Combination of Current and New Electrocardiographic-Based Criteria: A Novel Score for the Discrimination of Left Bundle Branch Capture

Authors, Journal, Affiliations, Type, DOI

• Authors: Sem Briongos-Figuero and Álvaro Estévez-Paniagua (joint first authors, corresponding); Ana Sánchez-Hernández; Roberto Muñoz-Aguilera
• Journal: Europace 2023;25(3):1051–1059. Received 6 July 2022; accepted 16 December 2022; published online 24 January 2023. Open access (CC BY-NC)
• Affiliations: Cardiology Department, Infanta Leonor Hospital, Madrid; Complutense University, Medicine Department, Madrid
• Type: Single-centre prospective cohort with retrospective re-analysis — 188 patients screened, 174 successful LBBAP; the validation cohort is the 71 patients with a documented QRS-transition capture criterion (127 QRS morphologies: 71 LBBP, 46 LVS)
• DOI: 10.1093/europace/euac276
• Funding: none declared. Disclosures: Briongos-Figuero and Estévez-Paniagua report personal fees from Medtronic Ibérica and Boston Scientific

Overview

The two leading ECG criteria for left bundle capture — V6-RWPT and the V6–V1 interpeak interval — had never been tested in an independent population. This single-centre study does that, and finds the originally reported cut-offs reproduce almost exactly: V6-RWPT ≤83 ms (SN 80.3%, SP 91.1%) and V6–V1 ≥33.5 ms (SN 77.1%, SP 84.6%), with >44 ms again 100% specific for capture. It then adds a new measure from the frontal plane — aVL-RWPT <79 ms (SN 71.2%, SP 88.4%; <74 ms 100% specific) — on the reasoning that precordial leads alone miss the superior–inferior component of activation, especially when the left posterior fascicle is captured and V6 adopts an rS pattern. Finally it combines the three into the "LBBP score" (0–8 points), where ≥3 points gave SN 89.2% and SP 100%, an AUC of 0.976, and correctly reclassified 10 of 14 cases misdiagnosed by the combined V6/V6–V1 criteria. The score is the practical contribution: a way to reach high sensitivity and high specificity simultaneously, which no single criterion achieves.

Keywords

Electrocardiographic-based criteria; Left bundle branch capture; Left ventricular septal capture; Conduction system pacing

Key Takeaways

What's new (the authors' own three points)

1. First external validation of V6-RWPT and V6–V1, reproducing the original optimal cut-offs (≤83 ms and ≥33 ms) almost exactly
2. aVL-RWPT <79 ms is a novel criterion adding frontal-axis information, "especially useful when R wave amplitude in V6 drops"
3. The LBBP score combines V6-RWPT, V6–V1 and aVL-RWPT into a rank score, with better joint sensitivity/specificity than any single criterion

The problem

• QRS-transition criteria during decremental pacing or deep septal programmed stimulation remain the non-invasive gold standard, but have low sensitivity, especially the threshold test, and may be time-consuming
• Existing ECG criteria overlap between ns-LBB and LVS morphologies, are influenced by stimulus latency and intraventricular conduction velocity, and depend on native QRS width and morphology
• All prior criteria use only V1–V6; loss of LBB capture produces a wider range of paced morphologies than loss of His capture, because lead position and septal depth vary — producing different forces in both superior–inferior (axis) and basal–apical (precordial transition) directions

Methods

• Population: all consecutive patients receiving LBBAP for bradycardia and/or heart failure (bail-out strategy), February 2020 – May 2022
• Implant: Medtronic C315His fixed-curve sheath and 4.1 F 3830 active-fixation lead deep in the septum
• Recording and measurement: digital EP system (GE), all 12 surface leads plus the endocardial channel simultaneously, digital calipers, 100 mm/s, mean of ≥3 complexes
• Capture definition (the gold standard for validation): unipolar paced QR/rSR′ in V1 plus at least one of:
  • (a) ns-LBBP → s-LBBP on output decrease — distinct isoelectric interval before the local EGM, M/rSR′ and wide notched R′ in V1, S in V5/V6, constant V6-RWPT
  • (b) ns-LBBP → LVS on output decrease — abrupt V6-RWPT prolongation ≥10 ms
• Descriptive-only categories (not used for validation): transition during lead screwing-in; and the combined V6-RWPT <75 ms or V6–V1 ≥33 ms rule when no transition could be demonstrated
• Measurements: aVL-RWPT, V6-RWPT from the pacing spike to the R peak; V6–V1 interpeak from V6 R peak to V1 R peak
• Statistics: ROC, multivariate logistic regression, DeLong test; score constructed heuristically — maximum points for the 100%-specific cut-off, intermediate points for the optimal cut-off, weighted by the Wald statistic

Results

Cohort

• 188 screened → 174 successful LBBAP (92.5%); 10 had no LBBAP morphology during screwing, 4 received HBP
• Capture: 128 LBB, 37 LVS, 9 indeterminate; morphology ns-LBBP 83, s-LBBP 48
• Capture diagnosis: transition ns→s 47, ns→LVS 30, both 15, combined criterion 36; transition during pacing 71, during screwing 21
• Baseline: age 79.1 ± 9.2; male 54%; hypertension 85.6%; AF 42.2%; LVEF 57.1 ± 10.8%; native QRS 119.0 ± 29.2 ms; wide QRS 48.4%; RBBB 33, RBBB + fascicular block 33, LBBB 25
• Procedural: fluoroscopy 10.4 ± 10.9 min, procedure 21.9 ± 19.7 min; paced QRS 115.6 ± 14.6 ms from onset, 146.3 ± 16.1 ms from stimulus; V6-RWPT 80.6 ± 11.3; aVL-RWPT 80.2 ± 15.6; V1-RWPT 119.9 ± 13.3; V6–V1 39.8 ± 12.8; LBB potential 74 (41.8%), potential-to-QRS 20.7 ± 5.9 ms

External validation (71 patients, 127 morphologies)

• The mean change at the ns-LBBP → LVS transition was 16.5 ± 5.2 ms in V6-RWPT and 15.6 ± 8.1 ms in aVL-RWPT — close to the arbitrary 15 ms thresholds used in the V6–V1 paper
• Subgroup performance of aVL-RWPT: non-diseased LBB (narrow QRS/isolated RBBB, n=51) — 79 ms, SN 77.3%, SP 85.6%, 100% SP at <74 ms; diseased LBB (LBBB or RBBB + fascicular block, n=20) — 83 ms, SN 77.8%, SP 91.7%, 100% SP at <76 ms
• Interobserver agreement for aVL-RWPT: ICC 0.948 (95% CI 0.897–0.973) over 60 measurements

The LBBP score

• The multivariate model (all three criteria independent, not redundant) had AUC 0.980
• The score had AUC 0.976; ≥3 points: SN 89.2%, SP 100%
• It reclassified 10/14 (71.4%) of cases the combined V6-RWPT <75 ms or V6–V1 ≥33 ms rule got wrong — 4/8 false negatives and 6/6 false positives
• Why the weighting: the interpeak interval's Wald weight was about twice the other two, so its points were doubled

How the score is meant to be used

• The logic is that single criteria cannot be simultaneously sensitive and specific. A 100%-specific cut-off (e.g. V6–V1 >44 ms) has a sensitivity of only ~37%, so many true captures are called LVS; combining three partially independent measures lets a patient accumulate points from different directions
• aVL adds a different axis of information — frontal (superior–inferior) rather than precordial (basal–apical) — which is why it can rescue cases in which the left posterior fascicle is captured and V6 shows an rS rather than an R pattern, or where the V6 R amplitude is too small to measure reliably

Limitations of the document

Authors' own stated limitations

• Too few patients with diseased LBBB, asystole or NIVCD to make strong subgroup conclusions about the criteria's performance in those substrates
• aVL-RWPT could not be measured in some patients because of a drop in R-wave amplitude — observed with a vertical or rightward basal QRS axis, and in the rare patients with a right paced axis, probably reflecting infrequent left anterior fascicle capture

Additional limitations evident in the study

• The score is not externally validated — it is derived and evaluated in the same 71 patients. The paper's external validation applies to the two pre-existing cut-offs (V6-RWPT and V6–V1); the aVL cut-off and the score's weights and thresholds are in-sample. The headline SN 89.2% / SP 100% and the "10/14 reclassified" are therefore optimistic, and no confidence intervals, calibration or bootstrap correction are reported. The score needs a second cohort before it can be adopted
• The validation cohort is still defined by the gold standard it is meant to replace. All 71 had a documented QRS transition. The 36 patients diagnosed by the combined rule were excluded from validation (correctly), but the score is intended for exactly those patients and is not tested in them. The criterion again lives in the minority with a transition and is recommended for the majority without one
• Measurement speed differs from the original papers. This study measured at 100 mm/s, whereas the V6–V1 study used 200 mm/s. That is a plausible explanation for the small differences in the 100%-specific V6-RWPT threshold (<80 vs <74 ms), but the paper does not discuss it
• The "100% specificity" of the score is a point estimate from 71 patients with no confidence interval. As with every small derivation study, the upper bound of specificity is uninformative and the lower bound is well below 100%; the reported SP is not a guarantee for a new population
• The cohort is elderly, bradycardic and mostly non-failing (LVEF 57%, CRT only 4.3%) — not the CRT population where capture adjudication matters most, and not the population with LBBB (25 patients, 14%) in whom the criteria are hardest
• The score's point allocation is heuristic, not derived from a decision curve or cost weighting. Doubling the interpeak weight because of its Wald statistic optimises statistical association, not clinical consequences of a false positive versus a false negative
• aVL-RWPT is unavailable precisely in some of the patients who need it — vertical/right axis and suspected LAF capture — so the score is undefined or biased in that subgroup
• Single centre, no follow-up, no outcomes. Whether using the score changes implant success, procedure time, lead revisions or remodeling is untested
• The 128 "LBB capture" patients include 36 diagnosed by the combined criterion rather than a transition; the validation is correctly restricted to the 71, but the overall cohort statistics mix definitions
• Rating rationale: rated high. It provides the first independent reproduction of the two dominant ECG criteria, adds a physiologically motivated frontal-plane measure, and delivers a practical combined score that materially improves the sensitivity/specificity trade-off. It is not very high because the score itself is derived and tested in the same small cohort, the validation population is selected by the reference standard, no confidence intervals or external score validation are provided, and there are no outcomes

Key Concepts Mentioned

• concepts/Left-Bundle-Branch-Area-Pacing — the technique; this paper supplies and validates its ECG capture criteria
• concepts/Conduction-System-Pacing — the umbrella category
• V6-RWPT — stimulus-to-R-peak in V6; ≤83 ms optimal, <80 ms 100% specific here
• V6–V1 interpeak interval — RV-to-LV R-peak difference; ≥33 ms optimal, >44 ms 100% specific
• aVL-RWPT — stimulus-to-R-peak in aVL; novel; <79 ms optimal, <74 ms 100% specific; adds frontal-axis information
• LBBP score — 0–8 points across the three criteria; ≥3 points = SN 89.2%, SP 100%
• QRS transition criteria — the non-invasive gold standard; ns-LBBP → s-LBBP or ns-LBBP → LVS during output decrement
• Non-selective vs selective LBB capture; LVS myocardial capture — the three states being separated
• Left posterior fascicle capture — produces an rS in V6 and reduces V6 R amplitude, motivating the aVL criterion
• Frontal versus precordial information — the rationale for adding aVL to V1–V6 criteria
• Stimulus latency — the confounder shared by all absolute RWPT measures

Key Entities Mentioned

• concepts/Left-Bundle-Branch-Area-Pacing — primary concept page
• Infanta Leonor Hospital, Madrid; Complutense University — the investigating institutions
• Sem Briongos-Figuero; Álvaro Estévez-Paniagua — the investigators
• Medtronic 3830 lead and C315His sheath — the implant system used in every case
• MELOS study — cited for the 24.8% incidence of rS morphology in V6
• Jastrzębski et al. — originators of the V6-RWPT and V6–V1 criteria being validated
• Medtronic Ibérica and Boston Scientific — disclosed personal fees
• Oxford University Press / European Society of Cardiology / Europace — publisher

Wiki Pages Updated

• Created: wiki/sources/ecg-score-lbb-capture-europace-2023.md
• Updated: wiki/concepts/Left-Bundle-Branch-Area-Pacing.md
• Updated: wiki/concepts/Conduction-System-Pacing.md
• Updated: wiki/sourceindex.md
• Updated: wiki/wikiindex.md