous malformations or coronary vein stenoses [4‐6]. Furthermore, approximately 30% of patients have nonresponse to BIVP [7, 8]. His-Purkinje system pacing is currently considered the optimal physiologic pacing method with the pacing lead directly implanted in the conduction system to narrow QRS wave and improve cardiac function by selective or nonselective His-bundle pacing (HBP) [9, 10]. Nevertheless, HBP has several shortcomings limiting its application, such as a relatively lower success rate, high corrective threshold and low R-wave amplitude due to the specific anatomic characteristics of the His bundle [11]. In addition, HBP implantation may easily injure the bundle branch and exacerbate occurrence of atrioventricular (AV) block [12‐14].

Huang et al. [15] first presented left bundle branch pacing (LBBP) in 2017, which targets pacing the proximal left bundle and its branches along with capture of LV septal myocardium. Selective LBBP (S-LBBP) only captures the LBB without myocardial capture, while nonselective LBBP (NS-LBBP) captures both the LBB and the local myocardium [16]. It is called LV septal endocardium pacing (LVSP) or deep septal pacing if only LV septal myocardium is captured [16]. Left bundle branch area pacing (LBBAP), with the lead implanted slightly distal to the His bundle and screwed deep in the LV septum ideally to capture LBB according to the ESC guidelines in 2021 [17], means LBBP or LVSP, without clear evidence for LBB capture [18]. Accumulating studies have shown that LBBAP can correct complete left bundle branch block (CLBBB), restore LV synchrony in HF patients, and improve cardiac function as well as symptoms in these patients, but the average period of follow-up for these studies was relatively short ranging from 6 to 12 months [15, 19, 20]. Therefore, we aimed to prospectively assess the long-term effects and safety in patients with HF and CLBBB after LBBAP and BIVP.

In this study, we analyzed the clinical characteristics of patients with HF and CLBBB after LBBAP and BIVP, respectively, and further performed a detailed comparison on the effect of the approaches on ECG and cardiac function between them. Our major findings were as follows: (1) LBBAP is a feasible and safe approach for successful correction of CLBBB in patients with HF and CLBBB; (2) The long-term follow-up revealed that LBBAP significantly improved LVEF and NYHA functional class and further lowered BNP level and LVEDD; (3) LBBAP significantly shortened QRS duration and exerted better cardiac electrical resynchronization to relieve symptoms of HF, compared with BIVP; and (4) In the LBBAP group, patients with moderately prolonged LVAT and QRS with a notch in the limb leads in preoperative ECG, hence may benefit more from CLBBB full correction.

Multiple studies have shown that LBBAP exerts fewer perioperative complications with no fatal adverse effects [15, 25‐28]. Results from our study revealed that neither procedural- nor device-related complications occurred in both groups, while LBBAP significantly reduced incidence of rehospitalization compared with BIVP. Our findings together with previous reports [25, 29] have indicated that LBBAP implanting is a relatively safe and effective procedure. Additionally, the capture thresholds of LBBAP were significantly lower than BIVP at implant, which was also significantly lower than that of HBP [25]. In fact, the LBBAP procedure had numerous advantages over BIVP, including shorter operation time and fluoroscopy duration, as well as lower and stable pacing thresholds. QRS duration is an established predictor of response to CRT [5], whereas its changes from preimplantation to post-implantation are also considered significant predictors of response to CRT [30]. Previous studies have demonstrated that LBBAP can significantly shorten the paced QRS duration [25, 31]. Results of the present study revealed significantly declined QRS durations in both LBBAP and BIVP groups, but the former were shorter than the latter. Additionally, LBBAP significantly improved LVEF, LVEDD, BNP and NYHA class at 24-month follow-up compared with baseline, while it also resulted in shorter QRS and lower BNP, than BIVP. Despite a lack of statistical significances, LBBAP caused a slight greater improvement in LVEF, as well as a greater reduction in LVEDD and NYHA class compared to BIVP. Collectively, these results demonstrated that LBBAP might be advantageous over BIVP in improving electrocardiographic and echocardiographic outcomes.

Although previous studies demonstrated that CRT therapy can significantly improve LVEF [32], only a few patients dramatically exhibited this effect, and have, therefore, been termed super-response [33]. We defined super-response as a final LVEF ≥ 50% at any point during follow-up [34]. The super-response rate observed in both groups in the present study was slightly lower than that previously reported [29, 35]. Additionally, the percentage of patients with a QRS less than 130 ms was significantly higher in the LBBAP than BIVP group, although the rate of super response was similar between the groups. The seemly paradoxical result may be due to differences in the mechanisms of LBBAP and BIVP therapy. Particularly, LBBAP directly paces the left bundle branch bypassing the block region, thereby generating a physiological cardiac conduction by synchronizing delayed LV activation and intrinsic activation in the right ventricle [19, 29]. In contrast, BIVP paces two non-physiological sides in the ventricles for resynchronization between the left and right ventricle, thereby prolonging LV activation time [36, 37]. Consequently, LBBAP can shorten QRS duration and significantly improve LVEF compared to BIVP. Another possible reason may be due to the small sample size of patients and the non-randomized study design.

Previous studies have reported successful correction of CLBBB by LBBAP, with high success rates achieved [20, 38]. For example, Vijayaraman et al. [35] reported that LBBAP resulted in a high success rate (88%) of CLBBB correction in patients. By comparison, our study obtained a lower success rate (61.90%) for CLBBB correction, partly due to operator experience at the early stage. In addition, some CLBBB patterns, such as left ventricular slow conduction caused by myocardial lesions and scar or conduction disorder caused by distal branch of left bundle branch, cannot be completely corrected using His-Purkinje-mediated ventricular activation [39]. We performed subgroup analysis based on CLBBB correction or failed correction by LBBAP, and found that patients with CLBBB correction exhibited significantly improved cardiac functions, such as LVEF, LVEDD, BNP, and NYHA class, compared with those with failed CLBBB correction. These results, together with those from a recent study [39], further demonstrated that correction of CLBBB is associated with improved cardiac function in patients after LBBAP. A possible explanation for this is that correction of CLBBB by LBBAP restores electrical and mechanical synchrony of the left ventricle. Taken together, these findings indicate that LBBAP significantly improves echocardiographic and clinical parameters compared with BIVP.

As previously discussed, LBBAP might be advantageous over BIVP in treating patients with HF and CLBBB. We attempted to explore which patients with HF are best suited for LBBAP. Interestingly, we found that the preoperative 12-ECG lead V6 ventricular activation time (VAT) in patients with CLBBB correction was significantly shorter than in those that with failed CLBBB correction. Our results are consistent with previous studies which have reported that patients with CLBBB exhibited LVAT greater than 60 ms in leads V5 and V6 [40, 41], indicating that prolonged LVAT may be an indicator of CLBBB. However, we observed that V6 VAT in patients with CLBBB correction was slightly shorter than that of patients with failed CLBBB correction by LBBAP, suggesting that patients with moderately prolonged LVAT in preoperative ECG are hence more easily to achieve CLBBB correction. Therefore, we speculate that moderately prolonged LVAT implies the presence of a block site in the proximal part of left bundle branch, hence LBBAP can immediately cross the conduction block site and effectively correct CLBBB. In contrast, cases with significantly prolonged LVAT may partly result from the conduction block site of distal part of left bundle branch or left ventricular conduction delay caused by myocardial lesion itself or myocardial scar, which can be failed to be corrected by LBBAP. We also found that the rate of full CLBBB correction in patients with QRS with a notch in preoperative limb leads was significantly higher than that of patients with no QRS with a notch in limb leads after pacing. A recent study indicated that about 1/3 of all LBBB cases, diagnosed by conventional criteria, may be a combination of LV hypertrophy and left anterior fascicular block, and not true CLBBB [42]. Moreover, Wagner et al. [42] proposed that true CLBBB had longer QRS duration and mid-QRS notching. Notches in limb leads represents the time when the electrical depolarization wave front reaches the endocardium of the LV and the epicardium of the posterolateral wall [42], indicating that notch in leads is an indicator of true CLBBB. Our results demonstrated that patients with moderately prolonged LVAT and mid-QRS notching in the limb leads in preoperative ECG could easily achieve full CLBBB correction with LBBAP. Taken together, these results showed that patients with heart failure and CLBBB, particularly those with moderately prolonged LVAT and QRS notch in limb leads, respond better to LBBAP, thereby exhibiting beneficial clinical outcomes.

There have been some articles published about the comparisons between LBBAP and BIVP [23, 43], which have demonstrated that LBBAP is an alternative method to BIVP in CRT treatment, our results were consistent with them. Compared with these studies, we have some advantages: (1) We had a longer duration of follow-up for 24-month; (2) We performed subgroup analyses by CLBBB correction and found that patients with CLBBB correction may get a better clinical outcome compared with those who with failed CLBBB correction; (3) We tried to analysis which type of patients would respond better to LBBAP through scanning baseline ECG characteristics. However, more larger sample size researches are needed to verify our conjecture.

There were several limitations in the present study. First, the small number of patients and non-randomized design may lead to inadequate power. LBBAP was in its early phase of clinical application in our hospital, and was only used in a small size of patients. Second, the definition of strict criteria for LBB capture was not used in the study. The characteristics of the ECG and the EGM in the LBBAP procedure, such as stim-LVAT, paced QRS morphology, and discrete component in the EGM, as the indirect criteria for LBB capture, were mainly used to distinguish LBBP from LVSP in this study. Indeed, it was difficult to distinguish them accurately in some cases. Currently, Wu et al. [16] proposed that retrograde His potential on the HBP lead and/or anterograde left conduction system potentials on the multielectrode catheter during LBBP were defined as the criteria for direct LBB capture, which could be used to distinguish LBBP from LVSP more accurately. Furthermore, the morphological characteristics of CLBBB in our study were not strictly met Strauss's criteria, which may lead to an underestimation of the efficacy of LBBAP compared to BIVP for CRT. Moreover, it was clear that the number of patients with renal dysfunction was significantly lower in the LBBAP group than that of the BIVP group, which may affect the lower hospitalization rate and mortality rate of LBBAP compared with BIVP. Since the renal dysfunction was recognized as a risk factor for higher HF hospitalization and all-cause mortality [44]. Collectively, we initially compared efficacy and clinical benefits of patients with HF and CLBBB, between LBBAP and BIVP groups, for 24-month follow-up. However, this was a small observational study with possible selection bias. Therefore, further studies using larger sample sizes are required to validate these findings.

In summary, LBBAP can significantly synchronize LV activation by correcting CLBBB and promote LV reverse remodeling. LBBAP is more effective in shortening QRS duration and enhancing echocardiographic and clinical responses than BIVP, therefore, may be a better candidate for resynchronization therapy for patients with HF and CLBBB.