Improved Risk Stratification for Ventricular Arrhythmias and Sudden Death in Patients With Nonischemic Dilated Cardiomyopathy

Aparna Irodi MD, FRCR (Vellore) reviewing Andrea Di Marco, Pamela Frances Brown, Joshua Bradley, Gaetano Nucifora et al, Improved Risk Stratification for Ventricular Arrhythmias and Sudden Death in Patients With Nonischemic Dilated Cardiomyopathy

 Journal of the American College of Cardiology, Volume 77, Issue 23, 15 June 2021, Pages 2890-2905.

Study Question:

Is it possible to provide an improved risk stratification algorithm for ventricular arrhythmia (VA) and sudden death in dilated cardiomyopathy (DCM) by combining the prognostic values of LGE and LVEF?


  • The goal of risk stratification for SCD in patients with cardiomyopathy is to identify which patients are at highest risk and most likely to benefit from ICD implantation.
  • Risk stratificationfor ventricular arrhythmias (VA) and sudden death in nonischemic dilated cardiomyopathy (DCM) remains suboptimal.
  • Traditionally, low LVEF (<35%) has been used in risk stratification. Although in ischemic heart disease, ICD implantation has been shown to reduce the rate of SCD and overall mortality in patients with LVEF < 35%, the benefit is less clear for nonischemic heart disease.
  • Myocardial fibrosis, detected via cardiac magnetic resonance (CMR) with late gadolinium enhancement (LGE) imaging, has emerged as a strong and consistent predictor of VA and sudden death. But, impact of LGE across different strata of LVEF is not very clear.
  • The authors sought to evaluate the impact of LGE on VA and sudden death across a range of LVEF strata and to develop a new algorithm for the risk stratification of VA and sudden death in patients with DCM.


  • Retrospective observational cohort study of 1,165 patients followed up for a median of 36 months.
  • Inclusion: DCM was considered as the presence of left ventricular systolic dysfunctionand dilatation, based on reduced LVEF and elevated left ventricular end-diastolic volume (LVEDV) indexed to body surface area; Non-dilated hypokinetic LV with LVEF< 45% were also included.
  • Exclusion: Primary valve disease or other cardiomyopathies such as hypertrophic cardiomyopathy, ARVC, cardiac amyloidosis /sarcoidosis, or congenital heart disease, ischemic heart disease. Patients with sustained VA or resuscitated cardiac arrest (rCA), which had occurred as the first manifestation of the disease were also excluded.
  • CMR: LV volumes and function was assessed in standard ways. LGE was assessed visually, and its extent was semi-quantified as the number of affected myocardial segments
  • Follow up and outcomes: Endpoint was a combined arrhythmic endpoint, which included appropriate ICD therapies, sustained monomorphic ventricular tachycardia (SMVT), sustained polymorphic ventricular tachycardia, rCA, and sudden death during follow-up.


  • 1,165 patients followed up for a median of 36 months.
  • LGE was present in 42% of patients.
  • The primary endpoint was reached by 74 patients (6%), with 89% of those events occurring in patients with LGE.
  • The proportion of patients with VA events was significantly higher in the LGE-positive group compared with the LGE negative group (14% vs. 1%; p < 0.001).
  • For prediction of the primary endpoint, the presence of LGE had a hazard ratio (HR) of 12.5 versus an HR of 5 without LGE in patients with LVEF ≤35%.
  • Secondary endpoint (sustained VA, rCA, or sudden death excluding appropriate ICD therapies) occurred in 44 patients: 93% of them had LGE, and the cumulative incidence of this event was 8.4% in LGE+ and 0.4% in LGE– patients (p < 0.001).
  • Specific high-risk LGE distributions (i.e, presence of any one of – epicardial LGE, transmural LGE, and combined septal and free-wall LGE) had a higher proportion of the primary endpoint compared with septal only or free-wall only.
  • An association between LGE and the combined arrhythmic endpoint was significant in the three LVEF slabs proposed in the study (≤20%, 21% to 35%, >35%)
  • LGE was the only independent predictor of the primary endpoint in patients with LVEF >35%.
  • The authors developed a 4-category risk predictive algorithm with progressively increasing risk of SCD or aborted SCD (p < 0.001).
  • Proposed algorithm for risk stratification which is called “the new clinical algorithm”:
RiskLVEFLGEPredicted yearly event rates
Intermediate- to low risk
Intermediate- to high-risk
Intermediate- to high-risk
High-risk group
> 20%
> 35%
< 20%
> 35%
< 35%
+ve, no high risk distribution
+ve, high risk distribution


  • Retrospective observational design with no validation cohort is the main drawback. Thus, despite strong internal validation and accuracy, applying this analysis to other patient populations could potentially yield different results.
  • A selection bias due to referral criteria for CMR cannot be excluded.
  • Only 8% of the study population had undergone genetic testing.


  • This paper adds to the evidence that the presence, extent and location of LGE on CMR provides important information that is additive to LVEF alone in determining which patients with DCMY are at higher risk of VA or SCD.
  • LGE was found to be a significant, consistent, and strong predictor of sustained VA and sudden death.

Patients with combined septal and free-wall LGE and those with epicardial or transmural LGE had a higher risk.


  • A clinical algorithm that integrates LVEF and LGE, as assessed by CMR imaging, improves the risk stratification for VA and sudden death in patients with nonischemic DCM.
  • Prospective studies would help in further advancing this knowledge and also to determine if and when SCD risk is low enough to forego ICD implantation in patients with DCM.