The use of cardiac magnetic resonance imaging (c-MRI) in risk stratification for clinical outcomes of patients with ischemic cardiomyopathy (ICM) remains low. This systematic review investigated the prognostic value of myocardial fibrosis as assessed by late gadolinium enhancement (LGE) on c-MRI in patients with ICM for ventricular tachyarrhythmia, sudden cardiac death (SCD), or all-cause mortality.

We conducted a systematic review of the electronic databases Pubmed and Embase for relevant prospective English-language studies published between January 1990 and February 2019. All included articles were prospective studies that comprised of human participants older than 18 years with ICM and a primary or secondary prevention implantable cardioverter/defibrillator (ICD); had a sample size >30 participants; had at least 6 months of follow-up; and reported on ventricular tachyarrhythmia, SCD, and all-cause mortality. A total of 90 articles related to ICM were identified and were subsequently screened independently by 2 authors. Pooled sensitivity and specificity of LGE were calculated using random-effects model.

Eight studies with 1,085 participants were included in the final analysis. The mean age of patients varied from 43 to 83 years, with most patients being men. The most common comorbidities reported included history of diabetes mellitus (22%-62%), hyperlipidemia (40%-86%), and hypertension (35%-88%). The ejection fraction of each study was reported as mean or median and varied from 22% to 35%. During a follow-up that ranged from 8.5 to 65 months, there were 110 ventricular arrhythmic events reported. The pooled sensitivity and specificity of LGE for ICD therapy delivered for ventricular arrhythmias were 0.79 (95% CI: 0.66-0.87) and 0.28 (95% CI: 0.14-0.46), respectively. For all-cause mortality, the pooled sensitivity and specificity of LGE were 0.76 (95% CI: 0.40-0.93) and 0.41 (95% CI: 0.14-0.75), respectively. Although SCD was of significant interest to our review, only 1 of the studies reported on the association between LGE and SCD, leading to the subsequent exclusion of SCD from the end point analysis.

LGE has high prognostic value in predicting adverse outcomes in patients with ICM and may provide helpful information for clinical decision making related to SCD prevention. Our findings illustrate how LGE may improve current risk stratification, prognostication, and selection of patients with ICM for ICD therapy.