Advertisement

Biventricular endocardial pacing and left bundle branch area pacing for cardiac resynchronization: Mechanistic insights from electrocardiographic imaging, acute hemodynamic response, and magnetic resonance imaging

Open AccessPublished:October 27, 2022DOI:https://doi.org/10.1016/j.hrthm.2022.10.019

      Background

      Biventricular endocardial pacing (BiV-endo) has demonstrated superior cardiac resynchronization compared to conventional biventricular epicardial pacing (BiV-epi). Left bundle branch area pacing (LBBAP) may also achieve effective cardiac resynchronization therapy (CRT).

      Objective

      The purpose of this study was to compare the acute electrical and hemodynamic effects of BiV-epi, BiV-endo, and LBBAP delivered from the LV endocardium and to assess how myocardial scar affects response.

      Methods

      Eleven patients with indications for CRT underwent a temporary pacing study with electrocardiographic imaging (ECGi) and hemodynamic assessment. BiV-endo was delivered by stimulation of the left ventricular (LV) lateral wall, and LBBAP was delivered by stimulation of the LV septum, at the site of a Purkinje potential. LV activation time (LVAT-95), LV dyssynchrony index (LVDI), biventricular activation time (BIVAT-90), and biventricular dyssynchrony index (BIVDI) were calculated. Myocardial scar was assessed using magnetic resonance imaging (MRI).

      Results

      The protocol was completed in 10 patients. Compared to BiV-epi (LVAT-95: 79.2 ± 13.1 ms; LVDI: 26.6 ± 3.4 ms) LV resynchronization was superior during BiV-endo (LVAT-95: 48.5 ± 14.9 ms; P = .001; LVDI: 16.6 ± 6.4 ms; P = .002) and LBBAP (LVAT-95: 48.9 ± 12.5 ms; P = .001; LVDI: 15.3 ± 3.4 ms; P = .001). Biventricular resynchronization was similarly superior during BiV-endo and LBBAP vs BiV-epi (BIVAT-90 and BIVDI; P <.05). The rate of acute hemodynamic responders was higher during BiV-endo (90%) and LBBAP (70%) vs BiV-epi (50%). The benefits of LBBAP (but not BiV-endo) on LV resynchronization were attenuated in a subset of 8 patients who underwent MRI.

      Conclusion

      Our findings suggest superior electrical resynchronization and a higher proportion of acute hemodynamic responders during BiV-endo and LBBAP compared to BiV-epi. Electrical resynchronization was similar between BiV-endo and LBBAP; however, septal scar seemed to attenuate response to LBBAP.

      Graphical abstract

      Keywords

      Introduction

      Cardiac resynchronization therapy (CRT) is an effective treatment for heart failure with electrical dyssynchrony. Left ventricular (LV) pacing is conventionally achieved epicardially via a lead in a lateral or posterolateral branch of the coronary sinus. However, about one-third of patients do not respond to conventional biventricular pacing (BiV-epi), and LV lead delivery is not achievable in all patients.
      • Sieniewicz B.J.
      • Gould J.
      • Porter B.
      • et al.
      Understanding non-response to cardiac resynchronisation therapy: common problems and potential solutions.
      Biventricular endocardial pacing (BiV-endo), His-bundle pacing (HBP), and left bundle branch area pacing (LBBAP) have emerged as alternatives to BiV-epi for patients who do not respond or fail coronary sinus lead implantation, with evidence these novel techniques may achieve superior resynchronization. BiV-endo is delivered by stimulation of the lateral LV endocardium via either a conventional pacing lead or the leadless WiSE CRT system (EBR Systems, Sunnyvale, CA).
      • Elliott M.K.
      • Mehta V.S.
      • Sidhu B.S.
      • Niederer S.
      • Rinaldi C.A.
      Endocardial left ventricular pacing.
      ,
      • Okabe T.
      • Hummel J.D.
      • Bank A.J.
      • et al.
      Leadless left ventricular stimulation with WiSE-CRT System—initial experience and results from phase I of SOLVE-CRT Study (nonrandomized, roll-in phase).
      Pacing the LV endocardium at the same site as the epicardium has been associated with improved acute hemodynamic response (AHR) and shorter QRS duration.
      • Behar J.M.
      • Jackson T.
      • Hyde E.
      • et al.
      Optimized left ventricular endocardial stimulation is superior to optimized epicardial stimulation in ischemic patients with poor response to cardiac resynchronization therapy.
      HBP may achieve superior electrical resynchronization and AHR compared to BiV-epi
      • Arnold A.D.
      • Shun-Shin M.J.
      • Keene D.
      • et al.
      His resynchronization versus biventricular pacing in patients with heart failure and left bundle branch block.
      ; however, delivery in clinical practice is limited by high pacing thresholds and a failure to correct left bundle branch block (LBBB) in patients with distal conduction disease.
      • Upadhyay G.A.
      • Cherian T.
      • Shatz D.Y.
      • et al.
      Intracardiac delineation of septal conduction in left bundle-branch block patterns: mechanistic evidence of left intrahisian block circumvented by His bundle pacing.
      ,
      • Teigeler T.
      • Kolominsky J.
      • Vo C.
      • et al.
      Intermediate-term performance and safety of His-bundle pacing leads: a single-center experience.
      LBBAP has potential advantages over HBP, including easier targeting of the conduction system, improved sensing, and lower capture thresholds.
      • Cano Ó.
      • Vijayaraman P.
      Left bundle branch area pacing: implant technique, definitions, outcomes, and complications.
      LBBAP typically is delivered from the right ventricle (RV) using lead-based techniques,
      • Arnold A.D.
      • Whinnett Z.I.
      • Vijayaraman P.
      His-Purkinje conduction system pacing: state of the art in 2020.
      but delivery from the LV endocardium using leadless technology has recently been described,
      • Elliott M.K.
      • Vergara P.
      • Wijesuriya N.
      • et al.
      Feasibility of leadless left ventricular septal pacing with the WiSE-CRT system to target the left bundle branch area: a porcine model and multicenter patient experience.
      and temporary LV septal pacing has been shown to achieve superior LV electrical resynchronization compared to BiV-epi.
      • Salden F.C.W.M.
      • Luermans J.G.L.M.
      • Westra S.W.
      • et al.
      Short-term hemodynamic and electrophysiological effects of cardiac resynchronization by left ventricular septal pacing.
      To date no study has compared electrical resynchronization and AHR during BiV-endo and LBBAP in a single patient cohort. The purpose of this study was to compare temporary BiV-epi, BiV-endo, and LBBAP using electrocardiographic imaging (ECGi) and AHR, and to assess their relationship to myocardial scar identified by cardiac magnetic resonance imaging (cMRI).

      Methods

      Study population

      Eleven patients who met criteria for CRT implant or upgrade were included (LV ejection fraction ≤35%, QRS duration ≥130 ms, symptomatic heart failure despite optimal medical therapy).
      • Brignole M.
      • Auricchio A.
      • Baron-Esquivias G.
      • et al.
      2013 ESC guidelines on cardiac pacing and cardiac resynchronization therapy.
      Both patients awaiting CRT and those with existing CRT devices at the time of enrollment were included. For patients with functioning CRT devices at the time of enrollment, CRT response was defined as an improvement in LV ejection fraction >5% or a reduction of LV end-systolic volume >15% on transthoracic echocardiogram at least 6 months after implant. Patients with atrial fibrillation and poor rate control, significant peripheral vascular disease, prosthetic aortic or tricuspid valve, LV thrombus, or a contraindication to anticoagulation were excluded. All patients provided written informed consent, and the study was conducted in accordance with the Declaration of Helsinki and approved by local research ethics committee (13/LO/1475). The results of the research study did not influence the subsequent care of the patients, and those without a functional CRT device at the time of recruitment underwent subsequent implantation of a conventional CRT system after the research was complete, according to clinical guidelines.
      • Brignole M.
      • Auricchio A.
      • Baron-Esquivias G.
      • et al.
      2013 ESC guidelines on cardiac pacing and cardiac resynchronization therapy.

      cMRI

      Patients underwent cMRI with late gadolinium enhancement, using a previously described protocol
      • Gould J.
      • Porter B.
      • Claridge S.
      • et al.
      Mean entropy predicts implantable cardioverter-defibrillator therapy using cardiac magnetic resonance texture analysis of scar heterogeneity.
      on a 1.5-T or 3-T scanner, before the temporary pacing procedure. Patients who had an existing non-MRI conditional device at the time of enrollment did not undergo cMRI. The presence and location of scar were determined by an independent cMRI expert who was blinded to the results of the temporary pacing procedure.

      Electrophysiological procedure

      A temporary pacing procedure was performed with patients under local anesthesia and sedation immediately before CRT implant or upgrade, or as a standalone procedure if the patient had a CRT device already in situ. A quadripolar catheter (5F Supreme™ 120 cm, St Jude Medical, St Paul, MN) was placed in the right atrium via femoral venous access. A decapolar catheter (6F Livewire 115 cm, St Jude Medical) was placed in the RV via femoral venous access for RV pacing. These temporary catheters were placed in close proximity to existing device right atrial and RV leads if already in situ. If the patient did not have a functional LV lead, a quadripolar catheter was placed targeting a posterolateral or lateral branch of the coronary sinus. A PressureWire X (Abbott, St Paul, MN) was placed in the LV cavity via a femoral arterial puncture using a 5Fr MPA catheter with a retrograde aortic approach. A decapolar catheter was inserted into the LV via the same arterial sheath and used for BiV-endo pacing and LBBAP. The temporary catheters were connected to a device programmer to allow dual-chamber and biventricular pacing with atrioventricular synchrony. BiV-epi was performed using the patient’s existing device, if present, or via the temporary catheters. BiV-endo was performed by pacing the LV endocardium at the basal or mid-lateral wall in conjunction with RV pacing. LBBAP was performed from the LV endocardium by positioning the decapolar catheter along the LV septum and identifying a Purkinje potential and pacing at this location. Example fluoroscopic images of catheter locations are shown in Figure 1. All pacing was performed at 10 bpm above intrinsic rhythm and kept at a fixed rate for all pacing modalities within each patient. BiV-epi and BiV-endo were performed with a ventriculoventricular (VV) delay of 0 ms. For patients in sinus rhythm, all pacing was performed with an atrioventricular (AV) delay of 100 ms. Patients were heparinized to achieve an activated clotting time >300 ms.
      Figure thumbnail gr1
      Figure 1Intraprocedural fluoroscopy demonstrating catheter positions during biventricular endocardial pacing (A) and left bundle branch area pacing (B). CS = coronary sinus; ICD = implantable cardioverter-defibrillator; LV = left ventricle; RA = right atrium; RV = right ventricle.

      ECGi

      Before the electrophysiological procedure, patients were fitted with a 252-electrode CardioInsight™ sensor array vest (Medtronic Inc., Minneapolis, MN) and underwent a computed tomographic scan to obtain electrode positions and cardiac anatomy as previously described.
      • Ramanathan C.
      • Ghanem R.N.
      • Jia P.
      • Ryu K.
      • Rudy Y.
      Noninvasive electrocardiographic imaging for cardiac electrophysiology and arrhythmia.
      Surface electrograms were continuously recorded throughout the electrophysiological procedure. The body surface potentials from individually selected beats were combined with the segmented cardiac anatomy from computed tomography to create reconstructed unipolar electrograms using the CardioInsight Workstation (Medtronic).
      • Ramanathan C.
      • Ghanem R.N.
      • Jia P.
      • Ryu K.
      • Rudy Y.
      Noninvasive electrocardiographic imaging for cardiac electrophysiology and arrhythmia.

      Calculation of dyssynchrony metrics

      Custom in-house code was used to create activation maps and calculate dyssynchrony metrics. Activation time for each point was defined as the maximal negative derivative (dV/dtmax) of the QRS segment of the unipolar electrogram. Poor-quality electrograms (amplitude <0.5 mV) and significant outliers (<1st quartile – [1.5 × interquartile range (IQR)]; or >3rd quartile + [1.5 × IQR]) were automatically removed. The outflow tracts and periannular regions were excluded. Calculated metrics from 3 beats were averaged for each pacing configuration. The following dyssynchrony metrics were calculated: (1) left ventricular activation time (LVAT-95)—time taken for activation of 95% of LV; (2) left ventricular dyssynchrony index (LVDI)—SD of LVATs; (3) biventricular activation time (BIVAT-90)—time taken for activation of 90% of the ventricles; and (4) biventricular dyssynchrony index (BIVDI)—SD of all activation times

      Acute hemodynamic measurements

      The hemodynamic study was performed using a previously described protocol.
      • Sohal M.
      • Hamid S.
      • Perego G.
      • et al.
      A multicenter prospective randomized controlled trial of cardiac resynchronization therapy guided by invasive dP/dt.
      ,
      • Elliott M.K.
      • Mehta V.S.
      • Rinaldi C.A.
      Pacing optimized by left ventricular dP/dt max.
      LV dP/dtmax measurements were recorded using CoroFlow (Coroventis, Uppsala, Sweden). For each pacing configuration, mean dP/dtmax was calculated over a 10-second period after 20 seconds of pacing. Baseline measurements (atrial pacing if in sinus rhythm and RV pacing if in atrial fibrillation or complete heart block) were recorded before and after each CRT pacing method, and AHR was expressed as percentage increase from the mean baseline measurement. Ectopic beats and the following 2 beats were excluded. Acute responders were defined as having AHR >10%.
      • Sohal M.
      • Hamid S.
      • Perego G.
      • et al.
      A multicenter prospective randomized controlled trial of cardiac resynchronization therapy guided by invasive dP/dt.
      • Elliott M.K.
      • Mehta V.S.
      • Rinaldi C.A.
      Pacing optimized by left ventricular dP/dt max.
      • Duckett S.G.
      • Ginks M.
      • Shetty A.K.
      • et al.
      Invasive acute hemodynamic response to guide left ventricular lead implantation predicts chronic remodeling in patients undergoing cardiac resynchronization therapy.

      Statistical analysis

      Continuous variables were tested for normality using the Shapiro-Wilk test and expressed as mean ± SD if normally distributed. Discrete variables are expressed as count (%). Dyssynchrony metrics and AHR for different pacing modalities were compared using a repeated measures analysis of variance. If a significant difference between groups was identified (P <.05), then pairwise comparisons between different CRT modalities were performed using the Tukey test for multiple comparisons. Relative change in LVAT-95 and LVDI between patients with and without septal scar were compared with an independent samples t test. Statistical analysis and creation of graphs were performed using the Stata Statistical Software Package Release 17 (StataCorp LLC, College Station, TX). P <.05 was considered significant.

      Results

      Patient characteristics

      Eleven patients were recruited. The procedure was abandoned because of difficult cardiac anatomy in 1 patient; therefore, data from 10 patients were included for analysis. Patient characteristics are summarized in Table 1. Four patients were naïve to CRT at the time of the research pacing protocol. Three patients had an existing CRT device but were not receiving CRT due to LV lead displacement (n = 2) or rapidly conducted atrial fibrillation (n = 1). In the latter case, atrioventricular nodal ablation was performed before the research protocol was commenced. Three patients had a functional CRT device at the time of the research procedure, and all had echocardiographic evidence of response. Mean age was 70.7 ± 5.6 years; 7 patients were male; and 3 were in atrial fibrillation. Three patients had ischemic etiology of heart failure. Mean LV ejection fraction before CRT implant was 28.7% ± 6.0%. Mean QRS duration at baseline was 164.8 ± 28.1 ms. Five patients had LBBB, 4 had an RV-paced rhythm, and 1 had right bundle branch block with QRS duration >150 ms. Eight patients underwent cMRI, and 5 had myocardial scar on late gadolinium imaging. BiV-epi, BiV-endo, and LBBAP were achieved in all 10 patients. One procedural complication occurred (femoral arterial pseudoaneurysm not requiring surgical intervention). Example 12-lead ECGs and activation maps during the different pacing modalities are shown in Figures 2 and 3, respectively.
      Table 1Baseline patient demographics
      Patient no.SexAge (y)LVEF (%)EtiologyRhythmQRS morphologyQRSdDevice at procedureCRT responderScar on cMRI
      1M6230NICMSRLBBB170CRT-DYesNone
      2M6319NICMSRRV-paced131PPMNASeptal midwall enhancement
      3M7233NICMSRLBBB184CRT-DYesNone
      4F7328NICMSRLBBB142NoneNASeptal midwall enhancement
      5F7434NICMSRRV-paced172CRT-P
      With nonfunctional LV lead.
      NANo cMRI
      6M8134ICMAFRV-paced159PPMNANo cMRI
      7F7325ICMSRLBBB138CRT-D
      With nonfunctional LV lead.
      NASubendocardial: Apical inferior and septum; true septum
      8M7127ICMAFRV-paced192CRT-D
      Not receiving CRT due to rapidly conducted AF; underwent atrioventricular nodal ablation before the research protocol.
      NASubendocardial: Basal inferolateral and anterolateral
      9M6620NICMAFRBBB153NoneNASeptal midwall enhancement
      10M7233NICMSRLBBB208CRT-DYesNone
      AF = atrial fibrillation; cMRI = cardiac magnetic resonance imaging; CRT = cardiac resynchronization therapy; CRT-D = cardiac resynchronization therapy–defibrillator; CRT-/P = cardiac resynchronization therapy–pacemaker; ICM = ischemic cardiomyopathy; LBBB = left bundle branch block; LVEF = left ventricular ejection fraction; NA = XXXXXX; NICM = nonischemic cardiomyopathy; QRSd = QRS duration; PPM = pacemaker; RBBB = right bundle branch block; RV = right ventricle; SR = sinus rhythm.
      With nonfunctional LV lead.
      Not receiving CRT due to rapidly conducted AF; underwent atrioventricular nodal ablation before the research protocol.
      Figure thumbnail gr2
      Figure 2Example 12-lead electrocardiograms recorded during baseline left bundle branch block (LBBB), conventional biventricular pacing epicardial (BiV-epi), biventricular endocardial pacing (BiV-endo), and left bundle branch area pacing (LBBAP). QRSd = QRS duration.
      Figure thumbnail gr3
      Figure 3Example activation maps from patient 1 during different pacing configurations. The septum is indicated by the left anterior descending artery (depicted in gray). Abbreviations as in and .

      LV activation metrics

      During baseline rhythm, mean LVAT-95 was 105.2 ± 22.8 ms. Compared to baseline, LVAT-95 was significantly shorter during BiV-epi (79.2 ± 13.1 ms; 95% confidence interval [CI] –45.7 to –6.3; P = .006) with additional shortening achieved with both BiV-endo (48.5 ± 14.9 ms; 95% CI –76.3 to –37.0; P <.001) and LBBAP (48.9 ± 12.5 ms; 95% CI –75.9 to –36.5; P <.001) (Figure 4A). LVAT-95 was significantly shorter during BiV-endo (95% CI –50.3 to –11.0; P = .001) and LBBAP (95% CI –49.9 to –10.5; P = .001) than during BiV-epi, in keeping with more rapid LV activation. At baseline, mean LVDI was 35.5 ± 8.0 ms and was significantly lower during BiV-epi (26.6 ± 3.4 ms; 95% CI –15.7 to –2.1; P = .006), BiV-endo (16.6 ± 6.4 ms; 95% CI –25.8 to –12.2 ms; P <.001), and LBBAP (15.3 ± 3.4 ms; 95% CI –27.0 to –13.4 ms; P <.001) (Figure 4B). LVDI was significantly lower during BiV-endo (95% CI –16.9 to –3.3; P = .002) and LBBAP (95% CI –18.1 to –4.5; P <.001) than during BiV-epi, suggesting reduced LV dyssynchrony. LVAT-95 and LVDI during BiV-endo and LBBAP were comparable (P = .961 and P = 1, respectively).
      Figure thumbnail gr4
      Figure 4Comparison of electrical resynchronization metrics. A: Left ventricular activation time (LVAT-95). B: Left ventricular dyssynchrony index (LVDI). C: Biventricular activation time (BIVAT-90). D: Biventricular dyssynchrony index (BIVDI). ∗P <.01 vs baseline; ˆP <.05 vs BiV-epi. Abbreviations as in .

      Biventricular activation metrics

      Mean BIVAT-90 at baseline was 118.5 ± 20.8 ms and was shorter during BiV-epi (79.1 ± 11.5 ms; 95% CI –59.2 to –19.4; P <.001), BiV-endo (57.2 ± 16.3 ms; 95% CI –81.2 to –41.4; P <.001), and LBBAP (59.2 ± 16.3 ms; 95% CI –79.2 to –39.4; P <.001) (Figure 4C). BIVAT-90 was lower during BiV-endo (95% CI –41.9 to –2.0; P = .026) and LBBAP (95% CI –39.9 to –0.002; P = .050) compared to BiV-epi, indicating more rapid biventricular activation. Mean BIVDI at baseline was 40.1 ± 7.0 ms and was lower during BiV-epi (28.2 ± 3.7 ms; 95% CI –18.9 to –5.0 ms; P <.001), BiV-endo (19.8 ± 6.3 ms; 95% CI –27.3 to –13.4; P <.001), and LBBAP (20.7 ± 5.5 ms; 95% CI –26.4 to –12.5; P <.001) (Figure 4D). BIVDI was lower during BiV-endo (95% CI –15.4 to –1.4; P = .013) and LBBAP (95% CI –14.5 to –0.5; P = .030) compared to BiV-epi, suggesting reduced biventricular dyssynchrony. BIVAT-90 and BIVDI during BiV-endo and LBBAP were comparable (P = .993 and P = .985, respectively).

      AHR

      Mean AHR during BiV-epi was 8.0% ± 3.6%. AHR was higher during BiV-endo (16.2% ± 8.8%; 95% CI –0.7 to 16.3; P = .08) and LBBAP (13.7% ± 7.1%; 95% CI –3.2 to 13.8; P = .28), although these improvements did not meet statistical significance (Figure 5). Using a cutoff >10% AHR from baseline, 50% of patients were acute responders during BiV-epi. In contrast, 90% of patients were acute responders during BiV-endo and 70% during LBBAP.
      Figure thumbnail gr5
      Figure 5Acute hemodynamic response. Abbreviations as in .

      Effect of myocardial scar on resynchronization

      cMRI data were available for 8 of 10 patients. Two patients did not undergo cMRI because of existing nonconditional cardiac devices at enrollment. Four of 8 patients (50%) had evidence of midwall or subendocardial scar in the septum. During LBBAP, the change from baseline in LVAT-95 was lower in patients with septal scar than in those without (–28.1% ± 22.4% vs –66.6% ± 5.0%; 95% CI –68.5 to –8.5; P = .02) (Figure 6A). Similarly, the change in LVDI was lower in patients with septal scar (–33.0% ± 23.4% vs –66.7% ± 6.7%; 95% CI –63.5 to –3.9; P = .03) (Figure 6B). During BiV-endo, the change from baseline in LVAT-95 and LVDI between patients with and those without septal scar was not significant (95% CI –38.9 to 4.8; P = .10; and 95% CI –32.8 to 24.2; P = .72, respectively) ()Figures 6C and 6D). Example activation maps during LBBAP and BiV-endo are shown for a patient with and a patient without septal scar in Figure 7. In the 8 patients with cMRI data, 75% of patients with septal scar demonstrated AHR >10% with LBBAP, whereas 100% of patients without scar had a significant response. There was no significant difference in absolute AHR during LBBAP between patients with and those without septal scar (17.7% ± 8.4% vs 13.9% ± 3.8%; 95% CI –7.5 to 15.1; P = .44). There was only 1 patient with lateral wall scar in whom AHR >10% was achieved with BiV-endo (but not BiV-epi).
      Figure thumbnail gr6
      Figure 6Change in left ventricular resynchronization during left bundle branch pacing (LBBAP) and biventricular endocardial pacing (BiV-endo) according to the presence of septal scar in a subgroup of 8 patients who underwent magnetic resonance imaging. Abbreviations as in .
      Figure thumbnail gr7
      Figure 7Example activation maps during left bundle branch area pacing (LBBAP) and biventricular endocardial pacing (BiV-endo) for a patient with septal scar (A) and without septal scar (B). Left: Magnetic resonance images with late gadolinium enhancement (short-axis, midventricular slice) with midwall fibrosis marked with an arrow. Right: Left ventricular activation maps and left ventricular activation times (LVAT-95) during LBBAP and BiV-endo.

      Discussion

      This is the first study to directly compare the electrical and hemodynamic effects of BiV-epi, BiV-endo, and LBBAP in a single patient cohort while defining underlying myocardial substrate on cMRI.
      The main findings of this study are as follows. (1) All methods of CRT delivery achieved significant reductions in LVAT-95, LVDI, BIVAT-90, and BIVDI compared to baseline rhythm. (2) BiV-endo and LBBAP produced superior LV resynchronization (LVAT-95 and LVDI) and biventricular resynchronization (BIVAT-90 and BIVDI) compared to BiV-epi. (3) Electrical resynchronization was similar with BiV-endo and LBBAP. (4) The proportion of acute hemodynamic responders (AHR >10%) was higher during BiV-endo and LBBAP compared to BiV-epi. (5) The electrical benefits of LBBAP (but not BiV-endo) were attenuated by the presence of septal scar in a subset of 8 patients who underwent cMRI.

      BiV-endo

      BiV-endo usually is delivered by stimulating the LV lateral wall with an endocardial lead or via leadless stimulation using the WiSE-CRT system.
      • Elliott M.K.
      • Mehta V.S.
      • Sidhu B.S.
      • Niederer S.
      • Rinaldi C.A.
      Endocardial left ventricular pacing.
      ,
      • Okabe T.
      • Hummel J.D.
      • Bank A.J.
      • et al.
      Leadless left ventricular stimulation with WiSE-CRT System—initial experience and results from phase I of SOLVE-CRT Study (nonrandomized, roll-in phase).
      This requires a coimplant in the RV to trigger LV pacing and in combination with a Micra leadless pacemaker (Medtronic) can allow completely leadless pacing.
      • Carabelli A.
      • Jabeur M.
      • Jacon P.
      • et al.
      European experience with a first totally leadless cardiac resynchronization therapy pacemaker system.
      There is evidence that pacing the LV endocardium achieves superior resynchronization compared to epicardial pacing. Unlike BiV-epi, BiV-endo allows LV pacing at any site, unrestricted by coronary sinus anatomy. This allows targeting of the latest site of activation and avoidance of scar.
      • Elliott M.K.
      • Mehta V.S.
      • Sidhu B.S.
      • Niederer S.
      • Rinaldi C.A.
      Endocardial left ventricular pacing.
      Behar at al
      • Behar J.M.
      • Jackson T.
      • Hyde E.
      • et al.
      Optimized left ventricular endocardial stimulation is superior to optimized epicardial stimulation in ischemic patients with poor response to cardiac resynchronization therapy.
      compared temporary LV endocardial and epicardial pacing at matching locations in 8 patients. BiV-endo was associated with a higher AHR (15.2% ± 10.7% vs 7.6% ± 6.3%; P = .014) and a shorter paced QRS duration (137 ± 22 ms vs 166 ± 30 ms; P <.001) compared to BiV-epi. The more rapid LV activation during LV endocardial pacing may be explained by access to fast conducting endocardial tissue or the distal Purkinje network, and this is supported by previous modeling studies.
      • Hyde E.R.
      • Behar J.M.
      • Claridge S.
      • et al.
      Beneficial effect on cardiac resynchronization from left ventricular endocardial pacing is mediated by early access to high conduction velocity tissue: electrophysiological simulation study.
      However, additional improvements in AHR are achievable with BiV-endo when pacing in the optimal LV location (ie, not the site of the coronary sinus lead),
      • Shetty A.K.
      • Sohal M.
      • Chen Z.
      • et al.
      A comparison of left ventricular endocardial, multisite, and multipolar epicardial cardiac resynchronization: an acute haemodynamic and electroanatomical study.
      so the benefits over BiV-epi are likely due to a combination of optimized pacing location and access to fast conducting endocardial tissue. Although the optimal endocardial pacing location likely varies between patients, in a study in which imaging or electroanatomic guidance was used to select the optimal location for the WiSE CRT electrode, the most common sites were the basal or mid-lateral segments.
      • Sieniewicz B.J.
      • Behar J.M.
      • Gould J.
      • et al.
      Guidance for optimal site selection of a leadless left ventricular endocardial electrode improves acute hemodynamic response and chronic remodeling.
      In the current study, we found that BiV-endo (with empirical pacing of the mid or lateral LV wall) achieved superior reductions in both LV and biventricular activation metrics compared to BiV-epi. Although no randomized studies comparing clinical outcomes between BiV-endo and BiV-epi have been performed, a previous study of WiSE CRTin patients who had not responded to BiV-epi showed echocardiographic improvement in 66.7%, supporting superior resynchronization with this technique.
      • Sidhu B.S.
      • Porter B.
      • Gould J.
      • et al.
      Leadless left ventricular endocardial pacing in nonresponders to conventional cardiac resynchronization therapy.

      Conduction system pacing and LBBAP delivered from the LV endocardium

      Conduction system pacing is a new paradigm in cardiac pacing and CRT. HBP is an attractive concept that can achieve physiological biventricular activation with beneficial effects on LV activation metrics and hemodynamics.
      • Arnold A.D.
      • Shun-Shin M.J.
      • Keene D.
      • et al.
      His resynchronization versus biventricular pacing in patients with heart failure and left bundle branch block.
      However, clinical delivery of HBP for CRT can be challenging, with procedural success rates of about 80%,
      • Qi J.
      • Jia X.
      • Wang Z.
      His bundle pacing for cardiac resynchronization therapy: a systematic literature review and meta-analysis.
      and is associated with high pacing thresholds,
      • Yuan Z.
      • Cheng L.
      • Wu Y.
      Meta-analysis comparing safety and efficacy of left bundle branch area pacing versus His bundle pacing.
      which can lead to loss of His capture on follow-up in up to 20% of cases.
      • Teigeler T.
      • Kolominsky J.
      • Vo C.
      • et al.
      Intermediate-term performance and safety of His-bundle pacing leads: a single-center experience.
      In randomized studies comparing HBP vs BiV-epi, significant crossover between arms led to a lack of improvement in QRS duration or LV ejection fraction on intention-to-treat analysis.
      • Upadhyay G.A.
      • Vijayaraman P.
      • Nayak H.M.
      • et al.
      His corrective pacing or biventricular pacing for cardiac resynchronization in heart failure.
      ,
      • Vinther M.
      • Risum N.
      • Svendsen J.H.
      • Møgelvang R.
      • Philbert B.T.
      A randomized trial of His pacing versus biventricular pacing in symptomatic HF patients with left bundle branch block (His-alternative).
      LBBAP has emerged as an alternative to HBP and currently is delivered using a right-sided approach with the pacing lead screwed into the interventricular septum.
      • Arnold A.D.
      • Whinnett Z.I.
      • Vijayaraman P.
      His-Purkinje conduction system pacing: state of the art in 2020.
      LBBAP can reverse LBBB at lower pacing thresholds and with improved sensing compared to HBP.
      • Yuan Z.
      • Cheng L.
      • Wu Y.
      Meta-analysis comparing safety and efficacy of left bundle branch area pacing versus His bundle pacing.
      Temporary delivery of LBBAP in an acute pacing protocol is not feasible from a right-sided approach because the conduction system tissue lies superficially under the LV septum.
      • Elliott M.K.
      • Vergara P.
      • Wijesuriya N.
      • et al.
      Feasibility of leadless left ventricular septal pacing with the WiSE-CRT system to target the left bundle branch area: a porcine model and multicenter patient experience.
      The effect of LBBAP delivered acutely from the LV was studied by Salden et al,
      • Salden F.C.W.M.
      • Luermans J.G.L.M.
      • Westra S.W.
      • et al.
      Short-term hemodynamic and electrophysiological effects of cardiac resynchronization by left ventricular septal pacing.
      who performed a temporary pacing procedure in 26 patients undergoing CRT implant using an ECG belt and acute hemodynamic study. LV septal pacing was performed by empirical pacing of the septum, without electrical mapping to target a Purkinje potential. Compared to BiV-epi, LV septal pacing was associated with significant reductions in QRS area (93 ± 26 μVs vs 73 ± 22 μVs; P <.05) and SD of activation times (31 ± 7 ms vs 26 ± 7 ms; P <.05), which is in keeping with the superior electrical resynchronization found with LBBAP in the current study. Salden et al
      • Salden F.C.W.M.
      • Luermans J.G.L.M.
      • Westra S.W.
      • et al.
      Short-term hemodynamic and electrophysiological effects of cardiac resynchronization by left ventricular septal pacing.
      did not find significant improvements in AHR with LBBAP over BiV-epi.
      The current study is unique in allowing electrical and hemodynamic comparisons of BiV-endo (delivered from the lateral wall) and LBBAP (delivered from the LV septum). In a previous study of 12 patients undergoing BiV-endo, AHRs during temporary BiV-endo (with lateral wall pacing), BiV-endo (with LV septal pacing), and BiV-epi were compared, with no significant differences found between modalities.
      • Rademakers L.M.
      • van Hunnik A.
      • Kuiper M.
      • et al.
      A possible role for pacing the left ventricular septum in cardiac resynchronization therapy.
      In contrast to the current study, LV septal pacing was performed along with RV pacing (without targeting of a Purkinje potential), and no assessments of electrical resynchronization were made. In the current study, we found superior LV and biventricular electrical resynchronization with BiV-endo and LBBAP compared to conventional BiV-epi. Furthermore, a higher proportion of patients were acute hemodynamic responders during BiV-endo and LBBAP. The electrical and hemodynamic performance of BiV-endo and LBBAP were very similar. AV optimization was not performed in our cohort due to a high proportion of patients with AV block or atrial fibrillation (50%); however, there is evidence that such performance may bring further reductions in biventricular resynchronization during LBBAP.
      • Strocchi M.
      • Lee A.W.C.
      • Neic A.
      • et al.
      His-bundle and left bundle pacing with optimized atrioventricular delay achieve superior electrical synchrony over endocardial and epicardial pacing in left bundle branch block patients.

      Effect of myocardial scar on LBBAP

      For the first time, we have demonstrated an attenuation of LV resynchronization with LBBAP when septal scar was present on cMRI. Whereas the degree of basal septal scar on cMRI has previously been associated with failure to deliver a lead into the septum for LBBAP,
      • Ali N.
      • Arnold A.
      • Miyazawa A.A.
      • et al.
      Po-673-01 Septal scar predicts failure of lead advancement to the left bundle area but not the ability to stimulate the left bundle.
      effects on LV activation have not previously been reported. Although the trends (albeit not statistically significant) of AHR during matched comparisons of CRT modalities within the same cohort generally followed the electrical resynchronization findings (Figure 5), this was not the case when comparing AHR between patients with and those without septal scar during LBBAP. This may be because comparisons between unmatched groups can be skewed by overall AHR during pacing for particular patients, which in turn may be affected by baseline LV systolic function and pre-existing LV reverse remodeling at the time of the research protocol. The proportion of patients with AHR >10% during LBBAP may have more clinical relevance and was found in all patients without septal scar and in only 3 of 4 patients with scar. Although our findings are limited by small patient numbers, this is an important hypothesis-generating finding that merits further investigation.

      Clinical relevance

      Our results confirm the superior electrical and hemodynamic effects of BiV-endo (with lateral wall stimulation) over BiV-epi. Furthermore, we demonstrated that LBBAP delivered from the LV septum achieves comparable electrical resynchronization to BiV-endo. Although the comparative effects of LBBAP delivered from a right-sided or left-sided approach are unclear, the findings from the current study support the growing body of evidence from observational data that LBBAP may provide superior cardiac resynchronization compared to BiV-epi.
      • Hua J.
      • Wang C.
      • Kong Q.
      • et al.
      Comparative effects of left bundle branch area pacing, His bundle pacing, biventricular pacing in patients requiring cardiac resynchronization therapy: a network meta-analysis.
      ,
      • Vijayaraman P.
      • Zalavadia D.
      • Haseeb A.
      • et al.
      Clinical outcomes of conduction system pacing compared to biventricular pacing in patients requiring cardiac resynchronization therapy.
      Although a lead-based right-sided approach presently is the conventional method to deliver LBBAP, the feasibility of permanent leadless LBBAP from the LV septum has recently been demonstrated in a porcine model and human observational study.
      • Elliott M.K.
      • Vergara P.
      • Wijesuriya N.
      • et al.
      Feasibility of leadless left ventricular septal pacing with the WiSE-CRT system to target the left bundle branch area: a porcine model and multicenter patient experience.
      The WiSE CRTsystem currently requires a coimplant to deliver RV pacing; however, with future modifications, completely leadless LBBAP with single-site pacing may be achievable. Although leadless LV pacing is in its infancy and implantation is associated with significant cost and periprocedural risk,
      • Sieniewicz B.J.
      • Betts T.R.
      • James S.
      • et al.
      Real-world experience of leadless left ventricular endocardial cardiac resynchronization therapy: a multicenter international registry of the WiSE-CRT pacing system.
      such techniques may avoid long-term lead-related complications and may play an increasing role as we move toward a future of leadless pacing.
      The finding of an attenuation of the benefits of LBBAP with septal scar has important clinical relevance. Midwall fibrosis is present on cMRI in up to one-third of patients with dilated cardiomyopathy,
      • Mikami Y.
      • Cornhill A.
      • Heydari B.
      • et al.
      Objective criteria for septal fibrosis in non-ischemic dilated cardiomyopathy: validation for the prediction of future cardiovascular events.
      and subendocardial scar may be present in patients with ischemic heart disease. In such patients, LBBAP may be limited by the effect of septal scar. The current study is hypothesis-generating and requires further investigation in larger studies. If septal scar does attenuate response to LBBAP, this may support a tailored approach to BiV-endo and LBBAP, and the use of imaging techniques such as cMRI to characterize the presence and distribution of scar. In patients with septal scar, BiV-endo with lateral wall stimulation may be more effective and in patients with lateral wall scar LBBAP may be the preferred option.

      Study limitations

      The main limitation of the study is the small sample size; however, this is typical for an invasive mechanistic study. Furthermore, all CRT modalities were performed and compared within the same patient group, thus allowing patients to act as their own controls. The included patients are a heterogeneous group including those with ischemic cardiomyopathy, non-LBBB, and atrial fibrillation. Although this may affect the outcomes in the study, we believe this improves the real-world applicability of the findings. Three patients had functional conventional CRT devices at the time of the research pacing protocol and had echocardiographic evidence of LV reverse remodeling. In these patients, the hemodynamic benefits of CRT may be underestimated compared to CRT-naïve patients. However, in a previous ECGi study assessing electrical remodeling after BiV-epi, there were significant changes in ventricular activation times between CRT-ON and CRT-OFF at 6-month follow-up,
      • Pereira H.
      • Jackson T.A.
      • Claridge S.
      • et al.
      Evidence of reverse electrical remodelling by non-invasive electrocardiographic imaging to assess acute and chronic changes in bulk ventricular activation following cardiac resynchronisation therapy.
      thus supporting the ability to compare the performance of different CRT modalities in these patients.
      The observed effects in the study were acute, and the longer-term effects of different LV endocardial stimulation techniques are unknown. However, the RADI-CRT (Prospective Randomised Controlled Trial of Cardiac Resynchronisation Therapy Guided by Invasive dP/dT) study confirmed that AHR >10% was predictive of reverse LV remodeling at follow-up.
      • Sohal M.
      • Hamid S.
      • Perego G.
      • et al.
      A multicenter prospective randomized controlled trial of cardiac resynchronization therapy guided by invasive dP/dt.
      Stimulation was performed without AV optimization and simultaneous VV stimulation. This may underestimate the beneficial effects of LBBAP pacing if AV delay was optimized.
      • Strocchi M.
      • Lee A.W.C.
      • Neic A.
      • et al.
      His-bundle and left bundle pacing with optimized atrioventricular delay achieve superior electrical synchrony over endocardial and epicardial pacing in left bundle branch block patients.
      The ability to perform repeated manipulations of timings is limited in invasive studies in which multiple different pacing modalities are compared. Likewise, the only commercially available system that can deliver leadless BiV-endo does not allow changes in VV timing; therefore, our findings are consistent with real-world techniques.
      • Okabe T.
      • Hummel J.D.
      • Bank A.J.
      • et al.
      Leadless left ventricular stimulation with WiSE-CRT System—initial experience and results from phase I of SOLVE-CRT Study (nonrandomized, roll-in phase).
      Similarly, multiple pacing locations for BiV-epi and BiV-endo were not performed, and empirical locations were chosen to maintain consistency between patients. Further optimization of BiV-epi or BiV-endo pacing sites may have obtained incremental benefits.

      Conclusion

      This is the first study to directly compare the acute electrical and hemodynamic effects of BiV-epi, BiV-endo, and LBBAP. Our results suggest that LBBAP and BiV-endo are the optimal resynchronization strategies with superior electrical performance and a higher rate of acute hemodynamic responders compared to BiV-epi. The choice of LBBAP vs BiV-endo pacing may relate to the presence of septal scar on cMRI, and this requires further investigation. Randomized studies are required to further determine the superiority of BiV-endo and LBBAP over BiV-epi.

      References

        • Sieniewicz B.J.
        • Gould J.
        • Porter B.
        • et al.
        Understanding non-response to cardiac resynchronisation therapy: common problems and potential solutions.
        Heart Fail Rev. 2019; 24: 41-54
        • Elliott M.K.
        • Mehta V.S.
        • Sidhu B.S.
        • Niederer S.
        • Rinaldi C.A.
        Endocardial left ventricular pacing.
        Herz. 2021; 46: 526-532
        • Okabe T.
        • Hummel J.D.
        • Bank A.J.
        • et al.
        Leadless left ventricular stimulation with WiSE-CRT System—initial experience and results from phase I of SOLVE-CRT Study (nonrandomized, roll-in phase).
        Heart Rhythm. 2022; 19: 22-29
        • Behar J.M.
        • Jackson T.
        • Hyde E.
        • et al.
        Optimized left ventricular endocardial stimulation is superior to optimized epicardial stimulation in ischemic patients with poor response to cardiac resynchronization therapy.
        JACC Clin Electrophysiol. 2016; 2: 799-809
        • Arnold A.D.
        • Shun-Shin M.J.
        • Keene D.
        • et al.
        His resynchronization versus biventricular pacing in patients with heart failure and left bundle branch block.
        J Am Coll Cardiol. 2018; 72: 3112-3122
        • Upadhyay G.A.
        • Cherian T.
        • Shatz D.Y.
        • et al.
        Intracardiac delineation of septal conduction in left bundle-branch block patterns: mechanistic evidence of left intrahisian block circumvented by His bundle pacing.
        Circulation. 2019; 139: 1876-1888
        • Teigeler T.
        • Kolominsky J.
        • Vo C.
        • et al.
        Intermediate-term performance and safety of His-bundle pacing leads: a single-center experience.
        Heart Rhythm. 2021; 18: 743-749
        • Cano Ó.
        • Vijayaraman P.
        Left bundle branch area pacing: implant technique, definitions, outcomes, and complications.
        Curr Cardiol Rep. 2021; 23: 1-9
        • Arnold A.D.
        • Whinnett Z.I.
        • Vijayaraman P.
        His-Purkinje conduction system pacing: state of the art in 2020.
        Arrhythm Electrophysiol Rev. 2020; 9: 136-145
        • Elliott M.K.
        • Vergara P.
        • Wijesuriya N.
        • et al.
        Feasibility of leadless left ventricular septal pacing with the WiSE-CRT system to target the left bundle branch area: a porcine model and multicenter patient experience.
        Heart Rhythm Published online. 2022 Aug 5; (02188-9): S1547-S5271https://doi.org/10.1016/j.hrthm.2022.07.017
        • Salden F.C.W.M.
        • Luermans J.G.L.M.
        • Westra S.W.
        • et al.
        Short-term hemodynamic and electrophysiological effects of cardiac resynchronization by left ventricular septal pacing.
        J Am Coll Cardiol. 2020; 75: 347-359
        • Brignole M.
        • Auricchio A.
        • Baron-Esquivias G.
        • et al.
        2013 ESC guidelines on cardiac pacing and cardiac resynchronization therapy.
        Eur Heart J. 2013; 34: 2281-2329
        • Gould J.
        • Porter B.
        • Claridge S.
        • et al.
        Mean entropy predicts implantable cardioverter-defibrillator therapy using cardiac magnetic resonance texture analysis of scar heterogeneity.
        Heart Rhythm. 2019; 16: 1242-1250
        • Ramanathan C.
        • Ghanem R.N.
        • Jia P.
        • Ryu K.
        • Rudy Y.
        Noninvasive electrocardiographic imaging for cardiac electrophysiology and arrhythmia.
        Nat Med. 2004; 10: 422-428
        • Sohal M.
        • Hamid S.
        • Perego G.
        • et al.
        A multicenter prospective randomized controlled trial of cardiac resynchronization therapy guided by invasive dP/dt.
        Heart Rhythm O2. 2021; 2: 19-27
        • Elliott M.K.
        • Mehta V.S.
        • Rinaldi C.A.
        Pacing optimized by left ventricular dP/dt max.
        Card Electrophysiol Clin. 2022; 14: 223-232
        • Duckett S.G.
        • Ginks M.
        • Shetty A.K.
        • et al.
        Invasive acute hemodynamic response to guide left ventricular lead implantation predicts chronic remodeling in patients undergoing cardiac resynchronization therapy.
        J Am Coll Cardiol. 2011; 58: 1128-1136
        • Carabelli A.
        • Jabeur M.
        • Jacon P.
        • et al.
        European experience with a first totally leadless cardiac resynchronization therapy pacemaker system.
        Europace. 2021; 23: 740-747
        • Hyde E.R.
        • Behar J.M.
        • Claridge S.
        • et al.
        Beneficial effect on cardiac resynchronization from left ventricular endocardial pacing is mediated by early access to high conduction velocity tissue: electrophysiological simulation study.
        Circ Arrhythm Electrophysiol. 2015; 8: 1164-1172
        • Shetty A.K.
        • Sohal M.
        • Chen Z.
        • et al.
        A comparison of left ventricular endocardial, multisite, and multipolar epicardial cardiac resynchronization: an acute haemodynamic and electroanatomical study.
        Europace. 2014; 16: 873-879
        • Sieniewicz B.J.
        • Behar J.M.
        • Gould J.
        • et al.
        Guidance for optimal site selection of a leadless left ventricular endocardial electrode improves acute hemodynamic response and chronic remodeling.
        JACC Clin Electrophysiol. 2018; 4: 860-868
        • Sidhu B.S.
        • Porter B.
        • Gould J.
        • et al.
        Leadless left ventricular endocardial pacing in nonresponders to conventional cardiac resynchronization therapy.
        Pacing Clin Electrophysiol. 2020; 43: 966-973
        • Qi J.
        • Jia X.
        • Wang Z.
        His bundle pacing for cardiac resynchronization therapy: a systematic literature review and meta-analysis.
        J Interven Card Electrophysiol. 2020; 59: 463-470
        • Yuan Z.
        • Cheng L.
        • Wu Y.
        Meta-analysis comparing safety and efficacy of left bundle branch area pacing versus His bundle pacing.
        Am J Cardiol. 2022; 164: 64-72
        • Upadhyay G.A.
        • Vijayaraman P.
        • Nayak H.M.
        • et al.
        His corrective pacing or biventricular pacing for cardiac resynchronization in heart failure.
        J Am Coll Cardiol. 2019; 74: 157-159
        • Vinther M.
        • Risum N.
        • Svendsen J.H.
        • Møgelvang R.
        • Philbert B.T.
        A randomized trial of His pacing versus biventricular pacing in symptomatic HF patients with left bundle branch block (His-alternative).
        JACC Clin Electrophysiol. 2021; 7: 1422-1432
        • Rademakers L.M.
        • van Hunnik A.
        • Kuiper M.
        • et al.
        A possible role for pacing the left ventricular septum in cardiac resynchronization therapy.
        JACC Clin Electrophysiol. 2016; 2: 413-422
        • Strocchi M.
        • Lee A.W.C.
        • Neic A.
        • et al.
        His-bundle and left bundle pacing with optimized atrioventricular delay achieve superior electrical synchrony over endocardial and epicardial pacing in left bundle branch block patients.
        Heart Rhythm. 2020; 17: 1922-1929
        • Ali N.
        • Arnold A.
        • Miyazawa A.A.
        • et al.
        Po-673-01 Septal scar predicts failure of lead advancement to the left bundle area but not the ability to stimulate the left bundle.
        Heart Rhythm. 2022; 19 (Abstract): S331
        • Hua J.
        • Wang C.
        • Kong Q.
        • et al.
        Comparative effects of left bundle branch area pacing, His bundle pacing, biventricular pacing in patients requiring cardiac resynchronization therapy: a network meta-analysis.
        Clin Cardiol. 2022; 45: 214-223
        • Vijayaraman P.
        • Zalavadia D.
        • Haseeb A.
        • et al.
        Clinical outcomes of conduction system pacing compared to biventricular pacing in patients requiring cardiac resynchronization therapy.
        Heart Rhythm. 2022; 19: 1263-1271
        • Sieniewicz B.J.
        • Betts T.R.
        • James S.
        • et al.
        Real-world experience of leadless left ventricular endocardial cardiac resynchronization therapy: a multicenter international registry of the WiSE-CRT pacing system.
        Heart Rhythm. 2020; 17: 1291-1297
        • Mikami Y.
        • Cornhill A.
        • Heydari B.
        • et al.
        Objective criteria for septal fibrosis in non-ischemic dilated cardiomyopathy: validation for the prediction of future cardiovascular events.
        J Cardiovasc Magn Reson. 2016; 18: 1-13
        • Pereira H.
        • Jackson T.A.
        • Claridge S.
        • et al.
        Evidence of reverse electrical remodelling by non-invasive electrocardiographic imaging to assess acute and chronic changes in bulk ventricular activation following cardiac resynchronisation therapy.
        J Electrocardiol. 2020; 58: 96-102