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(8–9); non-LBBB QRS ≥150ms → A(7); QRS 120–149ms + LBBB → M(5–6) 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)

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)

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; head-to-head trials ongoing concepts/Conduction-System-Pacing
• ~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: Whether LBBAP or His-bundle pacing provides equivalent or superior resynchronization with fewer leads is under active investigation; no large mortality RCT completed
• 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)

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

Sources

• sources/bradycardia-acc-aha-hrs-2018 — 2018 ACC/AHA/HRS guideline; CRT in LVEF 36–50% + AV block + expected ventricular pacing >40% (Class IIa)
• sources/crt-companion-nejm-2004 — COMPANION; NYHA III/IV; CRT-D mortality HR 0.64
• sources/crt-carehf-nejm-2005 — CARE-HF; NYHA III/IV; CRT-P definitive mortality trial
• sources/crt-maditcrt-nejm-2009 — MADIT-CRT; NYHA I/II; QRS ≥150ms threshold
• sources/crt-longterm-raft-nejm-2024 — RAFT Long-Term; 14-year survival; acceleration factor 0.80
• sources/icd-crt-auc-2025 — 2025 AUC; appropriateness ratings for CRT scenarios
• sources/lbbb-evg-ane-2019 — Strauss strict criteria, concordant/discordant LBBB, true vs pseudo-LBBB, CRT non-responder problem
• sources/vcg-fronphysiol-2022 — QRS area (Kors-derived VCG) as superior CRT predictor; combined CMR scar + VCG QRS area
• sources/rbbb-ccr-2021 — RBBB generally no CRT benefit; exceptions RBBB+LAFB, RBBB+prolonged PR
• sources/lbbb-europace-2017 — Imaging-based CRT selection; SF, apical rocking, systolic stretch index >9.7%; LV pressure-strain loops; TAVR-induced LBBB; 30% non-responder imaging approach