The Effect of Scar and His-Purkinje and Myocardium Conduction on Response to Conduction System Pacing

Authors, Journal, Affiliations, Type, DOI

• Authors: Marina Strocchi, Karli Gillette, Aurel Neic, Mark K. Elliott, Nadeev Wijesuriya, Vishal Mehta, Edward J. Vigmond, Gernot Plank, Christopher A. Rinaldi, Steven A. Niederer
• Journal: Journal of Cardiovascular Electrophysiology 2023;34(4):984–993
• Affiliations: King's College London; BioTechMed-Graz and Gottfried Schatz Research Center, Medical University of Graz; NumeriCor GmbH (Graz); Guy's and St Thomas' NHS Foundation Trust (London); University of Bordeaux, CNRS, and IHU Liryc (Bordeaux)
• Type: Computational modelling study (in silico trial)
• DOI: 10.1111/jce.15847

Overview

This in silico study simulated ventricular activation on 24 patient-specific four-chamber heart meshes with His-Purkinje systems and proximal LBBB, then tested how different conduction pathologies and scar patterns modify the response to CRT, His bundle pacing (HBP), left bundle branch pacing (LBBP), HOT-CRT and LOT-CRT. Three findings stand out. First, severe slowing of the left His-Purkinje system attenuates the benefit of CSP and makes CRT relatively better, while HOT-CRT and LOT-CRT restore synchrony. Second, slow myocardial conduction (with normal His-Purkinje conduction) favours CSP over CRT. Third, septal scar makes CSP ineffective while CRT still works, whereas lateral wall scar makes CRT ineffective while CSP still works. The study provides a mechanistic framework for selecting the resynchronisation strategy according to the underlying conduction pathology rather than applying one approach to all CRT candidates.

Keywords

Cardiac resynchronization therapy; Heart failure; His bundle pacing; Left bundle pacing; Conduction system pacing; Left bundle branch block

Key Takeaways

Methods

• 24 one-millimetre-resolution four-chamber heart meshes generated from heart failure patients in prior CRT trials, with synthetic His-Purkinje systems (three LV fascicles, one RV septal fascicle, RV moderator band).
• Eikonal model of electrical activation; myocardial conduction velocity 0.6 m/s along fibres and 0.24 m/s transverse; Purkinje velocity 3.0 m/s.
• Seven baseline scenarios combined with proximal LBBB:
  1. Normal His-Purkinje conduction.
  2. Mild diffuse LV His-Purkinje slowing (conduction velocity 2.1 m/s, 70% of healthy).
  3. Severe diffuse LV His-Purkinje slowing (1.05 m/s, 35% of healthy).
  4. Mild myocardial slowing (0.42 m/s, 70%).
  5. Severe myocardial slowing (0.21 m/s, 35%).
  6. Large septal non-conducting scar.
  7. Large lateral wall non-conducting scar.
• Pacing modes: CRT (RV apex + LV epicardial lateral wall at the latest-activated point, VV 0), selective HBP, selective LBBP with AV optimisation, HOT-CRT (HBP + LV epicardial lead), LOT-CRT (LBBP + LV epicardial lead).
• Electrical synchrony measured by BIVAT-90 (time to activate 90% of the ventricles), LVAT-95, and dyssynchrony indices.

Normal or mildly slowed His-Purkinje conduction

• With proximal LBBB and otherwise normal conduction, HBP and LBBP produced better LV and biventricular activation than CRT: LVAT-95 51.2 ± 4.9 ms (HBP) and 59.0 ± 6.5 ms (LBBP) vs 79.8 ± 7.7 ms (CRT); BIVAT-90 42.6 ± 3.8 and 48.0 ± 5.3 vs 68.0 ± 8.0 ms (all P < .01).
• Mild LV His-Purkinje slowing (2.1 m/s) preserved the CSP advantage over CRT (P < .01 for all metrics).

Severe LV His-Purkinje slowing

• At 1.05 m/s, CRT became comparable to HBP and LBBP for LV activation (LVAT-95: HBP 113.4 ± 8.7, LBBP 108.7 ± 10.2, CRT 108.5 ± 9.2 ms); BIVAT-90 was shorter with CRT than HBP (93.0 ± 8.9 vs 101.5 ± 7.8 ms; P < .05), while LBBP was comparable to CRT (89.1 ± 9.8 ms).
• HOT-CRT and LOT-CRT significantly improved both LV and biventricular activation: BIVAT-90 87.6 ± 6.7 ms (HOT-CRT) and 73.9 ± 7.6 ms (LOT-CRT) vs 93.0 ± 8.9 ms (CRT); LVAT-95 95.1 ± 7.6 and 89.9 ± 8.4 vs 108.5 ± 9.2 ms (P < .01).
• The authors attribute the HBP disadvantage to the longer distance the stimulus must travel along diseased LV fascicles; LBBP paces more distally.

Slow myocardial conduction

• With myocardial conduction velocity reduced to 0.42 m/s, CSP was superior to CRT: BIVAT-90 54.7 ± 5.1 ms (HBP) and 59.8 ± 6.3 ms (LBBP) vs 83.6 ± 11.8 ms (CRT); similar at 0.21 m/s (97.1 ± 9.9 and 101.5 ± 10.7 vs 134.5 ± 24.1 ms).
• HOT-CRT and LOT-CRT were not significantly better than HBP and LBBP in this scenario.

Scar

• Septal scar: HBP and LBBP were ineffective — activation remained at baseline levels (BIVAT-90 117.2 ± 14.9 and 118.1 ± 12.9 vs baseline 119.1 ± 10.8 ms; P = .9). CRT significantly shortened activation (85.1 ± 9.8 ms; P < .01). HOT-CRT and LOT-CRT approached CRT but did not exceed it.
• Lateral wall scar: CRT was ineffective (BIVAT-90 119.0 ± 17.4 vs baseline 124.8 ± 13.2 ms; P = .68), while HBP and LBBP reduced activation (105.3 ± 12.8 and 102.6 ± 12.5 ms; P < .01). HOT-CRT and LOT-CRT were equivalent to CSP alone because the LV lead sits near the scar.
• If Purkinje fibres within the septal scar remained viable, CSP became more effective than CRT — a sensitivity analysis that may explain some clinical heterogeneity.

Interpretation

• The response to CSP depends on whether the conduction system distal to the pacing site is viable and fast; septal scar creates a distal block that CSP cannot bypass.
• CRT is less dependent on the His-Purkinje system and therefore tolerates septal scar better, but it fails when the lateral wall is scarred.
• HOT-CRT and LOT-CRT combine the two mechanisms and are most useful when disease is severe and diffuse in the His-Purkinje system.
• The model predicts that slow myocardial conduction (a "myopathic" substrate) favours CSP, whereas slow conduction-system conduction favours CRT or hybrid approaches.

Limitations of the document

• Computational model, not clinical data — the His-Purkinje system is synthetically generated and not patient-specific, and the model assumes acute electrical synchrony correlates with long-term functional response.
• Perfect delivery assumed — HBP and LBBP are simulated as selective capture with no lead malposition, perforation, or non-selective capture; LBBP is always simulated with AV optimisation, which may overestimate benefit.
• Scar is modelled as uniformly non-conducting, without gradations of density or border-zone heterogeneity; Purkinje fibres within scar are assumed non-conducting except in a sensitivity analysis.
• His-Purkinje conduction velocity cannot be measured in vivo, so part of the model cannot be directly validated clinically.
• Limited clinical validation — the model reproduces LBBB activation and published metrics, but the pacing predictions are not tested prospectively.
• No long-term outcomes, and no consideration of fibrosis progression, heart failure remodelling, or arrhythmia.
• Funding: Wellcome/EPSRC, NIH, ERC, BHF and French government grants; one author is employed by NumeriCor GmbH.

Key Concepts Mentioned

• concepts/Conduction-System-Pacing — the umbrella technique
• concepts/His-Bundle-Pacing — one CSP modality
• concepts/Left-Bundle-Branch-Area-Pacing — the other CSP modality
• concepts/Cardiac-Resynchronization-Therapy — the comparator
• concepts/Late-Gadolinium-Enhancement — the key effect modifier
• concepts/Conduction-System-Pacing — the conduction network
• concepts/LOT-CRT — hybrid CSP plus LV pacing
• concepts/Conduction-System-Pacing — hybrid His plus LV pacing

Key Entities Mentioned

• King's College London — lead institution
• NumeriCor GmbH — commercial modelling collaborator
• Wellcome/EPSRC Centre for Medical Engineering — funding

Wiki Pages Updated

• Created: wiki/sources/scar-csp-jce-2023.md
• Updated: wiki/concepts/Conduction-System-Pacing.md
• Updated: wiki/concepts/LOT-CRT.md
• Updated: wiki/concepts/Left-Bundle-Branch-Area-Pacing.md
• Updated: wiki/sourceindex.md
• Updated: log.md