#cardiac-resynchronization-therapy #heart-failure-reduced-EF #biventricular-pacing #QRS-duration #device-therapy

Cardiac Resynchronization Therapy

Definition

Cardiac resynchronization therapy (CRT) delivers coordinated biventricular pacing to correct mechanical dyssynchrony in patients with heart failure, reduced ejection fraction, and wide QRS. Two device forms: CRT-P (pacing only) and CRT-D (pacing + defibrillator). CRT improves cardiac performance without increasing myocardial oxygen demand, triggers LV reverse remodeling, reduces ventricular arrhythmia burden, and extends survival.

Key Concepts

Patient Selection Criteria

Core indication (Class I in all major guidelines): NYHA II–III, LVEF ≤35%, LBBB morphology with QRS ≥150ms, on optimal medical therapy ≥3 months sources/crt-maditcrt-nejm-2009 (very high) sources/crt-carehf-nejm-2005 (very high)
NYHA III/IV + QRS ≥120ms (with echo dyssynchrony if QRS 120–149ms): established in CARE-HF (HR 0.64 for mortality; P<0.002) sources/crt-carehf-nejm-2005 (very high)
NYHA I/II + QRS ≥150ms (LBBB): supported by MADIT-CRT (HR 0.66 composite; QRS ≥150ms HR 0.48 vs 130–149ms HR 1.06; P=0.001 interaction) sources/crt-maditcrt-nejm-2009 (very high)
NYHA II/III + LVEF ≤30% + QRS ≥120ms: RAFT eligibility criteria; AF and non-LBBB morphology not excluded sources/crt-longterm-raft-nejm-2024 (very high)
CRT-D vs CRT-P: COMPANION showed CRT-D significant mortality reduction (HR 0.64; P=0.003); CRT-P borderline trend (HR 0.76; P=0.059); CARE-HF used CRT-P only (7% residual SCD rate) sources/crt-companion-nejm-2004 (very high) sources/crt-carehf-nejm-2005 (very high)
2025 AUC: NYHA II–III + LVEF ≤35% + LBBB + QRS ≥150ms → A(9); non-LBBB QRS ≥150ms → A(7); QRS 120–149ms + LBBB → A(7), a band ESC 2026 rates IIa/C at 130–149ms and does not recommend below 130ms — see entities/CRT sources/icd-crt-auc-2025 (high)

Evidence Base: Trial Hierarchy

COMPANION 2004 (NYHA III/IV; 1,520 pts; 16 months): primary composite death/any-cause hosp HR 0.81; CRT-D mortality HR 0.64 (P=0.003) — foundational CRT-D mortality evidence sources/crt-companion-nejm-2004 (very high)
CARE-HF 2005 (NYHA III/IV + dyssynchrony; 813 pts; 29.4 months): primary composite HR 0.63; mortality HR 0.64 — definitive CRT-P mortality trial; NNT≈9 sources/crt-carehf-nejm-2005 (very high)
MADIT-CRT 2009 (NYHA I/II; 1,820 pts; 2.4 years): composite HR 0.66; no mortality difference at 2.4 years; QRS ≥150ms drives benefit; women HR 0.37 — extends CRT to mild HF sources/crt-maditcrt-nejm-2009 (very high)
RAFT 2010 (NYHA II/III; 1,798 pts; 40 months): composite death/HF hosp HR 0.75 (P<0.001); death HR 0.75; confirmed CRT-D benefit in mild-to-moderate HF including AF and non-LBBB patients (parent trial of RAFT Long-Term) sources/crt-longterm-raft-nejm-2024 (very high)
RAFT Long-Term 2024 (1,050 pts from 8 RAFT sites; median 7.7 years all/13.9 years survivors): primary all-cause death acceleration factor 0.80 (95% CI 0.69–0.92; P=0.002); secondary composite AF 0.85 (0.74–0.98); curves converge after 12 years; ~80% mortality at 15 years — longest CRT survival data sources/crt-longterm-raft-nejm-2024 (very high)

Mechanisms of Benefit

Corrects intraventricular dyssynchrony → restores coordinated ventricular contraction → improves stroke volume and cardiac output without increasing myocardial work sources/crt-longterm-raft-nejm-2024 (very high)
LV reverse remodeling: reduction in LV volumes and mass documented by echo core labs in CARE-HF and MADIT-CRT sources/crt-carehf-nejm-2005 (very high) sources/crt-maditcrt-nejm-2009 (very high)
Ventricular arrhythmia reduction: CRT reduces VAs in primary (but not secondary) prevention ICD patients (RAFT arrhythmia substudy); LVEF normalisation with CRT reduces VA risk in MADIT-CRT sources/crt-longterm-raft-nejm-2024 (very high)
Early mechanistic gains (performance, remodeling, VA reduction) likely explain long-term sustained mortality benefit at 14 years sources/crt-longterm-raft-nejm-2024 (very high)

Long-Term Outcomes

At 7-year MADIT-CRT extended follow-up: mortality reduction confirmed in LBBB subgroup sources/crt-longterm-raft-nejm-2024 (very high)
At 4.4-year CARE-HF extended follow-up: original CRT-P mortality benefit persistent sources/crt-carehf-nejm-2005 (very high)
At 13.9-year RAFT Long-Term (survivor follow-up): acceleration factor 0.80 for death sustained; composite event curves begin converging after 12 years; overall 15-year mortality ~80% even with CRT-D sources/crt-longterm-raft-nejm-2024 (very high)
Benefit sustained despite AF, non-LBBB morphology, and less-widened QRS — groups known to derive less benefit sources/crt-longterm-raft-nejm-2024 (very high)

QRS Morphology and Duration as Effect Modifiers

LBBB + QRS ≥150ms: strongest benefit; Class I indication in all guidelines sources/crt-maditcrt-nejm-2009 (very high)
LBBB + QRS 120–149ms: weaker evidence; Class IIa/M in 2025 AUC sources/icd-crt-auc-2025 (high)
Non-LBBB (RBBB, NIVCD): substantially reduced or absent benefit; Class IIb/R in most scenarios sources/icd-crt-auc-2025 (high)
RAFT included 8.4% RBBB, 12.5% NIVCD, 9.0% paced QRS — yet overall benefit was significant, suggesting true LBBB patients drive a disproportionate signal sources/crt-longterm-raft-nejm-2024 (very high)
RBBB CRT exceptions: Generally no clinical benefit in RBBB + heart failure; however, RBBB + LAFB (bifascicular block) and RBBB + prolonged PR interval may represent electromechanical substrates where CRT provides benefit (Atwater 2017, Houston 2018) sources/rbbb-ccr-2021 (medium)

True vs Pseudo-LBBB and CRT Non-Responder Problem

~30% of CRT patients do not benefit; approximately 1/3 of conventionally-defined LBBB (QRS ≥120 ms) may be pseudo-LBBB — likely LVH + LAFB combination sources/lbbb-evg-ane-2019 (high)
Strauss' strict criteria (QRS ≥140/130 ms men/women + mid-QRS notching/slurring in ≥2 contiguous leads) better identify true LBBB with complete septal block — the mechanism that drives CRT benefit
True LBBB criteria (Strauss or PREDICT) associated with better echocardiographic CRT response and lower HF hospitalization than non-true LBBB (Mascioli 2012, Garcia-Seara 2018) sources/lbbb-evg-ane-2019 (high)
Controversy: Bertaglia 2017 showed stricter criteria did NOT improve CRT response vs current AHA definition
Vectorcardiography (VCG): mid-end conduction delay in QRS loop is pathognomonic of true LBBB; VCG can resolve ambiguous cases sources/lbbb-evg-ane-2019 (high)
See concepts/LBBB-Criteria for full criteria comparison
QRS area from VCG (Kors regression-derived): non-invasive alternative to intracardiac measurement; outperforms both QRS duration and conventional LBBB morphology for identifying delayed LV lateral wall activation and CRT response (van Deursen 2015, Rad 2016) sources/vcg-fronphysiol-2022 (medium)
Combined CMR scar (low focal scar) + high VCG QRS area = highest CRT response rate (Nguyên 2018); myocardial scar adversely affects QRS area and CRT hemodynamic response (Okafor 2020)

Concordant vs Discordant LBBB — CRT Prognosis

Discordant LBBB (~70%): appropriate ST/T discordance; lower LVEF (~36%), more severe disease, higher BNP, larger LV — derives greater CRT benefit sources/lbbb-evg-ane-2019 (high)
Concordant LBBB (~28–32%): T-wave concordant with QRS; higher LVEF (~51%), milder disease — derives less CRT benefit but has better baseline prognosis
Concordant LBBB in MADIT-CRT: CRT prognostic benefit greater in concordant vs discordant subgroup in that registry (Padeletti 2018) — see Contradictions
Discordant LBBB associated with more severe CAD and worse prognosis even after CRT-D (Padeletti 2018) sources/lbbb-evg-ane-2019 (high)

Imaging-Based CRT Selection — Beyond ECG Criteria

Echocardiographic dyssynchrony markers can identify CRT responders independent of ECG morphology: apical rocking (26%) and septal flash (20%) observed in CRT patients without typical LBBB — majority responded to CRT sources/lbbb-europace-2017 (high)
Septal flash (SF): Early rapid leftward IVS motion (driven by RV free wall contraction, not active IVS contraction); followed by paradoxical rightward motion; CRT predictor; SF alone in 8.4% of CRT candidates
Apical rocking: Back-and-forth apex motion from dyssynchronous IVS/lateral activation; CRT predictor; apical rocking alone in 8.6% of CRT candidates; both SF and apical rocking correction by CRT associated with LV reverse remodeling
1/3 of LBBB patients selected for CRT by ECG criteria do not have typical speckle-tracking contraction pattern; ECG-mechanics mismatch independently associated with adverse CRT outcomes sources/lbbb-europace-2017 (high)
Systolic stretch index (SSI) >9.7%: Sum of posterior-lateral systolic prestretch (pre-ejection) + septal systolic rebound stretch from 2D speckle-tracking radial strain (mid-LV short-axis view); independently predicts favorable CRT outcomes even at intermediate QRS duration (120–149ms) — where ECG criteria alone are insufficient for decision-making sources/lbbb-europace-2017 (high)
LV pressure-strain loop area (myocardial work): Non-invasive tool; septum in LBBB + chronic HF performs net negative work (systolic lengthening); after CRT, proportion of positive septal work increases dramatically — quantifies hemodynamic impact of dyssynchrony sources/lbbb-europace-2017 (high)
Additional factors for CRT response prediction beyond ECG: site of latest mechanical activation, extent and location of myocardial scar, LV lead positioning, venous anatomy
Nuclear SPECT gated perfusion imaging: phase analysis provides LV dyssynchrony indices; scar extent by LGE-CMR + VCG QRS area = highest CRT response prediction (Nguyên 2018) sources/vcg-fronphysiol-2022 (medium)

Electrical Metrics for LV Lead Placement and Optimisation

QRS area from the vectorcardiogram (Kors transform; √(Xarea² + Yarea² + Zarea²)) combines QRS duration and morphology, and its measurement variability is less than half that of QRS duration. See concepts/Electrical-Resynchronization-Metrics and entities/Vectorcardiography (sources/qrs-area-crt-jcm-2021, rating: medium)
Reduction in QRS area tracks acute haemodynamic improvement within the patient: across 188 pacing sites in 52 patients, ΔQRS area correlated with %ΔLVdP/dtmax (Spearman Rs = 0.47); in 21 patients with ≥3 sites, median within-patient R = 0.76 (IQR 0.35–0.89), and the largest ΔQRS area was within 5% of the maximal haemodynamic response in 36/43 patients (84%) (sources/qrs-area-crt-jcm-2021, rating: medium)
Where to pace: lateral (−48 µVs) and anterolateral (−43 µVs) LV pacing reduced QRS area most, versus anterior (−13 µVs) and posterior (+9 µVs); the proximal electrode pair reduced it more than the distal pair (−39 vs −17 µVs), and both findings mirrored %ΔLVdP/dtmax (sources/qrs-area-crt-jcm-2021, rating: medium)
The best anatomical site is not a population constant: the largest QRS-area reduction came from the lateral position in 48%, anterolateral in 24%, posterolateral in 19%, and anterior or posterior in the remainder — supporting individualised rather than protocolised lead placement (sources/qrs-area-crt-jcm-2021, rating: medium)
QRS duration cannot substitute: ΔQRS duration did not change across the five anatomical sites (p = 0.23) or between proximal and distal electrode pairs (p = 0.77), whereas ΔQRS area did (sources/qrs-area-crt-jcm-2021, rating: medium)
Body-surface activation mapping (ECG belt) measures SDAT, LVAT and LV/RV dispersion. During optimised LBBAP, SDAT fell 47% while QRS duration fell only 16%, and anodal capture shortened QRS but worsened LVAT and LV dispersion — so the metrics can rank the same configuration differently (sources/crp-lbbap-jice-2023, rating: low)
Neither metric is in routine use: automatic QRS-area software is not commercially available, and SDAT requires a proprietary research belt; current indications remain built on QRS duration and morphology (sources/qrs-area-crt-jcm-2021, rating: medium) (sources/crp-lbbap-jice-2023, rating: low)

Special Populations

Women: disproportionate benefit — MADIT-CRT HR 0.37 vs men HR 0.76; supports broader consideration of CRT in women, possibly at QRS 120–149ms sources/crt-maditcrt-nejm-2009 (very high)
Atrial fibrillation: historically excluded from trials; RAFT included 15.7% persistent atrial arrhythmia — overall benefit still significant sources/crt-longterm-raft-nejm-2024 (very high)
Non-ischemic CM: COMPANION non-ischemic CRT-D mortality HR 0.50 (P=0.015); CARE-HF included ischemic and non-ischemic sources/crt-companion-nejm-2004 (very high)

CRT in AV Block with Borderline LVEF (LVEF 36–50%) — ACC/AHA/HRS 2018 Guideline

Class IIa (B-R [SR]): In patients with AV block requiring PPM, LVEF 36–50%, and expected ventricular pacing >40% of the time: CRT (or His bundle pacing) preferred over RV pacing — prevents pacing-induced cardiomyopathy (sources/bradycardia-acc-aha-hrs-2018, very high)
Class IIa (B-R): LVEF 36–50% + expected ventricular pacing <40%: RV pacing is reasonable over CRT/His (lower risk of pacing-induced dysfunction at <40% pacing burden)
Evidence basis: systematic review "Impact of Physiologic Versus Right Ventricular Pacing Among Patients With Left Ventricular Ejection Fraction Greater Than 35%"; RV pacing >40% causes progressive LV dysfunction and adverse remodeling
This extends CRT use beyond LVEF ≤35% to an intermediate EF range when the pacing burden is high enough to risk CMP
Note: LVEF ≤35% with >40% anticipated ventricular pacing: CRT already addressed in 2013 ACCF/AHA HF guideline (outside scope of 2018 Bradycardia guideline)

Conduction System Pacing and LOT-CRT as CRT Alternatives

LBBAP vs BVP in LVEF ≤35% (n=1,778): composite death/HFH 21% vs 28% (adjusted HR 1.495 for BVP), HFH 12% vs 19%, mortality not significant after adjustment; LVEF +13% vs +10%; complications 3.8% vs 7.5%. (sources/lbbap-crt-jacc-2023, rating: high)
LBBAP vs BVP in LVEF ≤50%, propensity-matched (n=1,560): composite 22.2% vs 30.8% (HR 0.81), HFH 13.6% vs 21.8% (HR 0.63), mortality 12.4% vs 18.2% (HR 0.82, P = .156); LVEF +11.5% vs +8.9%. (sources/iclas-lbbap-crt-hrs-2025, rating: high)
CSP vs BVP in HFmrEF (n=1,004): composite 22% vs 34% (HR 0.64), HFH 13% vs 23%; LBBB subgroup composite HR 0.49; mortality not significant. (sources/iclas-hfmref-hrs-2025, rating: high)
LBBP vs LVSP vs BIVP (n=415): freedom from HFH/death 76.6% vs 48.4% vs 43.6%; LVSP and BIVP did not differ, suggesting that true LBB capture drives the benefit. (sources/lbbp-lvsp-crt-jaccep-2024, rating: high)
The first randomised trial in typical LBBB did not establish non-inferiority: LEFT-BUNDLE-CRT (n=176) — 94.6% BiVP vs 89.7% LBBAP in intention-to-treat (RR 0.95, 95% CI 0.88–1.02); on-treatment 92.8% vs 91.3% (RR 0.98, 95% CI 0.91–1.06) met the margin but was exploratory. Crossovers 20.7% vs 9.8%. (sources/left-bundle-crt-ehj-2026, rating: high)
LOT-CRT — LBBAP plus a coronary sinus lead — gives the greatest QRS narrowing and, in observational and acute studies, greater LVEF/LVESV improvement than BVP in IVCD and wide-QRS patients, particularly when baseline QRS ≥171 ms and LBBAP yields only deep septal pacing. See concepts/LOT-CRT. (sources/lot-crt-circep-2024, rating: high) (sources/lot-crt-hrs-2025, rating: medium)
MRI-guided risk stratification: in LBBB-associated non-ischaemic cardiomyopathy, scar burden <10% predicted LVEF normalisation in 82.2% at 2 years and a low arrhythmic risk, supporting CRT without a defibrillator in that subgroup. (sources/mri-lbbap-hrs-2025, rating: high)

Contemporary Considerations

Newer GDMT (sacubitril-valsartan/ARNi, SGLT2i): not available during most RAFT Long-Term follow-up; effect of CRT on top of optimal contemporary GDMT uncertain sources/crt-longterm-raft-nejm-2024 (very high)
Conduction system pacing (His-bundle pacing, LBBAP): emerging alternative to biventricular CRT for physiological resynchronization; the first randomised trial in typical LBBB did not establish non-inferiority in its intention-to-treat analysis, and long-term outcome data remain observational (sources/left-bundle-crt-ehj-2026, rating: high)
~80% mortality at 15 years in CRT-D patients underscores that device therapy complements but does not eliminate the natural history of advanced HFrEF

Contradictions / Open Questions

CRT benefit with newer GDMT: Original RAFT and MADIT-CRT predate ARNi and SGLT2i; whether LVEF improvement with modern GDMT reduces the population eligible for or benefiting from CRT remains unknown. DANISH caveat in 2025 AUC makes same point for ICD sources/crt-longterm-raft-nejm-2024 (very high) sources/icd-crt-auc-2025 (high)
Convergence of secondary composite curves after 12 years: RAFT Long-Term shows composite event curves (death/HTx/LVAD) begin converging after year 12, suggesting the relative benefit of CRT-D may attenuate at very long term — mechanism unclear; competing mortality from non-cardiac causes likely sources/crt-longterm-raft-nejm-2024 (very high)
CRT-P vs CRT-D in contemporary practice: COMPANION showed CRT-D mortality HR 0.64 vs CRT-P borderline trend (P=0.059) at 16 months; CARE-HF showed 7% SCD in CRT-P patients; modern GDMT may reduce SCD risk further, potentially narrowing CRT-D advantage
Non-LBBB CRT benefit: Persistent controversy about whether QRS ≥150ms non-LBBB patients benefit from CRT; individual patient meta-analysis (Cleland et al. 2013) suggests limited benefit; RAFT included these patients and still showed overall benefit sources/crt-longterm-raft-nejm-2024 (very high)
Strauss strict vs AHA criteria for CRT: Strauss strict criteria reduce false-positive LBBB by ~1/3 and better identify true LBBB mechanistically — but Bertaglia 2017 found no improvement in CRT response with strict criteria; the paradox is unresolved. True LBBB studies (Mascioli 2012, Garcia-Seara 2018) support strict criteria sources/lbbb-evg-ane-2019 (high)
Concordant/discordant LBBB and CRT benefit direction: Pérez-Riera 2019 and registry data suggest discordant LBBB (more severe disease) derives greater CRT benefit; however Padeletti/MADIT-CRT data suggest CRT prognostic benefit is greater in concordant subgroup — these findings are not directly contradictory (concordant has better baseline, discordant has more to gain) but create interpretive ambiguity for patient counseling sources/lbbb-evg-ane-2019 (high)
Sex-specific QRS thresholds: Women may benefit at QRS 120–149ms given MADIT-CRT data; 2025 AUC does not yet formally lower threshold for women sources/icd-crt-auc-2025 (high) sources/crt-maditcrt-nejm-2009 (very high)
Conduction system pacing vs biventricular CRT: Observational cohorts and one randomised non-inferiority trial have reported high CRT response rates with LBBAP and fewer heart failure hospitalisations, but the randomised trial (LEFT-BUNDLE-CRT) did not meet its intention-to-treat non-inferiority margin, and mortality differences have not been demonstrated. Whether LBBAP is equivalent, superior or slightly inferior to BVP in typical LBBB remains unresolved; a large mortality-powered RCT is still required. (sources/left-bundle-crt-ehj-2026, rating: high) (sources/iclas-lbbap-crt-hrs-2025, rating: high)
QRS area vs QRS duration for CRT selection: Van Deursen 2015 and Rad 2016 show VCG-derived QRS area outperforms QRS duration and LBBB morphology for identifying delayed LV lateral activation — yet current guidelines still rely on QRS duration thresholds; no RCT has validated QRS area-guided CRT selection strategy sources/vcg-fronphysiol-2022 (medium)
Is QRS-area-guided LV lead placement ready for practice? The supporting evidence is a 52-patient retrospective pooling of two acute haemodynamic studies with %ΔLVdP/dtmax as the endpoint — a surrogate the authors themselves note has a conflicting relationship with long-term outcome. The within-patient correlation is strong on average (median R 0.76) but its interquartile range (0.35–0.89) means that in a substantial minority the best electrical site is not the best haemodynamic site. No study has randomised lead placement by QRS area, and no commercial software computes it, so the proposal is mechanistically coherent and procedurally attractive but clinically untested. (sources/qrs-area-crt-jcm-2021, rating: medium)
Which electrical metric should define a successful configuration? QRS duration, QRS area and activation-time dispersion can rank the same pacing configuration differently — anodal capture narrows the QRS while worsening LVAT and LV dispersion, and selective LBBP gives the longest QRS of the LBBAP capture states. No source establishes which metric predicts outcome. (sources/crp-lbbap-jice-2023, rating: low) (sources/lbbp-vs-lvsp-criteria-jaccep-2021, rating: medium)

Connections

Related to concepts/Conduction-System-Pacing — emerging alternative resynchronization strategy
Related to concepts/His-Bundle-Pacing — physiological pacing approach
Related to concepts/Left-Bundle-Branch-Area-Pacing — LBBAP as CRT alternative
Related to concepts/Sudden-Cardiac-Death — CRT reduces VA burden; CRT-D adds defibrillation
Related to concepts/Arrhythmogenic-Cardiomyopathy — device therapy considerations overlap
Related to concepts/LBBB-Criteria — true vs pseudo-LBBB, Strauss strict criteria, concordant/discordant LBBB prognostic implications
Related to entities/Vectorcardiography — QRS area as superior CRT predictor; VCG transformation methods; combined CMR+VCG for CRT response prediction
Related to concepts/Electrical-Resynchronization-Metrics — QRS area, SDAT and dispersion as resynchronisation measures and lead-placement guides