Adopting Permanent his Bundle Pacing

Learning Curves and Medium-term Outcomes

Jhobeleen De Leon; Swee-Chong Seow; Elaine Boey; Rodney Soh; Eugene Tan; Hiong Hiong Gan; Jie Ying Lee; Lisa Jie Ting Teo; Colin Yeo; Vern Hsen Tan; Pipin Kojodjojo


Europace. 2022;24(4):606-613. 

In This Article


While HBP was acutely successful in 81.1% of patients, a further 12.4% of patients lost HBP capture during follow-up. Of those with persisting HBP capture, 11.3% experienced an equal to or greater than 1 V increase in HBP threshold, and 8.6% of patients had a 50% or greater decrease in R-wave amplitude. On an intention to treat basis, only 56.7% of patients in whom HBP was attempted, were HBP captured at thresholds of <2 V present after 259 days of follow-up. Five (2.6%) required ventricular lead repositioning. His bundle pacing was less likely to succeed in patients with broad QRS complex. There is a long learning curve of 30–40 cases associated with HBP, even for electrophysiologists experienced in pacing, before fluoroscopic, and procedural times plateaued.

Acute his Bundle Pacing Parameters

The overall acute procedural success rate in our study was 81% and the mean threshold on implantation was 1.3 V and 1.2 V on follow-up. This was comparable to the meta-analysis on His bundle pacing by Zanon et al.,[13] wherein they reported an average implant success rate of 84.8% (range between 35.4% and 100%), and acute (<3 months) and chronic (>3 months) thresholds of 1.76 and 1.79 V, respectively.

Sensing From his Bundle Pacing Lead

Although the R wave is <3 mV in 31.4% and 29.8% at implantation and at last follow-up, respectively, sensing issues requiring lead revision did not occur. Instead, ventricular sensing sensitivity or sensing configuration was adjusted. In the literature, oversensing and under-sensing issues requiring lead revision are uncommon (<1%).[10]

Loss of his Bundle Capture

Loss of HBP capture during a mean follow-up of 259 ± 213 days was recorded in 12.4% of our cohort. Majority (21 out of 29) of these patients had acceptable RV septal capture thresholds and therefore was deemed not to require lead revision. This was in large part due to the majority of our patients having normal LVEF and were receiving only dual-chamber pacemakers. Had such a loss of HBP capture rate occurred in an HF cohort receiving HBP for cardiac resynchronization therapy, the number of lead revisions required would be much higher. In the recommendations on HBP by an international collaborative group, there is no absolute HBP threshold cut-off, however, thresholds of 2.5 V at 1 ms were considered to be reasonable for non-dependent patients and lower thresholds for dependent patients.[12] In this study, the threshold of 3.5 V and above was considered as loss of capture since pacing with even a small safety margin, such as 0.5 V (therefore output of 4.0 V and above) would result in early battery depletion. The decision for lead revision was made upon considering patient factors including the need for physiological pacing, presence of RV septal capture, anticipated RV pacing burden, and battery longevity. In four SND patients, whom there was either loss of ventricular capture or high ventricular threshold, the pacing mode was switched to AAI, to avoid repeat surgery. The other factor that resulted in our low rate of ventricular lead revision could be our decision to perform HBP predominantly in patients with minimal ventricular pacing requirements such as SND at the start of our programme. With increasing experience, HBP was extended to patients with other pacing indications, such as AVB and CRT, such that eventually, more than half of our patients had AVB as their pacing indication.

Other studies have reported a loss of HBP capture rate between 4% and 17% during follow-up.[4,6,14–17] The exact aetiology is unknown but the proposed mechanisms include micro-dislodgement due to inadequate fixation of the lead, tricuspid valve motion or tricuspid regurgitation, excessive swinging motion of the lead, local fibrosis, progression of disease distal to the pacing site, and exit block.[6]

A single-centre study on the intermediate-term performance and safety of His bundle pacing leads reported an increase in threshold 1 V or more in 28% and thresholds of 2.5 V or greater in 24% of patients for whom HBP was acutely successful.[16] Our study showed a lower percentage of patients (11.3%) with ≥1 V increase in threshold but a similar 20.1% of patients (including those who had lost HBP capture) had HBP thresholds of 2.5 V or greater. Hence, in our practice, commercially available models with the largest battery capacity were used during HBP to in part mitigate this risk.

Learning Curve

Several studies have examined the learning curve of HBP. In a study by Dawson et al.,[18] they reported that procedural and fluoroscopy time improved with operator experience. Keene et al.,[15] reported a mean fluoroscopy time of 11.7 ± 12 min, and concluded that fluoroscopy times were shorter with increasing experience, levelling out after 30–50 cases. In this study, mean procedural and fluoroscopy times were 105.6 ± 36.8 and 13.9 ± 9.4 min, both of which plateaued after 30–40 procedures. Threshold on follow-up improved as the number of procedures increased and levelled after about 20 cases.

Clinical Predictors of Acute Success

Our study indicated that broad QRS complex was associated with failure in His bundle pacing. This was similar to a study by Bhatt et al.,[14] wherein their data showed that acute success is less likely in patients with complete heart block and bundle branch blocks. Barba-Pichardo et al.,[19] also reported lower HBP success in AVB patients with broad QRS complex compared to those with narrow QRS.


Although there were no pre-specified inclusion or exclusion criteria for HBP, the decision of whether to perform HBP and accept implant parameters was left to each operator's discretion. This would affect the eventual patient mix which may in turn influence HBP thresholds, medium-term lead performance, and need for lead revisions. The study is also limited by its non-randomized design.