Assessment of Myocardial Scarring Improves Risk Stratification in Patients Evaluated for Cardiac Defibrillator Implantation

Study Questions:

Does assessment of myocardial scar by magnetic resonance imaging (MRI) assist in risk stratification for implantable cardioverter defibrillator (ICD) therapy?


This was a single-center cohort study of patients referred for electrophysiology study (EPS) and/or ICD placement regardless of left ventricular ejection fraction (LVEF). All patients underwent cardiac MRI, and the percentage of scar encompassing with LVEF was measured. The primary endpoint was death or appropriate ICD discharge for sustained ventricular tachycardia (VT).


The mean cohort age was 59 ± 15 years, 53% had ischemic heart disease, 33%/5% were New York Heart Association (NYHA) class III/IV, 54% had no inducible VT on EPS, and mean LVEF was 35 ± 18%. Over 24 months of follow-up, 104 patients had an ICD placed. Of the 137 total patients enrolled, 39 (28%) met the primary endpoint of death/ICD discharge. Patients meeting the primary endpoint were more likely to have advanced NYHA symptoms and inducible VT on EPS. By cardiac MRI, patients meeting the primary endpoint had worse LVEF (30 ± 14% vs. 38 ± 19%, p = 0.002), larger LV volumes (LV end-diastolic volume 246 ± 156 vs. 192 ± 84 ml, p = 0.048), and greater percentage of LV scar (12.9 [6.3-19]% vs. 5 [0-15]%, p = 0.002) than those without events. On multivariable analysis, advancing NYHA class (hazard ratio [95% CI] = 1.7 [1.2-2.4]) and myocardial scar >5% (hazard ratio 4.6 [1.8-11.8]) were predictive of the primary endpoint. Myocardial scar >5% was associated with an adjusted increased mortality risk of 5.9 [1.4-26]. In patients with an LVEF >30%, myocardial scar >5% was associated with a 6.3 [1.4-28] increased risk of death when compared to those with similar EF and scar volume ≤5%. Event rates in those with an LVEF >30% and myocardial scar >5% were similar to patients with an LVEF ≤30% (hazard ratio 0.8 [0.4-1.6, p = 0.56).


The authors concluded that myocardial scarring is predictive of adverse events in patients considered for ICD implant, regardless of LVEF.


This was an excellent study demonstrating that LV scar is predictive of events in patients undergoing ICD evaluation. Study limitations were present; it was not blinded, ICDs were in part EPS guided, and 25% of patients did not have an ICD placed (and therefore ICD shocks were not possible to tally). Taking the limitations into account, LV scar percentage (while likely correlated with LVEF) appeared to be ‘better’ than LVEF at identifying those with events. This is intuitive since scar is likely the VT re-entrant nidus. It is also very interesting to know that patients with higher LVEF (>30%) and high scar burden are equally likely to suffer events. This begs the question: ‘Should all patients with myocardial infarction regardless of LVEF undergo cardiac MRI to assess scar percentage?’ A trial is needed.

Clinical Topics: Arrhythmias and Clinical EP, Heart Failure and Cardiomyopathies, Noninvasive Imaging, Atherosclerotic Disease (CAD/PAD), Implantable Devices, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias, Acute Heart Failure, Magnetic Resonance Imaging

Keywords: Risk, Coronary Artery Disease, Myocardial Ischemia, Defibrillators, Myocardial Infarction, Follow-Up Studies, Shock, New York, Magnetic Resonance Imaging, Tachycardia, Electrophysiology, Cicatrix, Death, Heart Failure, Cardiovascular Diseases, Stroke Volume

< Back to Listings