Left Ventricular Lead Placement Guided by Reduction in QRS Area

Authors, Journal, Affiliations, Type, DOI

• Authors: Mohammed Ali Ghossein, Francesco Zanon, Floor Salden, Antonius van Stipdonk, Lina Marcantoni, Elien Engels, Justin Luermans, Sjoerd Westra, Frits Prinzen, Kevin Vernooy
• Journal: Journal of Clinical Medicine 2021;10(24):5935 (published 17 December 2021)
• Affiliations: Cardiovascular Research Institute Maastricht (CARIM), Maastricht University; Santa Maria della Misericordia General Hospital, Rovigo, Italy; Maastricht University Medical Center; Yale New Haven Hospital; Radboud University Medical Center, Nijmegen
• Type: Retrospective pooled analysis of two prospective acute haemodynamic CRT studies
• DOI: 10.3390/jcm10245935

Overview

This study asks whether the reduction in QRS area during biventricular pacing can guide LV lead placement, by comparing it with the acute haemodynamic response (%ΔLVdP/dtmax) across different pacing sites within the same patient. In 52 patients and 188 pacing measurements pooled from two prior studies, lateral and anterolateral LV pacing produced both the largest QRS-area reduction and the largest haemodynamic improvement, while QRS duration did not change across sites. Within patients, ΔQRS area correlated strongly with %ΔLVdP/dtmax (median R 0.76), and the largest QRS-area reduction coincided with the largest haemodynamic gain in the great majority. The authors propose QRS area as an objective, non-invasive intraprocedural tool for choosing the LV pacing site and the active electrode on a quadripolar lead.

Keywords

Cardiac resynchronization therapy; LV lead placement; QRS area; invasive hemodynamic measurements; heart failure

Key Takeaways

Background and rationale

• CRT is indicated for reduced LV systolic function with a wide QRS and LBBB morphology, but the electrical substrate amenable to CRT — delayed activation of the LV lateral wall — is detected better by vectorcardiographic QRS area than by QRS duration or LBBB morphology.
• Baseline QRS area predicts echocardiographic response and is strongly associated with clinical outcomes, including mortality.
• Reduction in QRS area after CRT is also associated with echocardiographic and clinical response — but those data used one measurement per patient and did not test whether the reduction could guide lead positioning.
• Current LV lead positioning is guided by coronary venous anatomy, and several studies found no outcome difference between anatomical locations. A patient-specific approach is proposed instead.

Methods

• 52 patients from two prior prospective studies: 29 from Zanon et al. (Rovigo) and 27 from Salden et al. (Maastricht and Radboud); 4 excluded for unavailable VCG analysis.
• All had a Class Ia or IIa CRT indication and a quadripolar LV lead; RV lead in the apical septum, atrial lead in the right atrial appendage.
• Acute haemodynamics: %ΔLVdP/dtmax measured with a pressure wire in a stable LV position, averaging at least 10 beats at baseline (AAI pacing) and at each pacing site, excluding ventricular extrasystoles.
• ECG processing: 12-lead ECGs converted to orthogonal X-, Y-, Z-leads with the Kors transformation matrix; QRS area calculated as √(Xarea² + Yarea² + Zarea²); ΔQRS area and ΔQRS duration calculated between baseline and biventricular pacing.
• Statistics: within-patient paired comparisons, mixed-effects models with Geisser–Greenhouse correction, Pearson and Spearman correlations.

Baseline characteristics

• Age 70 ± 9 years; male 79%; ischaemic cardiomyopathy 50%; LVEF 29 ± 9%; NYHA II 54%, III 46%.
• Sinus rhythm 71%; QRS duration 169 ± 18 ms; QRS area 108 ± 44 µVs.
• LBBB 67%, non-specific IVCD 12%, RBBB 8%, upgrade from RV pacing 13%.

QRS-area reduction and haemodynamics by anatomical site

In 26 patients with measurements at different anatomical LV lead positions:

• ΔQRS area differed significantly by pacing site (p = 0.001), as did %ΔLVdP/dtmax (p = 0.0007).
• ΔQRS duration did not differ between pacing locations (p = 0.23).
• In the 21 patients with at least two anatomical pacing sites, the largest ΔQRS area came from the lateral position in 10 (48%) and the anterolateral position in 5 (24%); the largest LVdP/dtmax increase came from the lateral position in 11 (52%) and the anterolateral position in 6 (29%).

Proximal vs distal electrode pair

• In 43 patients, pacing from the proximal LV electrode pair produced a significantly larger ΔQRS area than the distal pair (−39 vs −17 µVs, p = 0.0005).
• %ΔLVdP/dtmax was also larger with proximal pacing (18% vs 16%, p = 0.0063) — a smaller difference than for QRS area.
• ΔQRS duration did not differ by electrode pair (p = 0.77).

Relationship between ΔQRS area and %ΔLVdP/dtmax

• Across 188 pacing measurements, Spearman correlation was Rs = 0.47 (p < 0.0001).
• Within patients with at least three pacing sites (n=21), the correlation was strong and varied between patients — median R = 0.76 (IQR 0.35–0.89) — with patient-specific slopes and intercepts.
• In 36/43 patients (84%), the %ΔLVdP/dtmax was within 5% of that patient's maximum when pacing was guided by the largest ΔQRS area.

Interpretation

• ΔQRS area reflects resynchronisation better than ΔQRS duration. QRS area combines both duration and morphology, and its measurement variability is less than half that of QRS duration (~20 ms observer variability for QRS duration).
• Conflicting ΔQRS duration results across PROSPECT and REVERSE are attributed in part to measurement variability and technique.
• The optimal anatomical LV pacing site does not exist as a population constant — the final lead position should be individualised, consistent with the argument of Wouters et al.
• Pacing in or near scar is a specific example of unfavourable placement: it produces worse resynchronisation and worse outcome than pacing viable tissue, and ischaemic cardiomyopathy yields smaller QRS-area reductions.
• Practical potential: QRS area can be calculated from a standard digital 12-lead ECG with software, making it objective and feasible intraprocedurally — though automatic commercial software is not yet available.

Limitations of the document

• Acute haemodynamic surrogate only: %ΔLVdP/dtmax is not a clinical or echocardiographic outcome, and evidence linking acute haemodynamic response to long-term outcome is explicitly described by the authors as conflicting; reverse remodelling involves more than haemodynamics.
• Retrospective pooled analysis of two studies with different protocols (different AV-delay rules, different pressure-wire systems, one using telescopic cannulation of all suitable veins and the other a single posterolateral position).
• Not all pacing sites were measured in every patient, so within-patient comparisons are incomplete and the mixed-effects model compensates for unbalanced sampling.
• Small sample (n=52; n=21 for the within-patient correlation) and observational design with no control arm and no randomisation of lead-placement strategy.
• The correlation is highly variable between patients — median R 0.76 but IQR 0.35–0.89 — so in a substantial minority, QRS-area reduction would not identify the best haemodynamic site.
• QRS area requires VCG transformation software that is not commercially available; the workflow in this study used custom Matlab code and digitised ECGs.
• Industry funding: part of the Salden study was financially supported by Medtronic, although the authors declare no other conflicts.
• Minor numerical inconsistencies in the results — the site-specific counts of largest ΔQRS area are reported as "2 (1%)" and "1 (0.5%)", which do not correspond to 2/21 and 1/21 (~10% and ~5%).
• No outcome data — the claim that QRS-area-guided placement improves CRT response remains untested prospectively.

Key Concepts Mentioned

• concepts/Cardiac-Resynchronization-Therapy — the therapy being optimised
• concepts/Electrical-Resynchronization-Metrics — QRS area as an electrical resynchronisation measure
• entities/Vectorcardiography — Kors transformation and QRS-area calculation
• concepts/LBBB-Criteria — LBBB morphology and its limits for identifying the CRT substrate
• entities/Heart-Failure — the underlying population

Key Entities Mentioned

• Maastricht University / CARIM — coordinating centre and QRS-area methodology
• Santa Maria della Misericordia General Hospital, Rovigo — contributing centre
• Radboud University Medical Center, Nijmegen — contributing centre
• Medtronic — partial funding of the Salden study
• Kors transformation matrix — the VCG derivation method

Wiki Pages Updated

• Created: wiki/sources/qrs-area-crt-jcm-2021.md
• Created: wiki/concepts/Electrical-Resynchronization-Metrics.md
• Updated: wiki/concepts/Cardiac-Resynchronization-Therapy.md
• Updated: wiki/entities/Vectorcardiography.md
• Updated: wiki/entities/CRT.md
• Updated: wiki/sourceindex.md
• Updated: log.md