Methods

Study population

This was a prospective, single center observational study. Consecutive patients referred to one electrophysiology team, between October 2019 to August 2021, for device implantation for bradycardia or cardiac resynchronization were offered conduction system pacing which was explained to them as a newer non-standard method which may have advantages but was associated with a research MRI scan. The study was approved by the local ethics committee (REC 19/YH/0174). Consenting patients were enrolled. They underwent a pre-implant cardiac MRI to which the cardiac catheterization laboratory team were kept blinded.

Cardiac MRI

Cardiovascular magnetic resonance was performed in a 1.5 Tesla (Aera, Siemens Medical Solutions, Erlangen, Germany) using a standard clinical scan protocol including LGE imaging following a bolus of 0.1mmol/kg of Gadobutrol (marketed as Gadovist, Bayer Pharma AG, Berlin, Germany). In the majority of cases, scar imaging was undertaken using a free-breathing motion corrected sequence as detailed by Captur et al8. (25/35, 71%) also had dark blood late gadolinium enhancement imaging which increases sensitivity for detecting subendocardial scar9.

Image Analysis

Cardiovascular magnetic resonance imaging analysis was performed using CVI42 software (Version 5.13.7, Calgary, Canada) blinded to clinical parameters. The presence and extent of myocardial scar in the basal septum (defined as the inferoseptal and anteroseptal segments in the most basal 3 complete slices) was assessed by an experience level 3 CMR operator. The observer was aware of the hypothesis regarding the basal septum but had no access to information on the procedural outcome. Scar burden was quantified either using full width half maximum semiautomated technique7with manual review, or manually (for dark blood images). In the patients who had both images, the greater of the two scar quantifications was used. Scar extent was quantified as the percentage of the amount of myocardium in those segments.

Left bundle area pacing

LBP was carried out in the electrophysiology laboratory (Hammersmith Hospital, UK) electrocardiograms were recorded using the electrophysiology system (Boston Scientific, Natick, Massachusetts).
The C315His fixed curve sheath (Medtronic, Minneapolis, MN) was used to position the SelectSecure 3830 lead (Medtronic, Minneapolis, MN) onto the bundle of His. We mapped the His bundle and stored fluoroscopic image in the right anterior oblique view, and this was used as a roadmap. The lead was then advanced to basal right ventricular septum about 2cm below the His bundle. From that position it was deployed deep into the ventricular septum (Figure 1). A single operator (ZW) assessed the deliverability of the left bundle lead and the electrical response while blinded to the results of the cardiac MRI.
During lead advancement, we monitored impedance and time to peak R wave (RWPT) (time from the stimulus to peak R wave in lead V5 or V6) with unipolar pacing. Left bundle capture was confirmed by a right bundle branch block morphology with a terminal R wave in lead V1 and any of: 1- RWPT time < 90ms, 2-The presence of left bundle branch potential (15-30ms before the QRS onset), 3-Transistion from non-selective to selective left bundle capture (Figure 2) or non-selective left bundle capture to myocardial only capture with change in pacing output or with programmed stimulation.10
Lead depth was assessed by placing the end of the sheath onto the septum and measuring the distance to the tip of the lead or by contrast injection. Furthermore, in cases where the lead was not advanced or left bundle was not captured, electrocardiographic monitoring showed left bundle branch morphology and the RWPT in V5/V6 remained > 80ms.
We defined two modes of failure. Unsuccessful lead delivery was inability to advance the lead deep enough into the septum to reach the left bundle area. Unsuccessful left bundle capture was failure to stimulate the left bundle despite successful lead delivery.
Paced QRS duration during LBAP was used to assess electrical response. The QRS was optimized by allowing fusion with intrinsic right bundle conduction if possible.

Statistical analysis

Continuous variables were expressed as mean and standard deviation if normally distributed and mean (IQR) otherwise. Correlation were assessed by the Spearman rank correlation coefficient. Comparisons between groups were performed using the Mann Whitney U test. A p value <0.05 was considered statistically significant. Statistical analysis was conducted in RStudio using the tidyverse package.