#cardiac-resynchronization-therapy #ventricular-dyssynchrony #qrs-area #body-surface-mapping #electrocardiography

Electrical Resynchronization Metrics

Definition

Electrical resynchronization metrics are non-invasive measures of how dyssynchronous ventricular activation is, and how completely pacing corrects it. They fall into three families: time-domain QRS measures (QRS duration, paced QRS duration), spatial/vectorcardiographic measures (QRS area), and body-surface mapping measures (standard deviation of activation times, regional activation times and dispersion). Their purpose is to do what QRS duration alone cannot — separate the electrical substrate that responds to CRT from the conduction delay on the surface ECG, and choose or optimise a pacing configuration within the individual patient. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/vcg-fronphysiol-2022, rating: medium)

Key Concepts

QRS duration — the guideline metric and its limits

QRS duration is the only electrical criterion embedded in CRT guidelines, but it is a time-domain summary that ignores the direction and morphology of activation. QRS area combines both duration and morphology and detects delayed LV lateral wall activation better than either QRS duration or LBBB morphology. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/vcg-fronphysiol-2022, rating: medium)
Measurement reproducibility is poor: inter- and intra-observer variability for paced QRS duration is around 20 ms, and the method used to measure it changes its association with CRT response. The variability of QRS area is less than half that of QRS duration. (sources/qrs-area-crt-jcm-2021, rating: medium)
Paced QRS duration does not discriminate pacing sites: across five anatomical LV lead positions and between proximal and distal electrode pairs, ΔQRS duration was unchanged (p = 0.23 and p = 0.77), while ΔQRS area and %ΔLVdP/dtmax both changed significantly. (sources/qrs-area-crt-jcm-2021, rating: medium)
QRS narrowing is evidence of resynchronisation, not of conduction-system capture. Selective LBBP — the capture state with the strongest evidence of direct LBB engagement — produces a longer paced QRS than nonselective LBBP or LV septal pacing, because septal myocardial capture in the nonselective state partially fuses with and neutralises delayed RV activation. (sources/lbbp-vs-lvsp-criteria-jaccep-2021, rating: medium)

QRS area — vectorcardiographic substrate measurement

How it is obtained: the 12-lead ECG is converted to orthogonal X-, Y-, Z-leads with the Kors transformation matrix, and QRS area is calculated as √(Xarea² + Yarea² + Zarea²). It is a single number summarising the three-dimensional QRS loop, and it does not necessarily track QRS duration — a wide QRS can have a small area and vice versa. (sources/qrs-area-crt-jcm-2021, rating: medium)
Baseline QRS area predicts response and outcome better than QRS duration or conventional LBBB morphology; it is strongly associated with mortality and major adverse cardiac events. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/vcg-fronphysiol-2022, rating: medium)
ΔQRS area after CRT predicts outcome: larger reductions are associated with echocardiographic response, survival and fewer ventricular arrhythmias, both in registries and in a 1,299-patient cohort. (sources/qrs-area-crt-jcm-2021, rating: medium)
ΔQRS area tracks acute haemodynamics within the patient: across 188 pacing measurements, Spearman Rs = 0.47; in patients with at least three pacing sites (n=21), median within-patient R = 0.76 (IQR 0.35–0.89). In 36/43 patients (84%), the site with the largest QRS-area reduction was within 5% of that patient's maximal %ΔLVdP/dtmax. (sources/qrs-area-crt-jcm-2021, rating: medium)
Site and electrode selection: lateral (−48 µVs) and anterolateral (−43 µVs) LV pacing reduced QRS area most, versus anterior (−13 µVs) and posterior (+9 µVs, i.e. worsening); the proximal electrode pair reduced it more than the distal pair (−39 vs −17 µVs). These differences mirrored the haemodynamic results. (sources/qrs-area-crt-jcm-2021, rating: medium)
The signal is not transportable as a population constant: the optimal anatomical site varied between patients — lateral in 48%, anterolateral in 24%, posterolateral in 19%, and occasionally anterior or posterior — which is the argument for individualised, not protocolised, lead placement. (sources/qrs-area-crt-jcm-2021, rating: medium)

Body-surface mapping — activation-time heterogeneity

The metrics: a multi-electrode ECG belt records unipolar ECGs around the torso and derives SDAT (standard deviation of activation times across all electrodes — global electrical heterogeneity), LVAT (mean activation time on left-sided electrodes), and LV and RV dispersion (SDAT of left- and right-sided electrodes — intra-chamber heterogeneity). (sources/crp-lbbap-jice-2023, rating: low)
What they add: SDAT reduction had previously been shown to predict 6-month reverse remodelling better than QRS-duration reduction, baseline QRS morphology or echocardiographic radial strain — so the metrics are proposed as a more mechanistically direct measure of resynchronisation. (sources/crp-lbbap-jice-2023, rating: low)
The QRS-duration divergence is the headline: optimised LBBAP reduced SDAT by 47% and LV dispersion by 61%, while reducing QRS duration by only 16% (157 → 133 ms). A metric can therefore register substantial electrical improvement that the surface QRS barely reflects. (sources/crp-lbbap-jice-2023, rating: low)
Metrics disagree about anodal capture: adding anodal (ring) capture to nonselective LBBP shortened QRS duration slightly but significantly worsened LVAT (51.9 vs 47.5 ms, p = 0.01) and LV dispersion (10.5 vs 8.7 ms, p = 0.01). If QRS duration were the target, anodal capture would look beneficial; by activation-time metrics it is not. (sources/crp-lbbap-jice-2023, rating: low)
LOT-CRT adds regional benefit: adding coronary-venous LV pacing to LBBAP shortened LVAT (22.7 vs 43.9 ms, p = 0.003) significantly, with non-significant improvements in other metrics. (sources/crp-lbbap-jice-2023, rating: low)

Acute haemodynamics — LV dP/dtmax and its relationship to QRS

LV dP/dtmax is the invasive contractility surrogate most often used to compare pacing configurations acutely. Unlike QRS duration it measures mechanical rather than electrical resynchronisation, and the two can diverge. (sources/lot-crt-circep-2024, rating: high)
The divergence is largest for LOT-CRT: LOT-CRT produced the greatest QRS shortening (29.5 ms) but LV dP/dtmax improvement similar to BVP (25.8% vs 26.4%), both above LBBAP alone (19.3%). Acute electrical superiority therefore did not translate into acute contractile superiority in this cohort. (sources/lot-crt-circep-2024, rating: high)
Anodal capture is a second example: bipolar LBBAP with anodal capture reduced LV dP/dtmax benefit (18.6% vs 23.7% with unipolar pacing; P < .001) despite a similar paced QRS. (sources/lot-crt-circep-2024, rating: high)
Acute metrics do not predict chronic remodelling equally well: in the CSPOT cohort, acute ΔLV dP/dtmax did not correlate with 6-month LVEF change (r = 0.01), whereas acute QRS shortening showed a directional relationship (r = 0.65). (sources/lot-crt-hrs-2025, rating: medium)

AV and VV timing — a fourth determinant of resynchronisation

AV delay is as important as pacing site: a 60 ms deviation from the optimal paced AV delay reduced the LV dP/dtmax benefit of LBBAP, LOT-CRT and BVP by approximately 30%. (sources/lot-europace-2026, rating: high)
Optimal AV delay depends on the metric: by LV dP/dtmax, LBBAP preferred a shorter delay (152 ms) than LOT-CRT (171 ms) or BVP (167 ms); by QRS duration the three were similar (164, 160 and 165 ms). (sources/lot-europace-2026, rating: high)
A patient-specific rule beats fixed programming: for LOT-CRT, PAV ≈ 0.56 × (Ap–LBBs) + 7.7 ms; for LBBAP, 0.54 × (Ap–LBBs) + 3.8 ms; for BVP, 0.48 × (Ap–RVs) + 28.0 ms. Regression-based delays reduced the mean error by about 10 ms compared with the best fixed delay. (sources/lot-europace-2026, rating: high)
VV = 0 is the best default for LOT-CRT and BVP in a predominantly IVCD cohort; pacing LBBAP first by 30 ms reduced LV dP/dtmax. The optimum is mechanism-dependent and may differ in pure conduction-system versus intramyocardial delay. (sources/lot-europace-2026, rating: high)

What the three families imply together

No single metric is the target. QRS duration is the guideline entry criterion, QRS area is the best-validated substrate measure, and SDAT/regional activation times are the most mechanistically direct — and they can rank pacing configurations differently, as the anodal-capture and LBB-capture examples show. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/crp-lbbap-jice-2023, rating: low) (sources/lbbp-vs-lvsp-criteria-jaccep-2021, rating: medium)
The metrics converge on the same physiological target — delayed LV lateral wall activation — whether measured as a vector (QRS area) or as a surface activation-time dispersion (SDAT). (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/crp-lbbap-jice-2023, rating: low)
Practical availability is the limiting step: QRS area currently needs custom VCG transformation software and SDAT needs a proprietary research belt, whereas QRS duration is on every ECG. Neither advanced metric is embedded in guidelines. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/crp-lbbap-jice-2023, rating: low)

Contradictions / Open Questions

Every advanced metric is validated against a surrogate, not an outcome. QRS area is validated against acute %ΔLVdP/dtmax (itself of uncertain relationship to long-term outcome) and against registry outcomes; SDAT is validated against echocardiographic reverse remodelling in a 20-patient study. No trial has randomised patients to a metric-guided strategy and shown improved hard outcomes — so the apparent superiority of QRS area over QRS duration is a prediction advantage, not a demonstrated clinical one. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/crp-lbbap-jice-2023, rating: low) (sources/vcg-fronphysiol-2022, rating: medium)
QRS area and QRS duration disagree, and the guidelines follow the weaker metric. QRS area outperforms duration for identifying the CRT substrate in multiple cohorts, yet current indications remain built on duration thresholds and LBBB morphology. This is a genuine evidence-practice gap, but the absence of a randomised metric-guided trial means it cannot yet be closed. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/vcg-fronphysiol-2022, rating: medium)
The within-patient relationship between QRS area and haemodynamics is highly variable. Median R = 0.76 is often quoted as "strong", but the interquartile range runs from 0.35 to 0.89 — in a substantial minority, the best electrical site is not the best haemodynamic site, and the metric would misguide lead placement. The study does not characterise which patients fall into that group. (sources/qrs-area-crt-jcm-2021, rating: medium)
QRS duration and activation-time heterogeneity can move in opposite directions. Optimised LBBAP cut SDAT by 47% but QRS duration by only 16%, and anodal capture narrowed the QRS while worsening LVAT and LV dispersion. Which metric should define "success" is unresolved, and the two could justify opposite programming choices. (sources/crp-lbbap-jice-2023, rating: low) (sources/lbbp-vs-lvsp-criteria-jaccep-2021, rating: medium)
The strongest QRS-area data come from one research group and one methodology. The QRS-area literature in this wiki draws on Maastricht/CARIM work using the Kors transform and custom Matlab analysis; the acute haemodynamic study pools two small cohorts (n=52) with different protocols. There is no multicentre validation with commercial software, and automatic QRS-area calculation is not commercially available. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/vcg-fronphysiol-2022, rating: medium)
The ECG-belt evidence is small, industry-linked and internally inconsistent. The 20-patient LBBAP study is funded by Medtronic, has a Medtronic employee author, is underpowered by the authors' own admission, and reports different p-values for the same LVAT comparison in the abstract and results. Its claim that all capture forms are electrically equivalent may reflect a 40-electrode resolution floor rather than true equivalence. (sources/crp-lbbap-jice-2023, rating: low)
Do these metrics identify the same patients? QRS area, SDAT and QRS duration have never been compared head-to-head in one cohort with outcomes. A patient with a narrow QRS but large QRS area, or a wide QRS with low SDAT, would be classified differently by different metrics — and no source says which classification is correct. (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/crp-lbbap-jice-2023, rating: low)
Electrical and mechanical resynchronisation can diverge, and the best acute electrical metric may not be the best predictor of chronic response. LOT-CRT achieved the greatest QRS shortening but did not improve acute LV dP/dtmax more than BVP, and acute ΔLV dP/dtmax did not correlate with 6-month LVEF change (r = 0.01), whereas acute QRS shortening did (r = 0.65). Which acute metric should guide intraprocedural decisions is unresolved. (sources/lot-crt-circep-2024, rating: high) (sources/lot-crt-hrs-2025, rating: medium)
AV timing is a metric in its own right. A 60 ms error in paced AV delay costs about 30% of the haemodynamic benefit, and the optimal delay differs by pacing modality and by the metric used to define it. Any comparison of pacing strategies that does not optimise AV delay is confounded. (sources/lot-europace-2026, rating: high)

Connections

Related to concepts/Cardiac-Resynchronization-Therapy — the therapy these metrics are used to select, guide and optimise
Related to entities/Vectorcardiography — the Kors transform and QRS-area derivation
Related to concepts/Left-Bundle-Branch-Area-Pacing — SDAT and regional activation times used to compare capture forms and LOT-CRT
Related to concepts/Conduction-System-Pacing — electrical resynchronisation as the rationale for physiological pacing
Related to concepts/LBBB-Criteria — true versus pseudo-LBBB, and the limits of morphology for identifying the CRT substrate
Related to entities/CRT — LV lead placement and device configuration