Author + information
- Received May 29, 2018
- Revision received July 10, 2018
- Accepted July 26, 2018
- Published online November 19, 2018.
- Advay G. Bhatt, MD∗ (, )
- Dan L. Musat, MD,
- Nicolle Milstein, MS,
- Jacqueline Pimienta, BA,
- Laura Flynn, NP,
- Tina Sichrovsky, MD,
- Mark W. Preminger, MD and
- Suneet Mittal, MD
- Valley Health System and the Snyder Center for Comprehensive Atrial Fibrillation, Ridgewood, New Jersey
- ↵∗Address for correspondence:
Dr. Advay G. Bhatt, Valley Health System, 970 Linwood Avenue, Paramus, New Jersey 07652.
Objectives This study sought to evaluate the clinical and procedural characteristics impacting outcomes during implementation of a His bundle pacing (HBP) program in a real-world setting.
Background Right ventricular pacing is associated with an elevated risk of heart failure, but device reprogramming and upgrades have significant challenges. HBP has emerged as an alternative and is reported to be highly successful in the hands of highly experienced centers.
Methods All patients referred for permanent pacemaker implantation at the Valley Hospital (Ridgewood, New Jersey) between October 2015 and October 2017 were evaluated; a subset of 24% was selected for HBP.
Results Permanent HBP was feasible with an acute implant success rate of 75%. HBP in the presence of bundle branch block (64% vs. 85%; p = 0.05) or complete heart block (56% vs. 83%; p = 0.03) was significantly less successful. The pattern of atrioventricular block in combination with bundle branch block (BBB) further affects outcomes. HBP is highly successful across the spectrum of atrioventricular block pattern severity in the absence of BBB. In the presence of BBB, Mobitz II AV block and complete heart block significantly attenuated HBP success compared with Mobitz I atrioventricular block (62% vs. 100%; p = 0.02). A rising threshold was observed in 30%, and 8% required lead intervention.
Conclusions HBP was feasible and readily learned with a high implant success in the hands of experienced electrophysiologists without prior exposure to the technique. BBB and atrioventricular block pattern appears to affect success. The technique is limited by a high rate of rising thresholds and lead intervention. These data have important implications for patient selection.
Approximately 1 million pacemakers are implanted annually throughout the world, primarily for sinus node dysfunction or heart block (1). A high burden of right ventricular (RV) pacing is associated with an elevated risk of heart failure (HF) (2–4) resulting from electromechanical dyssynchrony (5–7). The incidence and time course for developing HF remains poorly elucidated; however, complete heart block (CHB) is associated with an increased risk of HF within 6 months of pacemaker implant (8).
Device algorithms to minimize ventricular pacing (9,10) are not feasible with persistent advanced heart block and occur at the expense of physiologic atrioventricular synchrony. The common solution to mitigate these downstream concerns is biventricular pacing, but the approach is limited by implant failure, complications, and nonresponse (11,12).
His bundle pacing (HBP) has emerged as an alternative that facilitates physiologic pacing and therefore atrioventricular synchrony need not be sacrificed. The individual His bundle (HB) fibers are predestined for the right or left bundles (longitudinally dissociated) and intra-His block overcome with pacing the HB distal to the site of block but proximal to the bifurcation of the right and left bundles (13,14). The anatomy and current tools do not easily allow precision mapping and lead delivery distal to the site of block; however, HBP success remains high because of recruitment of the HB with higher output pacing (15).
To date, much of the HBP literature has emanated from a few select centers; in the hands of these highly experienced operators and centers, HBP is reported to be highly successful with improved clinical outcomes during several years of follow-up. These results may not be generalizable to laboratories implementing a HBP program, however. Therefore, the aims of this analysis were to evaluate the clinical and procedural characteristics during implementation of a HBP program in a real-world setting, identify electrophysiologic factors that affect HBP, and report on the outcomes in these patients.
We implemented our HBP at the Valley Hospital (Ridgewood, New Jersey) in October 2015, and all patients referred for permanent pacemaker implantation in accordance with American College of Cardiology Foundation/American Heart Association/Heart Rhythm Society guidelines were evaluated for HBP and included in this single-center, observational, and retrospective analysis (16).
Although no formal inclusion criteria were prespecified, HBP was preferred in patients in whom prolonged atrioventricular intervals were expected either during sinus rhythm or atrial pacing and in those expected to have a high burden of ventricular pacing, but was at the discretion of the implanting electrophysiologist. Initially, patients with CHB or who required temporary pacing were excluded; however, as our experience deepened, HBP was expanded to patients with intermittent CHB or stable escape rhythms. In such patients, nonselective His bundle pacing (NSHBP) was preferentially targeted to enhance safety by ensuring capture of some ventricular myocardium (17).
The institutional review board approved the observational study and all patients included in this analysis provided informed consent for inclusion of their data.
Assessment of underlying conduction disease
The pattern of atrioventricular block (Mobitz I, Mobitz II, or CHB) and presence of bundle branch block (BBB) were ascertained in all patients undergoing HBP. These features could not be reliably determined with permanent atrial fibrillation or pacing, however, and such patients were excluded from the analysis of the relationship of conduction disease and HBP outcomes.
The implant procedure was performed using a 69-cm Medtronic SelectSecure 3830 pacing lead (Medtronic, Minneapolis, Minneapolis), delivered through either fixed (C315-HIS) or deflectable (C304) curve sheaths (Medtronic) (17). HB mapping was performed using the atrial channel (gain setting of 0.05 mV/mm and sweep speed of 50 mm/s) on the Medtronic pacing system analyzer and electrophysiologic recording system. The region with a HB potential was targeted for detailed unipolar pace mapping to evaluate the paced QRS duration (QRSd) and morphology. After confirming HB capture, the lead was torqued 5 to 10 times to fix the lead to the target site. HBP thresholds <2.5 V at 1.0 ms was commonly accepted. The HB was not localized with a diagnostic electrophysiology catheter. A backup pacing lead was permissible but not routinely used.
In the event that acceptable HBP was not feasible, the following options were available: (1) implant an RV apical lead in patients with normal left ventricular (LV) function or (2) biventricular pacing in patients with LV dysfunction.
Selective His bundle pacing (SHBP) and NSHBP (Figure 1) was determined according to published standardized definitions (18). SHBP exclusively recruits the His-Purkinje system, resulting in identical intrinsic and paced QRS morphologies or QRS normalization (with BBB). NSHBP results from a fusion of septal and HB capture at high-output pacing; only septal myocardium is captured at low output.
HBP was not considered successful if: (1) the paced QRSd was >130 ms in the presence of a normal intrinsic QRSd or (2) the paced QRSd was greater than the intrinsic QRSd in patients with BBB. QRSd (paced) was measured at a sweep speed of 50 mm/s and double gain in leads I and V1 in a blinded manner by 2 electrophysiologists (A.G.B., D.L.M.) and a mean QRSd was calculated.
The patients were evaluated in device clinic according to standard institutional practice with a post-implant check at 1 to 2 weeks, 1 month, 6 months, 1 year, and annually thereafter. During each visit, a clinical assessment for HF, device interrogation, and 12-lead electrocardiogram analysis of QRS morphology and QRSd were performed. The performance of the Medtronic SelectSecure 3830 lead was evaluated for sensing abnormalities, rising capture thresholds, or loss of HB capture to determine if device reprogramming or lead revision was necessary. HF readmissions and death were ascertained by reviewing the available medical records.
Continuous variables were reported as mean ± SD and categorical variables were expressed as counts and percentages. Differences between the subgroups of HBP success versus failure were calculated either with the Student’s t-test or 2-tailed Fisher exact test for continuous and categorical variables, respectively. A p value <0.05 was considered statistically significant.
Between October 2015 and 2017, there were a total of 427 new pacemaker implants or upgrades, in which HBP was attempted in 101 (24%). The baseline characteristics of patients undergoing HBP versus non-HBP are compared in Table 1. HBP was attempted in younger patients less likely to have hypertension or antecedent atrial tachyarrhythmias, but more likely to be men, have a pacemaker, or have LV dysfunction. Finally, patients undergoing HBP were more likely to have Mobitz I or Mobitz II block and less likely to have sinus node dysfunction.
Of the 101 patients undergoing HBP, 47 (47%) had a normal QRS complex (QRSd: 92 ± 15 ms), 8 (8%) were 100% paced, and 46 (46%) had BBB (QRSd: 146 ± 16). There were 19 left BBB (QRSd: 153 ± 38 ms), 8 isolated right BBB (QRSd: 136 ± 15 ms), 18 bifascicular blocks (QRSd: 144 ± 16 ms), and 1 nonspecific intraventricular conduction block (QRSd: 146 ms).
Acute implant outcomes
HBP was initially thought to have been successful in 89 of 101 (88%) attempts; however, when we carefully adjudicated outcomes based on electrocardiograms and pacemaker interrogation, it was surprising that 13 in whom HBP was initially thought successful was in fact subjected to high septal capture and not nonselective HB capture. The actual success rate of HBP was 76 (75%). In the 12 patients in whom HBP was recognized as not successful during the procedure, RV apical pacing and biventricular pacing was performed in 11 and 1, respectively.
The mean fluoroscopy time was 10 ± 7 min with successful HBP; in contrast, the mean fluoroscopy time was 24 ± 15 min with unsuccessful HBP (p < 0.001). The latter includes the time required to implant an alternative pacing system. There was no significant difference in either the success of HBP or fluoroscopy usage among the first or second half of our HBP cohort, suggesting minimal impact of a learning curve. The only complication was an isolated pneumothorax that resolved without intervention. There were no lead dislodgements or cardiac perforations.
The sensed R wave, impedance, and HB capture threshold were 5.3 ± 3.8 mV, 507 ± 147 Ω, and 1.2 ± 0.8 V at 1.0 ms, respectively. Among patients with CHB, the nonselective HB and absolute (septal) capture thresholds were 1.4 ± 1.0 V at 1.0 ms and 0.5 ± 0.1 V at 1.0 ms, respectively. A HB capture threshold >2.5 to 3.3 V was accepted in 10 of 76 (13%) based on the clinical circumstance and inability to improve thresholds with extensive HB mapping; 5 (50%) were observed in the first 33 attempts.
A back-up pacing lead was considered necessary in 3 patients. The first was planned to undergo atrioventricular junction ablation, and we had no previous experience with HBP in this setting. The second developed CHB with ventricular asystole during lead fixation. Although the HB capture threshold was excellent, we were uncertain as to the clinical significance of this finding and, for safety, inserted a back-up pacing lead. The third patient with advanced conduction disease presented with syncope and became unstable during implant, requiring a back-up lead. In patients with pacing-induced cardiomyopathy, the RV lead was maintained and connected to the LV port of a biventricular pacemaker with the exception of 1 patient, in whom the RV lead was manually extracted without difficulty. There were 2 patients in whom the original intent was to implant a biventricular implantable cardioverter-defibrillator, but an LV lead could not successfully be delivered. In these patients, HBP was used for resynchronization; the implantable cardioverter-defibrillator lead remained in the RV apex.
Table 2 summarizes the effect of selective and nonselective HBP on QRSd. In the absence of BBB, selective and nonselective HBP minimally affected QRSd, whereas in the presence of BBB, QRSd markedly decreased with both selective and nonselective HBP. The QRSd with high septal pacing (167 ± 39 ms) was, on average, 34 ms longer than baseline. The interobserver correlation for measurement of QRSd was 96.1%.
When comparing clinical features between patients with and without successful HBP (Table 3), surprisingly, patients with successful HBP were older. We sought to determine the effect of QRS morphology and atrioventricular block pattern on the likelihood of achieving a successful HBP implant. HBP was significantly more successful in patients without BBB (n = 39 of 47, 83%) compared with patients with BBB (n = 30 of 46, 65%; p = 0.05, Figure 2); there was no difference in outcomes based on BBB subtype. HBP was successful in 40 of 48 (83%) patients with Mobitz I or II block, but only in 10 of 18 (56%; p = 0.03) patients with CHB (Figure 3). Finally, the combined effect of BBB and atrioventricular block pattern on HBP was evaluated (Figure 4). The lowest success rates were observed in patients with CHB and underlying BBB (5 of 12, 42%).
Other procedural characteristics
The highest HBP success was observed when the 3830 lead could be delivered using the fixed-curve Medtronic C315-HIS sheath. In these instances, HBP was achieved in 65 of 73 patients (89%). In contrast, when the deflectable-curve Medtronic C304 sheath was required following an unsuccessful attempt with a C315-HIS sheath, HBP was successfully achieved in only 11 of 28 patients (38%; p = 0.001).
Lead follow-up and intervention
One-month follow-up, including threshold testing, was performed in 72 (95%) of 76 patients following HBP implantation. Follow-up was completed in 56 (74%) at 6 months, 36 (47%) at 1 year, and 17 (22%) at 2 years. A total of 7 (9%) were considered lost to follow-up; this included 3 who did not follow-up after wound check and 4 who did not return following the 1-month post-implant check. One patient missed the 1-month post-implant check and thereafter resumed follow-up.
The HB capture threshold, impedance, and sensed R-wave at 1-month follow-up were 1.5 ± 1.2 V at 0.7 ± 0.3 ms, 379 ± 96 Ω, and 6.1 ± 4.5 mV, respectively. At 6 months’ follow-up, the values were 1.7 ± 1.4 at 0.7 ± 0.3 ms, 377 ± 97 Ω, and 6.1 ± 4.0 mV. At 12 months’ follow-up, the values were 1.7 ± 1.5 V at 0.6 ± 0.2 ms, 381 ± 81 Ω, and 6.0 ± 4.3 mV. Finally, at 24 months’ follow-up, the values were 1.8 ± 1.5 V at 0.6 ± 0.2 ms, 373 ± 83 Ω, and 4.4 ± 2.5 mV. Although the pacing threshold voltage increased significantly between the implant and 1-month follow-up (1.2 vs. 1.7 V; p = 0.04), the pulse width (1.0 vs. 0.7; p < 0.001) and impedance (507 vs. 379; p < 0.001) were significantly lower.
During follow-up, 24 of 76 (32%) patients experienced early or late thresholds ≥2.5 V @ 1.0 ms; however, in 16 (67%), HBP could be acceptably maintained by increasing the pacing output. Of the remaining 8 (33%), 6 required a lead revision; high septal pacing was later accepted in 2. The intraoperative pacing threshold in each of these 8 patients was <1.5 V. The time course from implant to lead revision was within 30 days in 3, 90 days in 2, and 8 months in 1. There were no overt lead dislodgements observed during follow-up. HBP was successfully reestablished in 2 patients; the remainder was converted to RV apical pacing.
Although there was no significant difference in the success of HBP among the first (76%), second (67%), and third (83%) tertiles of HBP attempts (Figure 5), the incidence of increasing pacing thresholds and need for lead revisions declined with experience. Among successful HBP attempts, a rising pacing threshold was encountered in 10 of 25 (40%) patients in the first tertile, 7 of 22 (32%) patients in the second tertile, and in 7 of 29 (24%) patients in the third tertile. Similarly, the need for lead revision occurred in 3 of 10 (30%) patients with a rising pacing threshold in the first tertile, 2 of 7 (29%) patients in the second tertile, and 1 of 7 (14%) patients in the third tertile.
There were 2 deaths among patients with successful HBP, with 1 from preexisting advanced HF and the other from a noncardiac cause. There were 3 deaths from noncardiac causes and 1 HF admission among patients with unsuccessful HBP. There was a trend toward lower mortality in patients with successful HBP (2 of 76 patients, 2.6%) compared with patients with unsuccessful HBP (3 of 25, 12%; p = 0.10).
Although follow-up echocardiographic data for the entire cohort is limited, we do have echocardiographic information on 13 (72%) of 18 patients with preexisting LV dysfunction in whom HBP was successful. In these patients, the mean LV ejection fraction improved from 32% to 47% (p = 0.002) at a mean follow-up of 273 days (range 41 to 701 days).
The main findings of this study are that: 1) establishing permanent HBP was feasible and readily learned with an acute implant success rate of 75% in the hands of experienced electrophysiologists without previous exposure to the technique; 2) the factors that significantly affected the success of HBP was the presence of BBB or CHB compared with Mobitz I or Mobitz II atrioventricular block; 3) the severity of atrioventricular block in combination with BBB further affected the success of HBP; and 4) a high proportion of patients (32%) experienced rising pacing thresholds upon follow-up and required either an increase in pacing output (two-thirds) or a lead revision (one-third).
The largest published experiences of HBP in the setting of atrioventricular block by Barba-Pichardo et al. (19) in 2010 (n = 182) and Sharma et al. (17) in 2015 (n = 94) achieved permanent HBP in up to 80%. More recently, HBP has been described to be successful in 92%, with a reduction in HF hospitalizations and a trend toward improved mortality (20). Although we successfully achieved HBP at a similarly high rate as the first published experiences and in general with excellent capture thresholds, there are several clinical challenges that need recognition when using the technique. First is that, despite HB mapping, a suboptimal threshold (>2.5 V at 1.0 ms) needed to be accepted in 13%. Second, a high proportion (32%) experienced rising HB capture thresholds during follow-up; the majority was managed with an increase in pacing output but at the expense of battery longevity. Last and most significantly, HBP appears to be associated with a higher need for lead intervention (8%), but, surprisingly, in each instance the implant thresholds were <1.5 V and there was no overt evidence of lead dislodgement. As such, this idiosyncratic failure rate is not well understood, other than possible microdislodgement.
These issues were more evident within the first 33 attempts; this suggests that although HBP may be readily achieved, there is a learning phase when using this technique and patients need systematic and close post-implant monitoring. The difficulty may be related to the use of new delivery sheaths and, more important, to torque the entire lead to achieve fixation with minimal tactile or visual feedback to ensure the fixation helix is fully embedded in the target region. HBP was reestablished in some patients requiring lead revision, suggesting that perhaps the lead was not appropriately fixed during the initial attempt. The challenges naturally lead to the speculation that the lead delivery tools and leads need further refinement and innovation to increase the short- and long-term success while reducing the need for lead intervention, although the published literature suggests a high success rate in experienced hands with the currently available tools.
To further understand the challenges of HBP, the anatomy of the HB must be fully appreciated and recognized that the target location is constrained; only a small window exists through which HBP may be achieved. The HB penetrates the central fibrous body and is, on average, 1.8 mm in diameter and 11 mm in length, with significant positional variation with respect to the membranous septum and coronary cusps (21). Kawashima et al. described 3 anatomic variants of the HB: 1) HB courses along the lower border on the membranous septum (47%); 2) HB coursed within the muscular septum (32%); and 3) the so-called “naked” HB in which the HB coursed onto the membranous septum (21%) (21). Kawashima et al. (22) further speculated that the proximity of the HB to the aortic root as the reason for high risk of heart block with transcatheter aortic valve replacement resulting from crushing of a calcified aortic valve and insertion of noncompliant or expanding prosthetic valve platform, which may be extrapolated to surgical aortic valve replacement from edema and fibrosis surrounding the sewing ring of the valve prosthesis. This may explain the trend toward reduced success of HBP in the presence of prior valve replacement, although this finding is at odds with a recent small case series published by Sharma et al. (23), in which HBP was highly successful (93%) in patients with prosthetic heart valves.
A speculative concern would be that an HB that courses within the muscular septum or more leftward may be less amenable to HBP given the need for deeper fixation of the lead (or higher pacing output) and accordingly more intense inflammatory and fibrotic reactions that then would lead to increasing thresholds or loss of capture. The evolution of HBP characteristics in the setting of antecedent conduction disease, resultant injury and remodeling from lead fixation, and higher output pacing are not fully known. In the context of anatomic constraints and evolving electrophysiologic properties of the His-Purkinje system, there may be an upper limit of success that may not be able to be practically overcome and there may be a significant rate of attrition from loss of capture or subtle changes in morphology from nonselective HB capture to RV septal capture that may go unrecognized over long-term follow-up.
Although the benefits of HBP are manifold (the potential for an upper limit of success, rate of attrition, and still nascent experience with HBP), better insight is needed to guide thoughtful decision-making surrounding individuals that stand to benefit from HBP compared with biventricular pacing or RV apical pacing with transvenous of leadless systems. The insight gained from this study in these regards is that HBP can be achieved in most patients with Mobitz I block and those without BBB; however, in patients with Mobitz II block and CHB, especially with underlying BBB, the success was more limited. Moreover, the presence of BBB and the pattern of atrioventricular block synergistically attenuated the success of HBP. These data have important implications for patient selection and suggest that HBP may not be an ideal first-line strategy in every patient by highlighting how the dynamic electrophysiologic substrate impacts HBP.
In our view, another important consideration in defining success may be the paced QRSd, especially with nonselective HB capture. Multiple studies have demonstrated that dyssynchrony and the benefits of CRT are not apparent when the QRSd is >130 ms (24). In this context, we strictly defined success to having a QRSd <130 ms in the absence of BBB or less than the native QRSd in the presence of BBB, which may account for the lower success rate compared with more contemporary reports. The width of the “pseudo-delta wave” resulting from septal activation in cases of nonselective HB capture and its contribution to overall QRSd, a so-called preexcitation index, may need to be factored in to minimize the risk of dyssynchrony.
Improved understanding of lead maturation dynamics in different subset of patients with varying levels of conduction and structural heart disease to guide patient selection is needed to identify factors that would predict loss of HBP and to model battery longevity and frequency of generator changes over a lifetime of pacing versus standard dual-chamber pacing. Although new tools may increase the acute success of HBP, the rate of attrition may be significant because of the previously mentioned anatomic and electrophysiologic factors that over time that would blunt or reverse the benefits of the technique. Ultimately, these issues need to be further evaluated in large multicenter trials of HBP.
The study is limited by the fact that it was performed at a single-center acute care hospital, with a small sample size, and individuals targeted for HBP were not randomly selected. The short and nonuniform follow-up may underestimate the concern of rising thresholds, loss of HB capture, and lead interventions. Moreover, echocardiographic data and hard clinical outcomes following HBP are limited.
HBP has emerged as a promising alternative form of physiologic pacing that is reported to be particularly successful in a select few and highly experienced centers. This study demonstrates that HBP may be readily learned by highly experienced electrophysiologists without prior exposure to the technique; however, the severity of heart block and presence of bundle branch block both impact acute procedural success and need to be factored in the clinical decision to pursue HBP. Moreover, the challenges of rising thresholds and lead revisions necessitate a systematic approach with close monitoring when learning the technique.
COMPETENCY IN MEDICAL KNOWLEDGE: A high burden of RV pacing is associated with an elevated risk of heart failure and mortality that is difficult to minimize with pacing algorithms and biventricular pacing is challenging. HBP has emerged as an alternative that allows physiologic pacing and thus atrioventricular synchrony may be maintained. HBP may be readily learned, but the challenges of rising thresholds and need for lead revisions necessitates a systematic approach with close implant monitoring when learning the technique.
TRANSLATIONAL OUTLOOK: Large randomized clinical trials are needed to better identify populations of patients that would benefit the most from HBP, confirm the clinical benefits, and obtain an improved understanding of lead maturation dynamics of the Medtronic 3830 lead in the His bundle position.
Dr. Mittal is a consultant to Abbott, Boston Scientific, and Medtronic. Dr. Bhatt is a consultant to Abbott and Medtronic. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- bundle branch block
- complete heart block
- His bundle
- His bundle pacing
- heart failure
- left ventricular
- nonselective His bundle pacing
- QRS diameter
- right ventricular
- selective His bundle pacing
- Received May 29, 2018.
- Revision received July 10, 2018.
- Accepted July 26, 2018.
- 2018 American College of Cardiology Foundation
- Tops L.F.,
- Schalij M.J.,
- Bax J.J.
- Kirk J.A.,
- Kass D.A.
- Merchant F.M.,
- Hoskins M.H.,
- Musat D.L.,
- et al.
- Wilkoff B.L.,
- Kudenchuk P.J.,
- Buxton A.E.,
- et al.
- Adabag S.,
- Roukoz H.,
- Anand I.S.,
- Moss A.J.
- Narula O.S.
- El-Sherif N.,
- Amay-Y-Leon F.,
- Schonfield C.,
- et al.
- Teng A.E.,
- Massoud L.,
- Ajijola O.A.
- Epstein A.E.,
- Ellenbogen K.A.,
- Freedman R.A.,
- et al.
- Vijayaraman P.,
- Dandamudi G.,
- Zanon F.,
- et al.
- Abdelrahman M.,
- Subzposh F.A.,
- Beer D.,
- et al.
- Sharma P.S.,
- Subzposh F.A.,
- Ellenbogen K.A.,
- Vijayaraman P.