#cardiac-resynchronization-therapy #heart-failure #HFrEF #left-bundle-branch-block #biventricular-pacing

Cardiac Resynchronization Therapy (CRT)

Details

CRT restores mechanical synchrony in heart failure with dyssynchronous ventricular contraction. Two platforms deliver it: biventricular pacing, with a right ventricular lead and a left ventricular lead in a coronary vein, and conduction system pacing, with a lead on the His bundle or in the left bundle branch area — used increasingly, but without a guideline recommendation, because no trial with patient-centred outcomes has tested it. Devices are implanted as CRT-P (pacing only) or CRT-D (pacing plus defibrillator). The primary indication is symptomatic HFrEF with LVEF ≤35%, LBBB morphology and QRS ≥150 ms on optimal medical therapy. (sources/HF-ESC-2026, rating: very high)

This page covers the device and the decisions around it: who receives CRT, which device, when to upgrade, how it compares with conduction system pacing, and where the lead goes. The mechanism, the response and non-response problem, and the full trial hierarchy including RAFT and RAFT Long-Term are on concepts/Cardiac-Resynchronization-Therapy.

Abbreviations: A / M / R, appropriate / may be appropriate / rarely appropriate (AUC scores); AF, atrial fibrillation; AUC, appropriate use criteria; AVJ, atrioventricular junction; BVP, biventricular pacing; CSP, conduction system pacing; HBP, His bundle pacing; HFrEF, heart failure with reduced ejection fraction; IVCD, intraventricular conduction delay; LBBAP, left bundle branch area pacing; LBBB, left bundle branch block; LOT-CRT, LBBAP-optimised CRT; LVEF, left ventricular ejection fraction; PICM, pacing-induced cardiomyopathy; RVP, right ventricular pacing.

Key Facts

Indications — ESC 2026

Scenario (LVEF ≤35%, symptomatic HFrEF, optimal FMT)	Class	Level
Sinus rhythm, LBBB, QRS ≥150 ms	I	A
Sinus rhythm, LBBB, QRS 130–149 ms	IIa	C
Sinus rhythm, non-LBBB, QRS ≥150 ms	IIa	C
Sinus rhythm, non-LBBB, QRS 130–149 ms	IIb	C
QRS <130 ms without a pacing indication	III	A
Upgrade to CRT in a paced/ICD patient with worsening HF and significant RV pacing	IIa	B1
CRT rather than RV pacing where ventricular pacing is needed for high-degree AV block	IIa	B1

(sources/HF-ESC-2026, rating: very high)

Two movements to note. The lower LBBB boundary rose from 120 to 130 ms, and LBBB 130–149 ms fell from Class I to IIa — the band was Class I at ≥120 ms in 2021. That leaves LBBB with QRS 120–129 ms below the ESC threshold altogether, where a QRS under 130 ms without a pacing indication is not recommended (III/A), while the AUC still rates the same band A(7): the two frameworks do not merely differ in strength, one recommends and the other advises against. Separately, CRT instead of RV pacing for AV block dropped from Class I/A (2021) to IIa/B1 with no new trial behind the change, which reads as a re-grading of certainty rather than a reversal. (sources/HF-ESC-2026, rating: very high)

Upgrade evidence — BUDAPEST-CRT Upgrade: upgrading to CRT-D in patients with ≥20% RV pacing burden and QRS ≥150 ms reduced death and HF hospitalisation (HR 0.27; P<0.001). It is the first RCT in this setting and enrolled only 360 patients, which is why upgrade remains IIa rather than Class I. (sources/HF-ESC-2026)

Conduction system pacing: no recommendation. The guideline states that no RCT with patient-centred outcomes has evaluated CSP in HFrEF, so it makes none — consistent with the AUC's M(5–6) for CSP-CRT. That gap is why PROTECT-HF and Left vs Left matter. (sources/HF-ESC-2026; sources/icd-crt-auc-2025)

Indications — AUC 2025

Scenario	Rating	Evidence basis
LVEF ≤35%, LBBB, QRS ≥150 ms, NYHA II–IV, sinus rhythm	A(9)	MADIT-CRT, RAFT, CARE-HF
LVEF ≤35%, LBBB, QRS 120–149 ms, NYHA II–III	A(7)	MADIT-CRT and RAFT subgroups
LVEF ≤35%, non-LBBB, QRS ≥150 ms, NYHA III–IV	A(7)	Attenuated but consistent
LVEF ≤35%, non-LBBB, QRS ≥150 ms, NYHA II	M(5)	Weaker in milder symptoms
LVEF ≤35%, non-LBBB, QRS 120–149 ms	M(4–5)	Weak base; echo-guided optimisation may help selection
LVEF ≤35%, QRS <120 ms (narrow)	R(1)	ECHO-CRT and RethinQ: no benefit, possible harm
LVEF 36–50%, LBBB, QRS ≥150 ms	M(4–6)	Limited trial evidence; PICM subgroup may benefit most

(sources/icd-crt-auc-2025, rating: high)

These are expert ratings, not a meta-analysis. The document is a 17-member modified Delphi process scoring 335 scenarios on a 1–9 scale, updating 2013, and an A or M score reflects consensus about evidence rather than the evidence itself.

CRT-D versus CRT-P

The two trials that settle this were not designed to compare the two devices. COMPANION randomised to medical therapy, CRT-P or CRT-D and found a significant mortality reduction with CRT-D (HR 0.64; P=0.003) but only a borderline one with CRT-P (HR 0.76; P=0.059), while the hospitalisation benefit was equal. CARE-HF, CRT-P only and followed longer, then found a significant mortality benefit (HR 0.64; P<0.002). (sources/crt-companion-nejm-2004, rating: very high; sources/crt-carehf-nejm-2005, rating: very high)
The residual arrhythmic risk is what the defibrillator buys. In CARE-HF, 7% of CRT-P patients died suddenly — 35% of all deaths in that arm — so resynchronisation lowers mortality without abolishing the arrhythmic component. No head-to-head trial powered for all-cause mortality has been done.
Magnetic resonance imaging can stratify the choice. In 225 patients with LBBB-associated non-ischaemic cardiomyopathy and LVEF ≤35%, LGE-CMR assigned those with under 10% scar burden to LBBAP without a defibrillator and those with 10% or more to LBBAP plus an ICD. At 30 months the low-scar group had a composite of death, heart failure hospitalisation or sustained VT/VF of 6.9% against 26.1% (HR 16.52; 95% CI 3.41–80.28), an echocardiographic response rate of 78.2% against 15.7%, and LVEF normalisation in 82.2% against 25% at two years. (sources/mri-lbbap-hrs-2025, rating: medium)
Guidelines still prefer CRT-D where primary-prevention criteria are also met (LVEF ≤35%, NYHA II–IV), which is most of this population. See entities/ICD.

Atrial Fibrillation

AF reduces CRT appropriateness by about one tier (A→M) against the sinus rhythm counterpart, because biventricular pacing capture is not guaranteed. (sources/icd-crt-auc-2025, rating: high)
AVJ ablation restores it: AF with CRT plus AV junction ablation returns to A(7–8), and ESC 2026 gives AV node ablation plus CRT IIa/B1 in severely symptomatic permanent AF with poor rate control and at least one heart failure hospitalisation. (sources/icd-crt-auc-2025; sources/HF-ESC-2026)
Without near-100% biventricular pacing the response is substantially reduced — this is the reason a pacing-percentage check belongs in every AF follow-up, and the reason ablation is offered rather than reserved.

Upgrade and Pacing-Induced Cardiomyopathy

Scenario	Criterion	Rating
RV pacing-induced cardiomyopathy	LVEF fall ≥10%, or LVEF <50%, attributable to RVP	A(7) for upgrade
High pacing burden before LVEF falls	>40% ventricular pacing with LVEF ≤50%	A(7) for pre-emptive upgrade
Worsening HF in a paced or ICD patient	Significant RV pacing burden; ESC qualifies with ≥20% burden and QRS ≥150 ms	IIa/B1

(sources/icd-crt-auc-2025, rating: high; sources/HF-ESC-2026, rating: very high)

PICM is common and its risk starts below the upgrade threshold. In 257 eligible patients, 19.5% developed PICM over a mean 3.3 years, with LVEF falling from 62.1% to 36.2%. Male sex (HR 2.15) and a wider native QRS (HR 1.03 per ms; ≥115 ms 90% specific) were independent predictors, and 13% of patients pacing only 20–39% of the time developed PICM. (sources/pacing-induced-cmp-hrs-2014, rating: medium)

Conduction System Pacing and LOT-CRT

Study	n	Design	Composite	HF hospitalisation	Mortality	Other
LBBP in LBBB with HF	88	Registry subgroup	—	—	—	Paced QRS 167→124 ms; LVEF 48.8→58.1% where baseline QRS ≥120 ms
LBBAP vs BVP, LVEF ≤35%	1,778	Observational case-control, adjusted	21% vs 28%	12% vs 19%	Not different after adjustment	LVEF +13% vs +10%; complications 3.8% vs 7.5%
iCLAS, LVEF ≤50%	1,560	Case-control, propensity-matched	22.2% vs 30.8% (HR 0.81)	13.6% vs 21.8% (HR 0.63)	12.4% vs 18.2% (P=0.156)	LVEF +11.5% vs +8.9%; complications 3.5% vs 6.5%
iCLAS, HFmrEF	1,004	Observational case-control, adjusted	22% vs 34% (HR 0.64)	13% vs 23%	Not significant	LBBB subgroup composite HR 0.49
LBBP vs LVSP vs BIVP	415	Prospective registry, three arms	Freedom from HFH or death 76.6% / 48.4% / 43.6%	—	—	LVSP indistinguishable from BIVP
LEFT-BUNDLE-CRT, typical LBBB	176	Randomised non-inferiority	Response 94.6% vs 89.7% (RR 0.95; CI 0.88–1.02)	—	—	Non-inferiority not met on intention-to-treat; crossovers 20.7% vs 9.8%
LBBP vs RVP, preserved LVEF	554	Non-randomised	No difference at 18 months	—	—	Complications equal; confounded by indication

(sources/longterm-lbbap-circep-2021, rating: high; sources/lbbap-crt-jacc-2023, rating: high; sources/iclas-lbbap-crt-hrs-2025, rating: high; sources/iclas-hfmref-hrs-2025, rating: high; sources/lbbp-lvsp-crt-jaccep-2024, rating: high; sources/left-bundle-crt-ehj-2026, rating: high; sources/lbbp-vs-rvp-europace-2020, rating: medium)

The pattern is consistent and the proof is missing. Every observational comparison favours conduction system pacing on heart failure hospitalisation; mortality differences vanish after adjustment, and the first randomised trial in typical LBBB did not establish non-inferiority.
LBBAP is preferred to HBP for CSP on electrical parameters and threshold stability, and true left bundle capture appears to be the active ingredient — in the three-arm comparison, left ventricular septal pacing was indistinguishable from biventricular pacing.
LOT-CRT — an LBBAP lead in place of the right ventricular lead alongside a coronary sinus lead — produces the greatest QRS narrowing and is most useful when baseline QRS is ≥171 ms or when LBBAP yields only deep septal pacing. Criteria, outcomes and the limitations are on concepts/LOT-CRT. (sources/lot-crt-circep-2024, rating: high; sources/lot-crt-hrs-2025, rating: medium)
The scar decides which strategy can work. Computational modelling found septal scar makes CSP ineffective while CRT still shortens activation, and lateral wall scar the reverse, with HOT-CRT and LOT-CRT approaching CRT in the first case and gaining nothing on CSP in the second. (sources/scar-csp-jce-2023, rating: medium)

LV Lead Placement by QRS Area

Lead position is still chosen by coronary venous anatomy, even though several studies found no outcome difference between anatomical sites. QRS area reduction is the candidate intraprocedural substitute. (sources/qrs-area-crt-jcm-2021, rating: medium)
In an acute study of 52 patients and 188 pacing sites, lateral pacing reduced QRS area most (−48 µVs) and anterolateral nearly as much (−43 µVs), against −13 µVs anterior and +9 µVs posterior; the proximal quadripolar electrode pair outperformed the distal (−39 vs −17 µVs). Haemodynamic change agreed: +24% LVdP/dtmax laterally against +12% and +11%.
The optimum is patient-specific but usually lateral: the best site was lateral in 48%, anterolateral in 24% and posterolateral in 19%, and the largest area reduction fell within 5% of the best haemodynamic site in 36 of 43 patients. QRS duration did not discriminate between sites at all (p=0.23).
Not deployable yet. Automatic QRS-area calculation from a standard 12-lead ECG is not commercially available, and no randomised trial has compared area-guided with anatomy-guided placement. The metric itself is on concepts/Electrical-Resynchronization-Metrics.

The Randomised Trials

Trial	n	Entry	Arms	Follow-up	Primary composite	All-cause mortality
COMPANION 2004	1,520	NYHA III–IV, LVEF ≤35%, QRS ≥120 ms + PR >150 ms, sinus rhythm	Optimal therapy vs CRT-P vs CRT-D	~16 months	HR 0.81 CRT-P; 0.80 CRT-D	HR 0.76 (P=0.059) CRT-P; 0.64 (P=0.003) CRT-D
CARE-HF 2005	813	NYHA III–IV, LVEF ≤35%, QRS ≥120 ms (echo dyssynchrony if 120–149 ms)	CRT-P vs medical therapy	29.4 months	HR 0.63 (P<0.001)	HR 0.64 (P<0.002); NNT ≈9
MADIT-CRT 2009	1,820	NYHA I–II, LVEF ≤30%, QRS ≥130 ms, sinus rhythm	CRT-D vs ICD alone	2.4 years	HR 0.66 (P=0.001), heart failure events	No difference; 3% per year in both arms

(sources/crt-companion-nejm-2004, rating: very high; sources/crt-carehf-nejm-2005, rating: very high; sources/crt-maditcrt-nejm-2009, rating: very high)

MADIT-CRT's benefit is a heart failure event reduction, not survival, which is why it compares CRT-D against ICD rather than against medical therapy, and why its positive result coexists with identical mortality.
The prespecified QRS interaction is the most consequential subgroup result in CRT: QRS ≥150 ms gave HR 0.48 (0.37–0.64) and QRS 130–149 ms gave HR 1.06 (0.74–1.52), so the entry criterion was broader than the population that benefited. The same trial found women HR 0.37 against men HR 0.76 (P=0.01) and no ischaemic-versus-non-ischaemic interaction.
COMPANION's mortality benefit was concentrated in non-ischaemic cardiomyopathy (CRT-D HR 0.50; P=0.015), with ischaemic patients at HR 0.73 and no significant interaction.
RAFT and RAFT Long-Term extend the evidence to a broader population and a 14-year horizon, and are summarised on concepts/Cardiac-Resynchronization-Therapy rather than duplicated here.

LVAD and CRT-D

De novo CRT-D at LVAD implantation is R(2–3), because the left ventricular lead is mechanically redundant once the pump unloads the ventricle — a rare instance of the AUC separating CRT from ICD, since a de novo ICD in the same setting is M(4–6). (sources/icd-crt-auc-2025, rating: high)

Contradictions / Open Questions

ESC and the AUC disagree about QRS 120–149 ms in LBBB, and in opposite directions. ESC 2026 puts 130–149 ms at IIa/C and leaves 120–129 ms below its threshold entirely, where QRS <130 ms is III/A; the AUC rates both bands A(7). A patient with LBBB at 125 ms is appropriate under one framework and not recommended under the other, and no new trial separates them. (sources/HF-ESC-2026, rating: very high; sources/icd-crt-auc-2025, rating: high)
CRT-P versus CRT-D rests on indirect comparison. COMPANION's CRT-P mortality result was borderline (P=0.059) and CARE-HF's was significant with longer follow-up; the difference is follow-up duration, not device efficacy. But CARE-HF also left 7% of CRT-P patients dying suddenly. Resynchronisation lowers mortality without removing the arrhythmic fraction, and no trial has been powered to compare the two devices head-to-head. (sources/crt-companion-nejm-2004, rating: very high; sources/crt-carehf-nejm-2005, rating: very high)
MADIT-CRT's entry criterion was wider than its benefit, and its stopping rule limits interpretation. It enrolled QRS ≥130 ms but the prespecified interaction (P=0.001) found benefit only at ≥150 ms (HR 0.48) and none at 130–149 ms (HR 1.06). It stopped at the ninth of twenty planned analyses after a mean 2.4 years, which usually overestimates effects — though extended follow-up later showed mortality benefit emerging, implying the short horizon understated rather than inflated the result. (sources/crt-maditcrt-nejm-2009, rating: very high)
CARE-HF enrolled NYHA III–IV; NYHA II rests on other trials. The Class I recommendation for NYHA II–IV combines CARE-HF's mortality evidence with MADIT-CRT and REVERSE, which addressed a milder population for a composite endpoint. (sources/crt-carehf-nejm-2005, rating: very high)
CARE-HF selected on QRS width and dyssynchrony, not on LBBB morphology. The modern LBBB-first hierarchy comes from meta-analyses and post-hoc analyses, so the AUC framework extends beyond what the mortality trial directly tested. (sources/crt-carehf-nejm-2005; sources/icd-crt-auc-2025)
Non-LBBB benefit is heterogeneous and concentrated where the AUC is least confident. A(7) applies to non-LBBB with QRS ≥150 ms in NYHA III–IV, but individual response varies widely, and the 120–149 ms non-LBBB band — rated M(4–5) — is a large clinical population. Three-dimensional echo and CMR feature tracking have not been validated for selection in a trial.
Conduction system pacing versus biventricular pacing remains unproven. Observational cohorts consistently favour CSP on hospitalisation, mortality differences disappear after adjustment, and the first randomised trial in typical LBBB missed non-inferiority on intention-to-treat. Superiority still rests on acute haemodynamics and QRS narrowing. (sources/left-bundle-crt-ehj-2026, rating: high; sources/iclas-lbbap-crt-hrs-2025, rating: high)
CRT in AF needs near-universal pacing, and achieving it requires a second procedure. Efficacy depends on biventricular capture that AF rate control cannot guarantee, so AVJ ablation is effectively a condition of appropriateness — an intervention not every patient accepts or every centre offers.
Is MRI-directed defibrillator avoidance ready for practice? The MADURAI strategy is single-centre and non-randomised, its effect estimate is unstable (HR 16.52, CI 3.41–80.28), and any scar predicted events even within the low-scar group. Withholding a defibrillator on imaging is plausible and untested. (sources/mri-lbbap-hrs-2025, rating: medium)
QRS-area-guided lead placement is mechanistically coherent and clinically untested. The evidence is a 52-patient acute pooling with a haemodynamic surrogate; the within-patient correlation is strong on average (median R 0.76) but its interquartile range (0.35–0.89) means it would misguide placement in a substantial minority, and no randomised comparison with anatomy-guided placement exists. (sources/qrs-area-crt-jcm-2021, rating: medium)

Connections

concepts/Cardiac-Resynchronization-Therapy — the therapy itself: mechanism, response and non-response, imaging selection, and the evidence hierarchy including RAFT
entities/ICD — combined as CRT-D; most patients meeting CRT criteria also meet ICD criteria
entities/Heart-Failure — the syndrome CRT treats
entities/Atrial-Fibrillation — reduces appropriateness; AVJ ablation restores it
concepts/Conduction-System-Pacing — CSP as the alternative platform, and why its evidence is not yet randomised
concepts/Left-Bundle-Branch-Area-Pacing — the lead position that replaced HBP in most CSP practice
concepts/His-Bundle-Pacing — the earlier CSP target
concepts/LOT-CRT — LBBAP plus a coronary sinus lead in IVCD and failed single-site pacing
concepts/LBBAP-Outcomes — lead performance, complications and long-term capture
concepts/Pacing-Induced-Cardiomyopathy — the condition that triggers upgrade
concepts/Electrical-Resynchronization-Metrics — QRS area and activation timing for lead and electrode selection
concepts/Secondary-Mitral-Regurgitation — functional MR that CRT can reduce