Cardiac MRI-Guided Assessment of Defibrillator Benefit in Cardiac Resynchronization Therapy for Non-Ischemic Cardiomyopathy

Background: Cardiac resynchronization therapy (CRT) improves cardiac function and outcomes in patients with heart failure and ventricular conduction abnormalities, but routine combination with a defibrillator (CRT-D) in non-ischemic cardiomyopathy (NICM) has recently come under debate. Updated Dutch guidelines recommend CRT-pacemaker (CRT-P) over CRT-D for NICM patients due to lack of overall survival benefit and higher complication rates [1]. Some patients may still benefit from defibrillator therapy, and imaging markers may assist in identifying these patients. This study assessed the association of late gadolinium enhancement (LGE) and mechanical dyssynchrony on cardiac MRI (CMR) with appropriate device therapy (ADT) and mortality in NICM patients with a CRT-D.

Methods: In this retrospective cohort, NICM patients with ejection fraction < 35% receiving de novo CRT-D between January 2009 - April 2022 and had a pre-implantation CMR scan were included. CMR was used to assess LGE presence, distribution and burden, and dyssynchrony was quantified on a mid-ventricular cine slice using the systolic stretch index (SSI) [2]. The primary outcome was time to first appropriate device therapy (ADT) and secondary outcome was all-cause mortality. Follow-up continued until March 2025.

Results: Of the 115 CRT-D patients, 101 had images available for LGE analysis, of whom 45 (44.6%) were LGE-positive. In LGE-positive patients, LGE was present in median 2 segments (IQR 2–3); 35.6% had an isolated midwall septal stripe, 8.9% only had inferior hinge-point LGE. In 86 patients, scans were available for SSI analysis, median SSI was 8.96 (IQR 5.2).

Over a median follow-up of 5.0 years (IQR 6.0), 9.7% experienced ADT, with no difference between LGE-positive (6.7%) and LGE-negative (11.7%, P=0.91), or between high (11.6%) and low SSI (9.3%, P=0.73). Among LGE-positive patients, those with ADT demonstrated a significantly greater LGE burden, reflected by the number of involved segments, than those without ADT (median 6 vs. 2; P=0.02).

During follow-up, 29.2% died; mortality did not differ by LGE presence (24.4% vs 32.1%, P=0.40) or SSI (37.2% vs 30.2%, P=0.49). Neither LGE (HR 0.753; P=0.48) nor SSI (HR 0.991; P=0.88) was associated with mortality. In multivariable analysis, age (HR 1.082; P=0.002) and creatinine (HR 1.010; P=0.01) were independently associated with mortality.

Conclusion: In NICM patients undergoing CRT-D implantation for primary prevention, neither pre-implantation LGE presence nor mechanical dyssynchrony was associated with ADT or all-cause mortality. However, among LGE-positive patients, those who experienced ADT had a significantly higher LGE burden. These findings suggest that the extent of LGE may be relevant for risk stratification between CRT-D and CRT-P in selected patients, but larger studies are needed to confirm this observation.

Univariable and Multivariable analysis Appropriate Device Therapy and mortality

 

Dependent variable appropriate device therapy during follow-up	Univariable		Multivariable		Dependent variable mortality during follow-up	Univariable		Multivariable
	HR (95% CI)	p-value	HR (95% CI)	p-value		HR (95% CI)	p-value	HR (95% CI)	p-value
Age (years)	0.993 (0.936 – 1.055)	0.83			Age (years)	1.098 (1.047 – 1.152)	< .001	1.082 (1.029 – 1.136)	0.002
Sex (male)	0.338 (0.091 – 1.253)	0.11			Sex (male)	0.663 (0.336 – 1.307)	0.24
BMI (kg/m2)	1.110 (0.981 – 1.255)	0.10			BMI (kg/m2)	1.022 (0.945 – 1.105)	0.59
Hypertension	1.331 (0.621– 2.852)	0.46			Hypertension	1.017 (0.646 – 1.601)	0.94
Diabetes	1.038 (0.281 – 3.837)	0.96			Diabetes	1.093 (0.491 – 2.435)	0.83
Hypercholesteremia	1.411 (0.723 – 2.755)	0.31			Hypercholesteremia	1.144 (0.770 – 1.700)	0.51
History to AF	1.598 (0.988 – 2.582)	0.06	3.044 (0.838 – 11.053)	0.09	History to AF	1.318 (0.934 – 1.861)	0.12
Creatinine (µmol/L)	1.008 (0.996 – 1.020)	0.21			Creatinine (µmol/L)	1.012 (1.005 – 1.019)	< .001	1.010 (1.002 – 1.018)	0.01
NYHA class	0.372 (0.151 – 0.917)	0.03	0.095 (0.007 – 1.396)	0.09	NYHA class	0.846 (0.525 – 1.365)	0.49
LVEF (%)	1.005 (0.934 – 1.081)	0.90			LVEF (%)	0.979 (0.938 – 1.021)	0.32
QRS duration (msec)	1.001 (0.970 – 1.032)	0.96			QRS duration (msec)	1.004 (0.986 – 1.023)	0.64
Non-LBBB	0.536 (0.161 – 1.789)	0.31			Non-LBBB	0.471 (0.233 – 0.954)	0.04	0.488 (0.221 – 1.075)	0.08
Amiodarone	1.263 (0.163 – 9.794)	0.82			Amiodarone	1.295 (0.395 – 4.249)	0.67
Beta blocker	0.660 (0.178 – 2.441)	0.53			Beta blocker	1.171 (0.484 – 2.834)	0.73
ACE or ATII antagonist	0.666 (0.143 – 3.043)	0.60			ACE or ATII antagonist	0.573 (0.250 – 1.316)	0.19
MRA	2.169 (0.651 – 7.222)	0.21			MRA	0.962 (0.488 – 1.894)	0.91
LGE on CMR	0.602 (0.150 – 2.410)	0.47			LGE on CMR	0.702 (0.331 – 1.492)	0.36
LGE pattern	2.458 (0.265 – 22.794)	0.43			LGE pattern	0.889 (0.327 – 2.419)	0.82
LGE burden	1.085 (0.898 – 1.310)	0.40			LGE burden	0.943 (0.781 – 1.138)	0.54
SSI on CMR	1.050 (0.870 – 1.266)	0.61			SSI on CMR	0.998 (0.895 – 1.113)	0.97

* Univariable and multivariable analyses of ADT and mortality during follow-up. Hazard ratios (HR) with 95% confidence intervals (CI) and p-values are presented for each variable.

Abbreviations: ADT = appropriate device therapy; BMI = body mass index; AF = atrial fibrillation; NYHA = New York Heart Association; LVEF = left ventricular ejection fraction; LBBB = left bundle branch block; ACE = angiotensin converting enzyme; ATII = angiotensin II; MRA = mineralocorticoid receptor antagonist; LGE = late gadolinium enhancement; CMR = cardiac magnetic resonance; SSI = systolic stretch index. Figure 1. LGE Patterns and Mechanical Dyssynchrony in NICM Patients Undergoing CRT-D Implantation. Distribution of late gadolinium enhancement (LGE) patterns and systolic stretch index (SSI) in non-ischemic cardiomyopathy patients undergoing CRT-D implantation. Pie chart shows the proportion of LGE patterns; scatter plot displays SSI distribution above and below the median of 8.96. Median follow-up was 5 years.
Figure 2. Kaplan Meier curves for ADT and mortality in NICM CRT-D patients. Panel A: Kaplan-Meier analysis stratified by LGE-positive and LGE-negative CRT-D patients, showing the occurrence of appropriate device therapy (ADT), p = 0.63. Panel B: Kaplan-Meier analysis stratified by LGE-positive and LGE-negative CRT-D patients, showing mortality, p = 0.35. Panel C: Kaplan-Meier analysis stratified by SSI high and SSI low CRT-D patients, showing the occurrence of ADT, p = 0.78. Panel D: Kaplan-Meier analysis stratified by SSI high and SSI low CRT-D patients, showing the occurrence of mortality, p = 0.69.