key: cord-0819050-3hcqm44z
authors: Chery, Godefroy; Kamp, Nicholas; Kosinski, Andrzej S.; Schmidler, Gillian Sanders; Lopes, Renato D.; Patel, Manesh; Al-Khatib, Sana M.
title: Prognostic value of myocardial fibrosis on cardiac MRI in patients with ischemic cardiomyopathy, a systematic review
date: 2020-08-11
journal: Am Heart J
DOI: 10.1016/j.ahj.2020.08.004
sha: ea5a5ff2f04b2de191de59ea373aa09d3d31c3d5
doc_id: 819050
cord_uid: 3hcqm44z

The use of cardiac magnetic resonance imaging (c-MRI) in risk stratification for clinical outcomes of patients with ischemic cardiomyopathy (ICM) remains low. OBJECTIVES: 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. METHODS: 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 greater than 18 years old with ischemic cardiomyopathy (ICM) and a primary or secondary prevention ICD, had a sample size >30 participants, had at least 6 months 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 two authors. Pooled sensitivity and specificity of LGE were calculated using random-effects model. RESULTS: Eight studies with 1085 participants were included in the final analysis. The mean age of patients varied from 43–83 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 22–35%. 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% Cl: 0.66–0.87) and 0.28 (95% Cl: 0.14–0.46) respectively. For all-cause mortality, the pooled sensitivity and specificity of LGE were 0.76 (95% Cl: 0.40–0.93) and 0.41 (95% Cl: 0.14–0.75) respectively. While SCD was of significant interest to our review, only one of the studies reported on the association between LGE and SCD leading to the subsequent exclusion of SCD from the endpoint analysis. CONCLUSION: 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.

Ischemic cardiomyopathy (ICM) is the most common cause of heart failure (HF) due to systolic dysfunction in the western hemisphere, and it results in a number of co-morbid complications and increased mortality. ICM also poses a high risk for sudden cardiac death (SCD), and ischemic heart disease (IHD) is the most common underlying substrate associated with SCD (42% of SCD are due to IHD) 1 . Implantation of an implantable cardioverter defibrillator (ICD) is recommended for both primary and secondary prevention in appropriately selected patients at high risk of SCD 1 . While ICD implantation has been shown in multiple trials to improve survival due its termination of ventricular arrhythmia, risk-stratification and identification of patients who stand to benefit the most from ICD placement remain a challenge as only one-third of patients with an ICD receive appropriate therapy (i.e. therapy delivered for ventricular arrhythmias) within 3 years after implantation 2 . Additionally, how to identify high-risk patients who do not meet current clinical indications for ICD therapy but yet remain at risk for SCD and SCA and stand to potentially benefit from appropriate ICD therapy is uncertain. Over the past decades, several clinical risk factors and tests have been investigated as potential tools to riskstratify vulnerable participants for SCD and ventricular arrhythmias; however, few have been found to have enough discriminative and predictive power.

One test that has been proposed for SCD risk stratification is the cardiac magnetic resonance imaging (c-MRI) which is used to detect myocardial fibrosis and can characterize scar burden and distribution 3 . Myocardial fibrosis in ICM signifies scar tissue from a prior MI 1 and is an important substrate for the genesis of spontaneous ventricular arrhythmias [4] [5] [6] .It has been shown that in patients with ICM, ventricular tachycardia (VT) results from scar-related re-entry, and the scar can be visualized and assessed using late gadolinium enhancement (LGE) on c-MRI 7,8 .

There is paucity of contemporary data on the utility of the detected myocardial fibrosis on c-MRI in the risk stratification for clinical outcomes, particularly in relation to its prognostic J o u r n a l P r e -p r o o f Journal Pre-proof significance in ICM. Although a few studies have suggested an association between myocardial fibrotic burden on cMRI in patients with ICM and mortality and other cardiovascular outcomes 9-11 , these studies were limited by their small sample size, limited follow-up, and retrospective design. Also, previous reviews on this topic have combined data on participants with inducible and those with spontaneous ventricular tachyarrhythmias making it difficult to derive meaningful potential conclusions about the role of LGE in predicting spontaneous ventricular arrhythmias. In this systematic review, we aim to investigate the prognostic value of myocardial fibrosis as assessed by LGE in patients with ICM for spontaneous ventricular tachyarrhythmias, sudden cardiac death or all-cause mortality.

We searched various electronic databases including Pubmed and Embase and published bibliographies for relevant prospective English-language studies from January 1990, to February 2019. Search terms included cardiac magnetic resonance imaging or c-MRI,

LGE, arrhythmia, hypertrophic cardiomyopathy, ICM, NICM, cardiac ventricular tachycardia, ventricular fibrillation/tachycardia or death, sudden death, cardiac death, cardiac defibrillators, implantable defibrillator or implantable device intervention. This initial search was conducted using the software Distiller, the world's most used systematic review software. The search was subsequently narrowed and limited to prospective studies comprised of human participants greater than 18 years old with ischemic cardiomyopathy (ICM) and a primary or secondary prevention ICD, had a sample size > 30 participants, had at least 6 months follow-up. Additionally, only studies on ICM that report on sustained and spontaneous ventricular arrhythmias were included in this analysis.

Editorials, case reports/series, editorial articles, session presentations, systematic review articles, letters to the editor, and comments were excluded. Patients or the public WERE NOT involved in the design, or conduct, or reporting, or dissemination of our research. Outcomes of interest were SCD, sustained and spontaneous ventricular tachyarrhythmias such as ventricular tachycardia, ventricular fibrillation, aborted SCD, appropriate ICD therapy including shocks and anti-tachycardia pacing, and all-cause mortality. We included studies with composite primary or secondary outcomes/events if they reported separately on the individual outcomes. Data from those studies were then extracted and collected for the analysis. Studies with composite primary or secondary outcomes of interest that did not report separately on the individual outcomes of interest were not included in the pooled analysis. Studies that reported on inducible arrhythmias during an electrophysiology study were also excluded. Of special Interest to us were aspects of the myocardial scar that the investigators found to be most predictive of outcomes and the reported pattern of scar.

Patient's baseline characteristics are reported using median with interquartile range and Heterogeneity statistics for the meta-analysis were also generated. Individual study sensitivities and specificities were displayed with exact 95% confidence intervals. Metaanalysis was performed with the "metafor" package (version 2.0) within R statistical software version 3.4.2 (The R Foundation, Vienna, Austria, 2017). IRB approval was not needed for this study.

No extramural funding was used to support this work.

Eight studies comprised of 1,085 participants with median/mean follow-up of 8.5-64.8 months, were included in the final analysis. The total population age and male gender percentage ranged from 43-83 years and 64-88% respectively (table 1).

The most commonly reported co-morbidities were prior history of diabetes (22-62%), hyperlipidemia (40-86%), and hypertension (35-88%). The ejection fraction was reported as mean or median in each study, and it varied from 22-35% (table 1). The rate of use of beta-blockers ranged from 49% to 94% while the rate of use of angiotensin-convertingenzyme inhibitors (ACE-i) or angiotensin receptor blockers (ARB) ranged from 51% to 100%. The rate of use of aldosterone antagonists was reported in five of the eight studies and ranged from 13.8% to 56% [12] [13] [14] [15] [16] . All studies included participants meeting criteria for primary prevention ICDs [12] [13] [14] [15] [16] [17] [18] [19] . Three of the eight studies also enrolled patients receiving secondary prevention ICD 4, 15, 18 . The study that reported the number of patients with primary and secondary prevention ICDs included only 10 out of 91 patients with a secondary prevention ICDs 4 . The remaining two studies did not provide specific breakdown of patients with primary vs secondary prevention ICDs 15, 18 . Interestingly, all of the studies were conducted in European countries or Australia with one study also enrolling participants from North America 15 .

LGE c-MRI characteristics Various scar parameters were utilized to describe aspects of the myocardial fibrotic burden visualized on cMRI. Five studies use the core scar extent, two studies use the peri-infarct zone, one study uses the relative infarct transmularity (table 1) . For our analysis, scar parameters found by the investigators to be the strongest predictors of primary and/or secondary endpoints were included. Additionally, studies noted that all included subjects underwent c-MRI testing prior to ICD implantation and that follow-up was conducted during clinic visits and device interrogation via remote transmission. Study participants did not undergo repeat c-MRI following ICD implantation.

A total of 110 arrhythmic events occurred among the 1,085 participants where arrhythmic event is described as sustained and spontaneous ventricular tachyarrhythmia such as for appropriate ICD therapy/discharge and mortality are displayed in figure 2 and 3, respectively. Heterogeneity testing summary is reported above using the I² symbol.

Through our meta-analysis which included 1,085 participants, we have shown that LGE has high prognostic value in predicting adverse outcomes in patients with ICM. While In all the studies included in our systematic review, increased myocardial fibrotic burden strongly correlated with adverse events regardless of the scar parameters examined.

While LGE appears to be associated with outcome, it is not clear if it is additive to other factors (e.g. LVEF). Nonetheless, it does appear to confer a more complete prognostic implication beyond existing conventional parameters and continued contemporary improvement of ICM management. For instance, LGE correlated with worse outcomes in two of the included studies that consisted of ICM subjects on optimal medical therapy 12, 13 and in the other studies with subjects being optimized on various components of GDMT.

The specific rates of use of medical therapy of the studies, as informed by contemporary guidelines, are provided above. Moreover, a recent study revealed that infarcted myocardial tissue as measured by c-MRI may help to better identify patients at risk for monomorphic VT when combined with LVEF 13 .The key finding in our study is that While most recent studies use a binary approach by detecting the presence of scar to investigate the role of LGE in predicting outcomes, a more granular approach characterizing and grading the severity of diseased myocardium of infarct tissue heterogeneity may add to the prognostic discriminating power of LGE beyond that noted in our study. Presently, it remains unknown if certain scar parameters provide more discriminating power beyond the absolute presence/absence, as compared to others.

Specifically, does the extent of scar matter more than the location? Or does transmurality confer worse prognosis than subendocardial scar? Or does the density matter more than the transmurality of the scar? What threshold of the scar or fibrotic aspect of T1 mapping confer worse prognosis? Also, how does the selection of contrast agents or MRI vendors impact quantification of scar? Can LGE eventually be also used for CRT selection? Now that our study has illustrated that the mere presence of LGE correlates with adverse outcomes in prospective ICM participants, future randomized studies are needed to address the aforementioned questions and potential prognostic role of various scar parameters.

In relatively few small studies, both myocardial infarct size and peri-infarct border size were linked to mortality in patients with ICM 11, 24 . However, as previously noted, studies use various methods to characterize scar parameters and to quantify fibrotic burden.

These methods range from presence or absence of fibrosis to characterization of scar using transmural scar percentage, core infarct zone, total scar zone, peri-infarct zone or "grey" zone, and distribution of the scar. One reason for the heterogeneity in reporting is the lack of a national and international standardized protocol for reporting c-MRI scar.

Among studies in our review, some have found the core infarct zone to be more predictive of worse outcomes while others have found the peri-infarct zone to be more important. Recently, myocardial fibrosis noted on other imaging modalities has been correlated with poor outcomes. Borger van der Burg et al. reported that extensive scar tissue visualized on technetium-99m tetrofosmin scintigraphy is an independent predictor of death or recurrent VT in patients with ICM 25 . However, use of those other imaging modalities has been limited due to their inability to adequately assess the size, transmural extent, distribution and density of myocardial scar. c-MRI is the best imaging modality to assess and to accurately characterize myocardial scar burden as it provides information on the aforementioned scar parameters that other modalities cannot provide. As such, it is unclear what additional diagnostic or prognostic information other imaging modalities would provide beyond that of c-MRI.

Our study demonstrated that LGE provides prognostic value in the prediction of ventricular tachyarrhythmia and all-cause mortality in participants with ICM. Specifically, LGE has a high sensitivity for the prediction of the aforementioned outcomes. While the specificity of LGE in our findings is low and indiscriminate, if the LGE can be coupled with other tests with low sensitivity but high specificity, false positives may be correctly identified as negative. Therefore, LGE is not the end-all factor but rather one that contributes to the prediction of bad outcomes in patients with ICM and could potentially help improve current used of various different definitions of variables and defined outcomes. We attempted to account for this considerable amount of heterogeneity by conducting and reporting heterogeneity testing summaries in the result section above, which was negative. Fourth, our endpoint analysis is a composite of arrhythmic events which is rather a heterogeneous composite given that some studies report arrhythmic events as any appropriate ICD therapy including anti-tachycardia pacing (which itself is dependent on device settings) while others report only appropriate ICD shock for ventricular arrhythmias. This inherent issue with report of arrhythmic events is unavoidable and one that is difficult to adjust for as manuscripts do not always provide specific details regarding "appropriate" ICD therapy, which in turn may significantly impact the results of systematic review. Fifth, our endpoint analysis was calculated using binary findings (presence/absence) as reported from the original studies and thus, a dose-response relationship was not assessed. As noted above, file for more details. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents.

There is no funding source to report for this study. Bernhardt et al. 13 

AHA/ACC/HRS Guideline for Management of Patients With Ventricular Arrhythmias and the Prevention of Sudden Cardiac Death: Executive Summary. Circulation

Long-term clinical course of patients after termination of ventricular tachyarrhythmia by an implanted defibrillator

Late Gadolinium Enhancement Cardiac Magnetic Resonance Identifies Post Infarction Myocardial Fibrosis and the Border Zone at the Near Cellular Level in ex vivo Rat Heart

Infarct tissue heterogeneity assessed with contrast-enhanced mri predicts spontaneous ventricular arrhythmia in patients with ischemic cardiomyopathy and implantable cardioverter-defibrillator

Infarct morphology identifies patients with substrate for sustained ventricular tachycardia

Infarct tissue heterogeneity by magnetic survival benefit with implantable cardioverter defibrillator implantation in patients with severe ischaemic cardiomyopathy: Influence of gender

Quantitative Tissue Characterization of Infarct Core and Border Zone in Patients With Ischemic Cardiomyopathy by Magnetic Resonance Is Associated With Future Cardiovascular Events

Extent of Left Ventricular Scar Predicts Outcomes in Ischemic Cardiomyopathy Patients With Significantly Reduced Systolic Function. A Delayed Hyperenhancement Cardiac Magnetic Resonance Study

Myocardial scar characteristics based on cardiac magnetic resonance imaging is associated with ventricular tachyarrhythmia in patients with ischemic cardiomyopathy

Myocardial scar extent evaluated by cardiac magnetic resonance imaging in ICD patients: Relationship to spontaneous VT during long-term follow-up

Myocardial fibrosis predicts appropriate device therapy in patients with implantable cardioverter-defibrillators for primary prevention of sudden cardiac death

Native T1 and ECV of Noninfarcted Myocardium and Outcome in Patients With Coronary Artery Disease

Cardiovasc Imaging

Prediction of arrhythmic events in ischemic and dilated cardiomyopathy patients referred for implantable cardiac defibrillator evaluation of multiple scar quantification measures for late gadolinium enhancement magnetic resonance imaging

Prediction of life-threatening arrhythmic events in patients with chronic myocardial infarction by contrast-enhanced CMR

Late gadolinium enhancement cardiac magnetic resonance imaging for the prediction of ventricular tachyarrhythmic events: A meta-analysis

Myocardial Fibrosis Assessment by LGE Is a Powerful Predictor of Ventricular Tachyarrhythmias in Ischemic and Nonischemic LV Dysfunction: A Meta-Analysis

Myocardial Fibrosis Assessment by LGE Is a Powerful Predictor of Ventricular Tachyarrhythmias in Ischemic and Nonischemic LV Dysfunction: A Meta-Analysis

Assessment of myocardial scarring improves risk stratification in patients evaluated for cardiac defibrillator implantation

Quantitative Tissue Characterization of Infarct Heterogeneity in Patients with Ischemic Cardiomyopathy by Magnetic Resonance Predicts Future Cardiovascular Events

Impact of Viability, Ischemia, Scar Tissue

Inappropriate implantable cardioverter-defibrillator shocks: Incidence, predictors, and impact on mortality

Inappropriate Implantable Cardioverter-Defibrillator Shocks in MADIT II. Frequency, Mechanisms, Predictors, and Survival Impact

Prognostic importance of defibrillator shocks in patients with heart failure

Appropriate and inappropriate ventricular therapies, quality of life, and mortality among primary and secondary prevention implantable cardioverter defibrillator patients: Results from the pacing fast VT REduces shock ThErapies (PainFREE Rx II) trial. Circulation

Frequency and Causes of Implantable Cardioverter-Defibrillator Therapies: Is Device Therapy Proarrhythmic?

Reduction in inappropriate therapy and mortality through ICD programming

Core/total scar: <2 SD above that of a remote non-infarcted myocardium

Grey or peri-infarct zone: difference of

Core/total scar: ≥2 SD above that of a remote non-infarcted myocardium

Core/total scar: ≥2 SD above the mean of the reference range (for normal/abnormal myocardium)

core /total scar aka presence of LGE

Godefroy Chery: Conceptualization; Data curation; Formal analysis; Investigation; Methodology. Roles/Writing -original draft

Conceptualization; Data curation; Formal analysis; Investigation; Methodology

Gillian Sanders Schmidler: Methodology; Software; Supervision; Validation; Visualization

Al-Khatib: Conceptualization; Data curation; Investigation; Methodology