- Email: [email protected]
Impact of the presence and amount of myocardial fibrosis by cardiac magnetic resonance on arrhythmic outcome and sudden cardiac death in nonischemic dilated cardiomyopathy
Impact of the presence and amount of myocardial fibrosis by cardiac magnetic resonance on arrhythmic outcome and sudden cardiac death in nonischemic dilated cardiomyopathy Martina Perazzolo Marra, MD, PhD,* Manuel De Lazzari, MD,* Alessandro Zorzi, MD,* Federico Migliore, MD, PhD,* Filippo Zilio, MD,* Chiara Calore, MD, PhD,* Giulia Vettor, MD,* Francesco Tona, MD, PhD,* Giuseppe Tarantini, MD, PhD,* Luisa Cacciavillani, MD, PhD,* Francesco Corbetti, MD,† Benedetta Giorgi, MD,† Diego Miotto, MD,† Gaetano Thiene, MD,* Cristina Basso, MD, PhD,* Sabino Iliceto, MD,* Domenico Corrado, MD, PhD* From the *Department of Cardiac, Thoracic and Vascular Sciences, and †Department of Medical Diagnostic Sciences and Special Therapies; University of Padova, Padova, Italy. BACKGROUND Current risk stratification for sudden cardiac death (SCD) in nonischemic dilated cardiomyopathy (NIDC) relies on left ventricular (LV) dysfunction, a poor marker of ventricular electrical instability. Contrast-enhanced cardiac magnetic resonance has the ability to accurately identify and quantify ventricular myocardial fibrosis (late gadolinium enhancement [LGE]). OBJECTIVE To evaluate the impact of the presence and amount of myocardial fibrosis on arrhythmogenic risk prediction in NIDC. METHODS One hundred thirty-seven consecutive patients with angiographically proven NIDC were enrolled for this study. All patients were followed up for a combined arrhythmic end point including sustained ventricular tachycardia (VT), appropriate implantable cardioverter-defibrillator (ICD) intervention, ventricular fibrillation (VF), and SCD. RESULTS LV-LGE was identified in 76 (55.5%) patients. During a median follow-up of 3 years, the combined arrhythmic end point occurred in 22 (16.1%) patients: 8 (5.8%) sustained VT, 9 (6.6%) appropriate ICD intervention, either against VF (n ¼ 5; 3.6%) or VT (n ¼ 4; 2.9%), 3 (2.2%) aborted SCD, and 2 (1.5%) died suddenly. Kaplan-Meier analysis revealed a significant correlation between the LV-LGE presence (not the amount and distribution) and malignant arrhythmic events (P o .001). In univariate Cox regression analysis, LV-LGE (hazard ratio [HR] 4.17; 95% confidence interval [CI] 1.56–11.2; P ¼ .005) and left bundle branch
block (HR 2.43; 95% CI 1.01–5.41; P ¼ .048) were found to be associated with arrhythmias. In multivariable analysis, the presence of LGE was the only independent predictor of arrhythmias (HR 3.8; 95% CI 1.3–10.4; P ¼ .01). CONCLUSIONS LV-LGE is a powerful and independent predictor of malignant arrhythmic prognosis, while its amount and distribution do not provide additional prognostic value. Contrast-enhanced cardiac magnetic resonance may contribute to identify candidates for ICD therapy not fulfilling the current criteria based on left ventricular ejection fraction. KEYWORDS Risk stratification; Dilated cardiomyopathy; Cardiac magnetic resonance; Arrhythmias; Heart failure; Late gadolinium enhancement ABBREVIATIONS CE-CMR ¼ contrast-enhanced cardiac magnetic resonance; CI ¼ confidence interval; CMR ¼ cardiac magnetic resonance; ECG ¼ electrocardiogram; EF ¼ ejection fraction; HR ¼ hazard ratio; ICD ¼ implantable cardioveter-defibrillator; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular; NIDC ¼ nonischemic dilated cardiomyopathy; SCD ¼ sudden cardiac death; VF ¼ ventricular fibrillation; VT ¼ ventricular tachycardia; WHO ¼ World Health Organization (Heart Rhythm 2014;11:856–863) I 2014 Heart Rhythm Society. All rights reserved.
Introduction This work was supported in part by research grant TRANSACSTPD113ZKJ from the University of Padova. Address reprint requests and correspondence: Dr Domenico Corrado, Inherited Arrhythmogenic Cardiomyopathy Unit, Department of Cardiac, Thoracic and Vascular Sciences, University of Padova, Via Giustiniani 2, 35121 Padova, Italy. E-mail address: [email protected].
1547-5271/$-see front matter B 2014 Heart Rhythm Society. All rights reserved.
Nonischemic dilated cardiomyopathy (NIDC) is a common heart muscle disorder, which is responsible for approximately one-third of the heart failure and is associated with significant morbidity and mortality.1,2 Sudden cardiac death (SCD) is an important mode of death accounting for 30% of all fatalities in patients with NIDC.3 Current risk stratification for SCD and http://dx.doi.org/10.1016/j.hrthm.2014.01.014
Perazzolo Marra et al
LV-LGE by CE-CMR in NIDC
indication to implant an implantable cardioveter-defibrillator (ICD) rely on reduction of left ventricular (LV) ejection fraction (EF).4,5 However, LVEF is a recognized poor method for arrhythmic risk stratification, as evidenced by the finding that the majority of patient with ICD implantation never receive appropriate therapy.6 The most important reason is that EF represents a global assessment of LV systolic function, but it does not necessarily correlate with the myocardial substrate needed for ventricular tachycardia (VT)/ventricular fibrillation (VF). Advances in our understanding of pathobiologic processes underlying both ventricular function impairment and electrical instability may offer novel predictors of arrhythmic outcome in NIDC. Clinicopathologic correlation studies have showed that the presence of myocardial fibrosis may provide a substrate for malignant ventricular arrhythmias and SCD.7,8 The clinical assessment of myocardial fibrosis has traditionally been based on endomyocardial biopsy, with inherent limitations related to the invasive approach, the small myocardial sample size, and the possible complications. Cardiac magnetic resonance (CMR) with late gadolinium enhancement (LGE) is an emerging noninvasive tool that has the ability to accurately identify and quantify ventricular myocardial fibrosis.9 Previous CMR studies in patients with NIDC have demonstrated that ventricular LGE heralds an adverse prognosis with regard to heart failure– related hospitalization and death.10–13 In these studies, arrhythmic outcome was most often evaluated as a secondary end point and a statistically significant direct association between LV-LGE and arrhythmic risk in NIDC has not been demonstrated consistently. Moreover, the existence of a dose-response relationship between the amount of nonischemic myocardial fibrosis and rate of malignant arrhythmic events remains to be assessed. The present investigation was primarily designed to assess the value of the presence and extent of myocardial fibrosis as evidenced by contrast-enhanced cardiac magnetic resonance (CE-CMR) for predicting major arrhythmic events and SCD in a consecutive series of patients with a diagnosis of NIDC according to World Health Organization (WHO) criteria.14 We rigorously characterized the disease pathogenesis by systematic coronary angiography for the exclusion of ischemic heart disease and our study population included the entire spectrum of disease phenotype ranging from early to advanced hemodynamically compromised stages.
Methods Study population and design The present study included a consecutive series of patients referred to the Heart Failure and Heart Transplantation Unit of the University Hospital of Padua for unexplained LV dilatation and dysfunction. All patients underwent detailed clinical evaluation, including 12-lead electrocardiogram (ECG), 2-dimensional transthoracic echocardiography, CECMR, and coronary angiography. The diagnosis of NIDC
857 was based on the 1995 WHO/International Society and Federation of Cardiology criteria.14 To be enrolled, patients had to have (1) a depressed LV systolic function (LVEF o50%) on a non-CMR study; (2) an angiographic study showing the absence of flow-limiting coronary artery disease (defined as Z50% luminal stenosis on coronary angiography); and (3) the absence of either valvular or hypertensive heart disease and congenital heart abnormalities. Exclusion criteria were as follows: recent onset of heart failure (o1 month); diagnosis of hypertrophic cardiomyopathy, restrictive cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, suspected infiltrative heart disease, or other specific cardiomyopathies; hemodynamic unstable conditions; contraindication to CMR (claustrophobia, pacemaker, ICD, metallic clips, atrial fibrillation, severe obesity preventing the patient from entering the scanner bore, and pregnancy); and chronic renal failure with an estimated glomerular filtration rate of o30 mL/min. The study was approved by the institutional review board, and all patients gave their informed consent.
Follow-up The follow-up data were obtained during regular outpatient visits at 3-, 6-, and 12-month intervals. All events were adjudicated by the consensus of an independent committee blinded to the CMR results. A 24-hour ECG Holter recordings were obtained for each patient without ICD every 6 months; the analysis of ECG Holter recordings focused on VT, either nonsustained or sustained. Routine ICD interrogation (at 1, 3, and 6 months and every 6 months thereafter) and ECG recordings at the time of symptoms were used to document the occurrence of spontaneous VT during follow-up in all patients. All interrogations were identified and adjudicated by 2 electrophysiologists (D.C. and F.M.) blinded to other clinical data including CMR analysis. The study outcome was the index combined end point of major arrhythmic events such as SCD, cardiac arrest due to VF, sustained VT, or appropriate ICD intervention. Sudden death was defined as any natural death occurring instantaneously or within 1 hour from the onset of symptoms. Sustained VT was defined as tachycardia originating in the ventricle with rate 4100 beats/min and lasting 430 seconds or requiring an intervention for termination. Nonsustained VT was defined as Z3 consecutive ventricular premature beats with a rate of 4100 beats/min that lasted o30 seconds during 24-hour Holter monitoring. Appropriate ICD intervention was defined as a device shock or antitachycardia overdrive pacing delivered in response to a ventricular tachyarrhythmia and documented by stored intracardiac ECG data. VF or ventricular flutter was defined as irregular or regular tachycardia with regard to polarity, amplitude, morphology, and sequence of intracardiac electrograms, with a mean cycle length of r240 ms (ie, Z250 beats/min). VT was defined as regular tachycardia with a mean cycle length of 4240 ms. For each patient, rate
858 cutoff criteria for antitachycardia pacing or shock discharge were programmed to avoid inappropriate ICD interventions and to guarantee successful therapy for ventricular tachyarrhythmias. ICD was routinely programmed to include a monitoring zone that identified VT with a rate of 4160 beats/min. Patients received a conventional device programming, with a 2.5-second delay at 170–199 beats/min and a 1.0-second delay at Z200 beats/min.
CE-CMR CMR was performed by using a comprehensive dedicated protocol including postcontrast sequences (see Online Supplemental Methods). The presence, location, and extent of LGE were independently assessed by 2 experienced observers (M.P.M. and F.C.) who were blinded to patient clinical data and outcome; ambiguous cases were reviewed by a third expert. To exclude artifact, LGE was deemed present only if visible in 2 orthogonal views (short-axis and long-axis views). Myocardial LGE was quantified by a semiautomatic detection and LGE mass (in grams) expressed as a percentage of total LV mass, according to previously validated methods.15,16 LGE was quantified by using a signal intensity threshold of 42 SD above a remote reference region.10–12,17–19 The pattern of LGE distribution was characterized as either epicardial, mid-wall, or patchy/junctional.12,20 If more than 1 pattern was present, the distribution was characterized on the basis of the predominant pattern. Patients diagnosed as having NIDC on the basis of coronary angiography, who showed at CE-CMR an “ischemic LGE pattern,” either subendocardial or transmural (most likely due to a transient occlusive coronary event, thrombosis or vasospasm, or embolization from nonstenotic but unstable coronary plaque), were excluded from the final analysis.
Heart Rhythm, Vol 11, No 5, May 2014 correlation of traditional clinical predictors of arrhythmic risk in NIDC with the index arrhythmic end point during follow-up was determined by means of the Cox regression analysis. Variables with a P value of o.15 in the univariate analysis were integrated into multivariable analysis to estimate the hazard ratio (HR) and to identify independent predictors of major arrhythmic events. The Cox model was used to calculate the relation between the amount of LGE and hazard ratios. HRs and confidence intervals (CIs) are presented both in univariate and in multivariable analysis. Linear regression was used to assess the correlation between the 2 independent observers performing LGE planimetry. A P value of o.05 was considered statistically significant. Statistics were analyzed with SPSS version 19 (SPSS Inc, Chicago, IL).
Results Baseline characteristics Of a total of 276 patients initially referred to our department for clinical evaluation of LV dilatation and dysfunction, 27 refused or had contraindication to CE-CMR. One hundred eight patients were excluded from the study because of a diagnosis of “ischemic” dilated cardiomyopathy on the basis of angiographic demonstration of flow-limiting coronary artery disease (n = 102), CE-CMR findings of “ischemic pattern” of LGE in the absence of coronary artery stenosis (n = 4), or an LV dilatation/dysfunction “out of proportion” to concomitant single vessel coronary artery disease (n = 2). Four patients were lost over the follow-up owing to no contact. The remaining 137 patients (median age 49 years; 108 men) fulfilled the criteria for NIDC and constituted the final study population. Baselines clinical characteristics of the enrolled patients are reported in Online Supplemental Table 1.
LV-LGE Statistical analysis Data are expressed as mean ⫾ SD and as median with 25–75 percentiles for normally distributed and skewed variables, respectively. Normal distribution was assessed by using the Shapiro-Wilk test. Categorical differences between groups were evaluated by using the χ2 test or the Fisher exact test, as appropriate. Paired and unpaired t tests were used to compare normally distributed continuous variables obtained, respectively, from the same patient and different patients; paired and unpaired rank sum tests were used for skewed continuous variables. Kaplan-Meier analysis was used to estimate the survival distributions of the arrhythmic end point and to show the differences in survival between patients with vs without LGE. The start of follow-up was defined as the date of the initial CMR. Patients were censored at the time of their first event or the time of their last clinical follow-up. The mean event rate per year was evaluated by the number of events occurring during the follow-up divided by the number of patients multiplied by the average duration of follow-up. The
In postcontrast sequences, LV-LGE was identified in 76 (55.5%) patients. The LGE pattern of distribution was epicardial in 23 (30.3%) patients, mid-wall in 46 (60.5%) patients, or patchy/junctional in 7 (9.2%) patients. In 8 patients, the predominant epicardial pattern was associated with either septal mid-wall (n ¼ 5) or junctional (n ¼ 3) LGE patterns. The median extent of LGE was 9% (4%–15%) of LV mass. The representative cases of LV-LGE regional distribution patterns and histopathologic correlations are, respectively, shown in Figures 1A–1C and Online Supplemental Figure 1. The presence of LV-LGE was unrelated to sex, age, New York Heart Association functional class, LVEF, or any other clinical baseline characteristic (see Online Supplemental Table 1). No significant correlation between the presence of LV-LGE and any of the CMR functional parameters was found. Linear regression analysis showed a high correlation between the 2 observers for planimetry of %LGE (r ¼ .94; P o .01).
Perazzolo Marra et al
LV-LGE by CE-CMR in NIDC
859
Figure 1 Representative cases of myocardial fibrosis on CE-CMR underlying major arrhythmic events during follow-up. A: Postcontrast short-axis view of a patient with focal mid-wall anteroseptal LGE who experienced aborted SCD owing to ventricular fibrillation, which was interrupted by ICD shock (D). B: Circumferential mid-wall LGE distribution on postcontrast short-axis view in a patient who was hospitalized for syncope, without ICD, and had in-hospital VF that required external defibrillation (E). C: Epicardial LGE distributing to anterior, septal, and inferior LV walls on postcontrast short-axis view in a patient who experienced a sustained ventricular tachycardia (F). CE-CMR ¼ contrast-enhanced cardiac magnetic resonance; ICD ¼ implantable cardioveter-defibrillator; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular.
Arrhythmic outcome During a median follow-up of 3 years (range 31 days to 9.6 years), a total of 69 (50.3%) patients underwent ICD implantation, either for primary (N ¼ 62) or secondary (N ¼ 7) prevention, which was combined with cardiac resynchronization therapy in 14 (20%) patients. The primary arrhythmic end point occurred in 22 (16.1%) patients, with a 5.4% annual rate of major arrhythmic events (see Online Supplemental Table 2). Eight (5.8%) patients had an episode of sustained VT, 9 (6.6%) had Z1 appropriate ICD interventions, either against VF (n ¼ 5) or VT (n ¼ 4), and 3 (2.2%) experienced aborted SCD and 2 (1.5%) died suddenly. Representative examples of major arrhythmic events that occurred in patients with LV-LGE are shown in Figures 1D–1F. Table 1 summarizes the clinical characteristics of patients with or without major arrhythmic events during follow-up. Patients who experienced arrhythmic events significantly more often showed a left bundle branch block ECG pattern (54.5% vs 28.9%; P ¼ .02) and the presence of LV-LGE on CE-CMR (77% vs 51%; P ¼ .04). Figures 2A and 2B show Kaplan-Meier analysis of survival from the index combined arrhythmic end point stratified by the presence of LGE and LVEF. Overall, the annual event rate was 7% in patients with LV-LGE vs 3.3% in those without LGE on CE-CMR (log-rank P ¼ .002)
(Figure 2A). The annual event rate was 5.1% in patients with reduced EF (r35%) vs 2.4% in those with EF 435% (logrank, P ¼ .467; Figure 2B). The arrhythmic event-free survival did not differ between patients whose %LV-LGE mass measured below the median and patients with values above the median (log-rank, P ¼ .295; Figure 3). The prevalence of major arrhythmic events was not associated with a specific LGE pattern. During follow-up, 27 of 137 (19.7%) patients experienced an episode of heart failure, requiring intravenous administration of diuretics and inotropes (13 in the LGE-negative group and 14 in the LGE-positive group) and 12 of 137 (8.8%) had non-SCD or heart transplantation (6 in the LGEnegative group and 6 in the LGE-positive group). Of the 68 patients undergoing serial 24-hour ECG Holter monitoring during follow-up, 30 (44%) had Z1 episode of nonsustained VT. There was an association of borderline statistical significance between nonsustained VT and LVLGE by CMR (P ¼ .071), while patients with (n ¼ 19) and without (n ¼ 11) nonsustained VT did not differ significantly with regard to LGE amount and pattern.
Predictors of arrhythmic events The results of univariate and multivariate analyses for predictors of major arrhythmic events during follow-up are summarized in Table 2. Univariate predictors of events were the presence of left bundle branch block on ECG and the
860
Heart Rhythm, Vol 11, No 5, May 2014
Table 1 Characteristics of patients with and without arrhythmic events during follow-up Characteristic Age (y) Sex: male, n (%) NYHA class, n (%) I II III IV Medications, n (%) ACEI or ARB therapy β-Blockers Spironolactone Diuretic Amiodarone Digoxin Statin LBBB, n (%) EDV (mL/m2) EF (%) EF o35%, n (%) LGE, n (%)
No arrhythmic events Arrhythmic events (n ¼ 115 [84%]) (n ¼ 22 [16%]) 47 (37–60) 90 (78.3)
59 (43–70) 18 (81.8)
29 37 45 4
(25) (32) (39) (4)
4 (18) 11 (50) 7 (32) 0 (0)
102 (87)
18 (82)
88 51 87 22 14 28 33 109 33 67 59
(76) (44) (76) (19) (12) (24) (28.9) (87–140) (28–40) (58.3) (51)
19 (86) 11 (50) 20 (91) 3 (14) 4 (18) 9 (41) 12 (54.5) 123 (105–143) 30 (29–40) 15 (68.2) 17 (77)
Continuous values are expressed as median (25th-75th percentiles). ACEI ¼ angiotensin-converting enzyme inhibitor; ARB ¼ angiotensin receptor blocker; EF ¼ echocardiographic ejection fraction; EDV ¼ echocardiographic end-diastolic volume; LBBB ¼ left bundle branch block; LGE ¼ late gadolinium enhancement; NIDC ¼ nonischemic dilated cardiomyopathy; NYHA ¼ New York Heart Association.
presence of LV-LGE on CE-CMR. Neither LVEF (HR 0.99; 95% CI 0.94–1.04; P ¼ .62) nor the percentage of LV-LGE extent (HR 1.04; 95% CI 0.98–1.09; P ¼ .18) significantly predicted arrhythmic outcome (Table 2). In multivariable analysis, the presence of LV-LGE remained the only independent predictor of arrhythmic events (HR 3.8; 95% CI 1.3–10.4; P ¼ .01).
Figure 3 Kaplan-Meier analysis of freedom from major arrhythmic events stratified by the extent of LGE (median value). LGE ¼ late gadolinium enhancement.
Discussion Our study was designed to evaluate the value of the presence and extent of LV-LGE by CE-CMR for predicting major arrhythmic events and SCD in a large population of patients with NIDC diagnosed according to the WHO criteria.14 By study design, the hemodynamic patient profile encompassed the entire disease spectrum including patients with severe LV dysfunction who meet criteria for ICD implantation and those with mild or moderate disease forms who did not. Dilated cardiomyopathy was classified as “nonischemic” by coronary angiography findings in association with the LGE distribution pattern, which is consistent with current clinical practice and allows the most accurate etiology definition for clinical research purposes. The major study results are as follows: (1) the presence of LV-LGE is a powerful and independent predictor of arrhythmic
Figure 2 Kaplan-Meier analysis of freedom from major arrhythmic events stratified by the presence of LGE (A) and reduced LV ejection fraction (B) on CECMR. CE-CMR ¼ contrast-enhanced cardiac magnetic resonance; EF ¼ ejection fraction; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular.
Perazzolo Marra et al Table 2
LV-LGE by CE-CMR in NIDC
861
Predictors of arrhythmic events during follow-up Univariate analysis
Multivariable analysis
Characteristic
HR
CI
P
Age Sex: male Family history of DCM NYHA class: I-II vs III-IV CRT LBBB LVEVD (mL/m2) LVEF (%) LGE presence LGE extent (% of LV mass)
1.02 0.89 0.39 1.34 1.06 2.34 1.01 0.98 4.17 1.04
0.99–1.04 0.3–2.6 0.11–1.41 0.55–3.29 0.98–1.08 1.01–5.41 0.99–1.02 0.94–1.08 1.56–11.2 0.98–1.09
0.22 .80 .38 .524 .51 .048 .48 .63 .005 .18
HR
CI
P
2.3
0.9–5.1
.06
3.8
1.3–10.4
.01
CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; DCM ¼ dilated cardiomyopathy; EF ¼ ejection fraction; HR ¼ hazard ratio; LBBB ¼ left bundle branch block; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular; LVEDV ¼ left ventricular end-diastolic volume; NYHA ¼ New York Heart Association.
outcome during a median follow-up period of 3 years; (2) the presence of LV-LGE appears to be superior to traditional LVEF r35% in predicting major arrhythmic events and SCD; (3) the absence of LGE characterizes a low arrhythmic subgroup of patients with no SCD events during follow-up.
Previous studies See Online Supplemental Discussion.
Arrhythmic prognostic value of LGE In the present study, LGE was found in 76 of 137 (55.5%) patients with NIDC, a figure comparable to that of previous CMR studies which reported a prevalence of 30%–65% and similar to that of pathologic studies on NIDC hearts in which macroscopic LV scarring was detected in up to half of the cases.21,22 The present CMR study is in agreement with previous investigations, which consistently demonstrated the prognostic value of LGE in patients with NIDC.10–13 Our study results significantly showed that LGE on CE-CMR identifies patients at increased risk of major arrhythmic events and SCD during a long-term follow-up. Although in univariate Cox regression analysis, both abnormal CE-CMR and left bundle branch block were significant predictors for the composite arrhythmic end point, yielding an HR of 3.8 and 3.4, respectively, the presence of LGE remained the only independent predictor of adverse arrhythmic outcome in multivariable analysis (HR 3.8). Among patients with LGE, we found a relatively small amount of LGE with a median percentage of 9%. Moreover, in our study a greater extent of LGE did not appear to confer a higher arrhythmic risk. The annual rates of arrhythmic events did not differ significantly between patients whose % LV-LGE mass measured below the median (5.7%/y) compared with those with values above the median (8.6%/y). This finding diverged from that of previous studies, which reported a prognostic significance of LV-LGE size, with the arrhythmic risk being proportional to the increasing percentage of LV scar.10,16 In the Gulati study,23 the LGE size was
an independent predictor of malignant arrhythmic events with an HR of 1.08 per 1% increase. This discrepancy may be explained by the difference in study population, considering that we included only patients with angiographically proven NIDC while in the Gulati study coronary artery disease was not rigorously ruled out by coronary angiography. The fact that in the latter study a proportion of enrolled patients could have an ischemic myocardial scar is relevant if one considers that previous CMR studies actually showed a direct association between the amount of postinfarction scar and arrhythmic risk.24–28 In this regard, Klem et al24 found a significant increase in the risk when the LGE size exceeded 5% of LV mass, while Kwong et al25 reported a significant risk increase even for the lowest tertile of LVLGE amount (1.4%). Thus, we can speculate that the lack of a dose-response relationship between LGE and arrhythmic risk in our study may be explained by the different pathophysiologic significance of myocardial fibrosis in ischemic dilated cardiomyopathy vs NIDC. While ventricular tachyarrhythmias are mechanistically (ie, scar-related macroreentry) and, thus, quantitatively related to postinfarction scar, LV-LGE in NIDC may not be the sole source of ventricular arrhythmias. We can speculate that compact myocardial fibrosis, as evidenced by LGE, is part of a more complex fibrotic process also involving patchy and interstitial fibrosis, which may fall under the resolution power of traditional LGE (Online Supplemental Figure 1). Interstitial fibrosis, in which short collagen septa are interspersed with myocardial fibers, may play an important arrhythmogenic role in the setting of nonischemic heart disease by predisposing to slow conduction and leading to unidirectional block.26 The emerging T1 mapping sequence technique will offer the potential to improve the clinical assessment of interstitial myocardial fibrosis and to refine the prognostic value of LGE. Although we observed significantly higher rates of major arrhythmic events in LV-LGE– positive patients with NIDC, the absence of LV-LGE did not guarantee an uneventful arrhythmic outcome, given that 5 (8%) LV-LGE–negative patients experienced arrhythmic events such as VT (n ¼ 3
862 [4.9%]) and appropriate ICD intervention (n ¼ 2 [3.2%]). These findings may be explained by various mechanisms of ventricular tachyarrhythmias in NIDC, partly unrelated to myocardial fibrosis, which include functional bundle branch/ interfascicular macroreentry, anisotropic conduction, or spatial dispersion of ventricular repolarization.27–29 It is noteworthy that in our NIDC cohort, both aborted SCD and SCD occurred exclusively in the subgroup of patients with LGE, suggesting that VF mostly occurs in the context of scarring myocardium.
Study limitations Because endomyocardial biopsy was not performed systematically, an infiltrative disease could not be excluded definitively at the time of the study enrollment; however, no patients had a CMR pattern consistent with either infiltrative or storage myocardial disease. In the setting of ischemic heart disease, the “infarct gray zone” on CE-CMR has been demonstrated to reflect the infarct tissue heterogeneity and predict the occurrence of spontaneous ventricular arrhythmias.30 The prevalence and prognostic meaning of scar heterogeneity in patients with NIDC was not investigated by both the previous and the present studies because of methodological limitations related to the smaller amount of myocardial fibrosis and the specific scar pattern, not comparable with that of post-infarction scar. Moreover, the quantification of hyperenhanced signals of nonischemic myocardial scar is not yet standardized.31 We quantified LGE by using a signal intensity threshold of 42 SD above a remote reference region, according to the current guidelines17 and most of the previous studies on CE-CMR in NIDC. Although relatively large areas of scarring can be identified by using the traditional LGE technique, small patches of fibrosis as well as interstitial fibrosis, which may have a potential role in arrhythmia susceptibility, cannot be detected. In the near future, the use of T1 mapping system will offer the potential to improve clinical assessment of diffuse interstitial myocardial fibrosis and to further refine the prognostic value of LGE. Finally, our finding that scar amount does not provide additional prognostic value for malignant arrhythmic events and SCD needs to be confirmed by further studies showing a broader range of LV scar.
Clinical implications An ICD implantation has been proven to significantly reduce mortality in NIDC. The current guidelines recommend ICD implantation for primary prevention in the subgroup of patients with reduced LVEF (r35%).32 However, risk assessment based on LVEF alone is not accurate enough, given that the majority of patients receiving an ICD for primary prevention never experience appropriate therapy. In this regard, EF represents a global assessment of LV systolic function and does not necessarily correlate with myocardial lesions underlying ventricular electrical instability.
Heart Rhythm, Vol 11, No 5, May 2014
Conclusions Our study showed that LV-LGE by CE-CMR is independently associated with an adverse arrhythmic prognosis in patients with NIDC. This suggests that risk stratification based on the demonstration of the myocardial scar substrate has the potential to better identify patients with NIDC at risk of SCD and to improve accuracy of indication of ICD implantation. Unlike traditional imaging techniques such echocardiography that disclose LV mechanical dysfunction (either regional or global), CE-CMR allows the detection of myocardial fibrosis that may act as a potentially arrhythmogenic myocardial substrate. This explains why the presence of LV-LGE is a stronger predictor of adverse arrhythmic outcome rather than traditional LVEF. There is a high level of concern among physicians about stratification of the risk of SCD in patients with mild to moderate NIDC who do not meet ICD implantation criteria by EF. Our study population included such patients, and the results remained statistically significant after adjustment for LVEF. Thus, the detection of myocardial scar may identify a subgroup of patients with NIDC and with a high arrhythmic risk and less severe LV dysfunction, not currently fulfilling criteria for ICD implantation. In our study, a subanalysis of the predictive value of LGE for arrhythmic outcome in the subgroup of patients with LVEF 435% was not performed because of the limited number of events. Further studies on larger population with longer follow-up are needed to confirm the arrhythmic prognostic value of CE-CMR in patients with NIDC and with early disease stage and to substantiate the indication of prophylactic ICD implantation based on LV-LGE demonstration.
Appendix Supplementary data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.hrthm. 2014.01.014.
References 1. Jefferies JL, Towbin JA. Dilated cardiomyopathy. Lancet 2010;375:752–762. 2. Maron BJ, Towbin JA, Thiene G, Antzelevitch C, Corrado D, Arnett D, Moss AJ, Seidman CE, Young JB. Contemporary definitions and classification of the cardiomyopathies. Circulation 2006;113:1807–1816. 3. Felker GM, Thompson RE, Hare JM, Hruban RH, Clemetson DE, Howard DL, Baughman KL, Kasper EK. Underlying causes and long-term survival in patients with initially unexplained cardiomyopathy. New Engl J Med 2000;342: 1077–1084. 4. Goldberger JJ, Buxton AE, Cain M, et al. Risk stratification for arrhythmic sudden cardiac death: identifying the roadblocks. Circulation 2011;123: 2423–2430. 5. Passman R, Goldberger JJ. Predicting the future: risk stratification for sudden cardiac death in patients with left ventricular dysfunction. Circulation 2012;125: 3031–3037. 6. Ezekowitz JA, Rowe BH, Dryden DM, Hooton N, Vandermeer B, Spooner C, McAlister FA. Systematic review: implantable cardioverter-defibrillators for adults with left ventricular systolic dysfunction. Ann Intern Med 2007;147: 251–262. 7. Wu TJ, Ong JJ, Hwang C, Lee JJ, Fishbein MC, Czer L, Trento A, Blanche C, Kass RM, Mandel WJ, Karagueuzian HS, Chen PS. Characteristics of wave fronts during ventricular fibrillation in human hearts with dilated
Perazzolo Marra et al
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
LV-LGE by CE-CMR in NIDC
cardiomyopathy: role of increased fibrosis in the generation of reentry. J Am Coll Cardiol 1998;32:187–196. Pogwizd SM, McKenzie JP, Cain ME. Mechanisms underlying spontaneous and induced ventricular arrhythmias in patients with idiopathic dilated cardiomyopathy. Circulation 1998;98:2404–2014 Mewton N, Liu CY, Croisille P, Bluemke D, Lima JA. Assessment of myocardial fibrosis with cardiovascular magnetic resonance. J Am Coll Cardiol 2011;57: 891–903. Assomull RG, Prasad SK, Lyne J, Smith G, Burman ED, Khan M, Sheppard MN, Poole-Wilson PA, Pennell DJ. Cardiovascular magnetic resonance, fibrosis, and prognosis in dilated cardiomyopathy. J Am Coll Cardiol 2006;48:1977–1985. Wu KC, Weiss RG, Thiemann DR, Kitagawa K, Schmidt A, Dalal D, Lai S, Bluemke DA, Gerstenblith G, Marbán E, Tomaselli GF, Lima JA. Late gadolinium enhancement by cardiovascular magnetic resonance heralds an adverse prognosis in nonischemic cardiomyopathy. J Am Coll Cardiol 2008;51: 2414–2421. Lehrke S, Lossnitzer D, Schöb M, Steen H, Merten C, Kemmling H, Pribe R, Ehlermann P, Zugck C, Korosoglou G, Giannitsis E, Katus HA. Use of cardiovascular magnetic resonance for risk stratification in chronic heart failure: prognostic value of late gadolinium enhancement in patients with non-ischaemic dilated cardiomyopathy. Heart 2011;97:727–732. Masci PG, Barison A, Aquaro GD, Pingitore A, Mariotti R, Balbarini A, Passino C, Lombardi M, Emdin M. Myocardial delayed enhancement in paucisymptomatic nonischemic dilated cardiomyopathy. Int J Cardiol 2012;157:43–47. Richardson P, McKenna W, Bristow M, Maisch B, Mautner B, O'Connell J, Olsen E, Thiene G, Goodwin J, Gyarfas I, Martin I, Nordet P. Report of the 1995 World Health Organization/International Society and Federation of Cardiology Task Force on the Definition and Classification of Cardiomyopathies. Circulation 1996;93:841–842. Amado LC, Gerber BL, Gupta SN, Rettmann DW, Szarf G, Schock R, Nasir K, Kraitchman DL, Lima JA. Accurate and objective infarct sizing by contrastenhanced magnetic resonance imaging in a canine myocardial infarction model. J Am Coll Cardiol 2004;44:2383–2389. Kim RJ, Fieno DS, Parrish TB, Harris K, Chen EL, Simonetti O, Bundy J, Finn JP, Klocke FJ, Judd RM. Relationship of MRI delayed contrast enhancement to irreversible injury, infarct age, and contractile function. Circulation 1999;100: 1992–2002. Friedrich MG, Sechtem U, Schulz-Menger J, et al. Cardiovascular magnetic resonance in myocarditis: a JACC White Paper. J Am Coll Cardiol 2009;53: 1475–1487. Iles L, Pfluger H, Lefkovits L, Butler MJ, Kistler PM, Kaye DM, Taylor AJ. Myocardial fibrosis predicts appropriate device therapy in patients with implantable cardioverter-defibrillators for primary prevention of sudden cardiac death. J Am Coll Cardiol 2011;57:821–828. Neilan TG, Coelho-Filho OR, Danik SB, et al. CMR quantification of myocardial scar provides additive prognostic information in nonischemic cardiomyopathy. J Am Coll Cardiol Imaging 2013;6:944–954.
863 20. McCrohon JA, Moon JC, Prasad SK, McKenna WJ, Lorenz CH, Coats AJ, Pennell DJ. Differentiation of heart failure related to dilated cardiomyopathy and coronary artery disease using gadolinium-enhanced cardiovascular magnetic resonance. Circulation 2003;108:54–59. 21. Roberts WC, Siegel RJ, McManus BM. Idiopathic dilated cardiomyopathy: analysis of 152 necropsy patients. Am J Cardiol 1987;60:1340–1355. 22. Waller TA, Hiser WL, Capehart JE, Roberts WC. Comparison of clinical and morphologic cardiac findings in patients having cardiac transplantation for ischemic cardiomyopathy, idiopathic dilated cardiomyopathy, and dilated hypertrophic cardiomyopathy. Am J Cardiol 1998;81:884–894. 23. Gulati A, Jabbour A, Ismail TF, et al. Association of fibrosis with mortality and sudden cardiac death in patients with nonischemic dilated cardiomyopathy. JAMA 2013;309:896–908. 24. Klem I, Weinsaft JW, Bahnson TD, Hegland D, Kim HW, Hayes B, Parker MA, Judd RM, Kim RJ. Assessment of myocardial scarring improves risk stratification in patients evaluated for cardiac defibrillator implantation. J Am Coll Cardiol 2012;60:408–420. 25. Kwong RY, Chan AK, Brown KA, Chan CW, Reynolds HG, Tsang S, Davis RB. Impact of unrecognized myocardial scar detected by cardiac magnetic resonance imaging on event-free survival in patients presenting with signs or symptoms of coronary artery disease. Circulation 2006;113:2733–2743. 26. Nguyen TP, Qu Z, Weiss JN. Cardiac fibrosis and arrhythmogenesis: the road to repair is paved with perils. J Mol Cell Cardiol 2013. Epub ahead of print. 27. Gao P, Yee R, Gula L, et al. 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. Circ Cardiovasc Imaging 2012;5:448–456. 28. Anderson KP, Walker R, Urie P, Ershler PR, Lux RL, Karwandee SV. Myocardial electrical propagation in patients with idiopathic dilated cardiomyopathy. J Clin Invest 1993;92:122–140. 29. Fenoglio JJ Jr, Pham TD, Harken AH, Horowitz LN, Josephson ME, Wit AL. Recurrent sustained ventricular tachycardia: structure and ultrastructure of subendocardial regions in which tachycardia originates. Circulation 1983;68: 518–533. 30. Epstein AE, DiMarco JP, Ellenbogen KA, et al. ACC/AHA/HRS 2008 Guidelines for Device-Based Therapy of Cardiac Rhythm Abnormalities. J Am Coll Cardiol 2008;51:e1–e62. 31. Roes SD, Borleffs CJ, van der Geest RJ, Westenberg JJ, Marsan NA, Kaandorp TA, Reiber JH, Zeppenfeld K, Lamb HJ, de Roos A, Schalij MJ, Bax JJ. Infarct tissue heterogeneity assessed with contrast-enhanced MRI predicts spontaneous ventricular arrhythmia in patients with ischemic cardiomyopathy and implantable cardioverter-defibrillator. Circ Cardiovasc Imaging 2009;2:183–190. 32. AlJaroudi WA, Flamm SD, Saliba W, Wilkoff BL, Kwon D. Role of CMR imaging in risk stratification for sudden cardiac death. JACC Cardiovasc Imaging 2013;6:392–406.
block (HR 2.43; 95% CI 1.01–5.41; P ¼ .048) were found to be associated with arrhythmias. In multivariable analysis, the presence of LGE was the only independent predictor of arrhythmias (HR 3.8; 95% CI 1.3–10.4; P ¼ .01). CONCLUSIONS LV-LGE is a powerful and independent predictor of malignant arrhythmic prognosis, while its amount and distribution do not provide additional prognostic value. Contrast-enhanced cardiac magnetic resonance may contribute to identify candidates for ICD therapy not fulfilling the current criteria based on left ventricular ejection fraction. KEYWORDS Risk stratification; Dilated cardiomyopathy; Cardiac magnetic resonance; Arrhythmias; Heart failure; Late gadolinium enhancement ABBREVIATIONS CE-CMR ¼ contrast-enhanced cardiac magnetic resonance; CI ¼ confidence interval; CMR ¼ cardiac magnetic resonance; ECG ¼ electrocardiogram; EF ¼ ejection fraction; HR ¼ hazard ratio; ICD ¼ implantable cardioveter-defibrillator; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular; NIDC ¼ nonischemic dilated cardiomyopathy; SCD ¼ sudden cardiac death; VF ¼ ventricular fibrillation; VT ¼ ventricular tachycardia; WHO ¼ World Health Organization (Heart Rhythm 2014;11:856–863) I 2014 Heart Rhythm Society. All rights reserved.
Introduction This work was supported in part by research grant TRANSACSTPD113ZKJ from the University of Padova. Address reprint requests and correspondence: Dr Domenico Corrado, Inherited Arrhythmogenic Cardiomyopathy Unit, Department of Cardiac, Thoracic and Vascular Sciences, University of Padova, Via Giustiniani 2, 35121 Padova, Italy. E-mail address: [email protected].
1547-5271/$-see front matter B 2014 Heart Rhythm Society. All rights reserved.
Nonischemic dilated cardiomyopathy (NIDC) is a common heart muscle disorder, which is responsible for approximately one-third of the heart failure and is associated with significant morbidity and mortality.1,2 Sudden cardiac death (SCD) is an important mode of death accounting for 30% of all fatalities in patients with NIDC.3 Current risk stratification for SCD and http://dx.doi.org/10.1016/j.hrthm.2014.01.014
Perazzolo Marra et al
LV-LGE by CE-CMR in NIDC
indication to implant an implantable cardioveter-defibrillator (ICD) rely on reduction of left ventricular (LV) ejection fraction (EF).4,5 However, LVEF is a recognized poor method for arrhythmic risk stratification, as evidenced by the finding that the majority of patient with ICD implantation never receive appropriate therapy.6 The most important reason is that EF represents a global assessment of LV systolic function, but it does not necessarily correlate with the myocardial substrate needed for ventricular tachycardia (VT)/ventricular fibrillation (VF). Advances in our understanding of pathobiologic processes underlying both ventricular function impairment and electrical instability may offer novel predictors of arrhythmic outcome in NIDC. Clinicopathologic correlation studies have showed that the presence of myocardial fibrosis may provide a substrate for malignant ventricular arrhythmias and SCD.7,8 The clinical assessment of myocardial fibrosis has traditionally been based on endomyocardial biopsy, with inherent limitations related to the invasive approach, the small myocardial sample size, and the possible complications. Cardiac magnetic resonance (CMR) with late gadolinium enhancement (LGE) is an emerging noninvasive tool that has the ability to accurately identify and quantify ventricular myocardial fibrosis.9 Previous CMR studies in patients with NIDC have demonstrated that ventricular LGE heralds an adverse prognosis with regard to heart failure– related hospitalization and death.10–13 In these studies, arrhythmic outcome was most often evaluated as a secondary end point and a statistically significant direct association between LV-LGE and arrhythmic risk in NIDC has not been demonstrated consistently. Moreover, the existence of a dose-response relationship between the amount of nonischemic myocardial fibrosis and rate of malignant arrhythmic events remains to be assessed. The present investigation was primarily designed to assess the value of the presence and extent of myocardial fibrosis as evidenced by contrast-enhanced cardiac magnetic resonance (CE-CMR) for predicting major arrhythmic events and SCD in a consecutive series of patients with a diagnosis of NIDC according to World Health Organization (WHO) criteria.14 We rigorously characterized the disease pathogenesis by systematic coronary angiography for the exclusion of ischemic heart disease and our study population included the entire spectrum of disease phenotype ranging from early to advanced hemodynamically compromised stages.
Methods Study population and design The present study included a consecutive series of patients referred to the Heart Failure and Heart Transplantation Unit of the University Hospital of Padua for unexplained LV dilatation and dysfunction. All patients underwent detailed clinical evaluation, including 12-lead electrocardiogram (ECG), 2-dimensional transthoracic echocardiography, CECMR, and coronary angiography. The diagnosis of NIDC
857 was based on the 1995 WHO/International Society and Federation of Cardiology criteria.14 To be enrolled, patients had to have (1) a depressed LV systolic function (LVEF o50%) on a non-CMR study; (2) an angiographic study showing the absence of flow-limiting coronary artery disease (defined as Z50% luminal stenosis on coronary angiography); and (3) the absence of either valvular or hypertensive heart disease and congenital heart abnormalities. Exclusion criteria were as follows: recent onset of heart failure (o1 month); diagnosis of hypertrophic cardiomyopathy, restrictive cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, suspected infiltrative heart disease, or other specific cardiomyopathies; hemodynamic unstable conditions; contraindication to CMR (claustrophobia, pacemaker, ICD, metallic clips, atrial fibrillation, severe obesity preventing the patient from entering the scanner bore, and pregnancy); and chronic renal failure with an estimated glomerular filtration rate of o30 mL/min. The study was approved by the institutional review board, and all patients gave their informed consent.
Follow-up The follow-up data were obtained during regular outpatient visits at 3-, 6-, and 12-month intervals. All events were adjudicated by the consensus of an independent committee blinded to the CMR results. A 24-hour ECG Holter recordings were obtained for each patient without ICD every 6 months; the analysis of ECG Holter recordings focused on VT, either nonsustained or sustained. Routine ICD interrogation (at 1, 3, and 6 months and every 6 months thereafter) and ECG recordings at the time of symptoms were used to document the occurrence of spontaneous VT during follow-up in all patients. All interrogations were identified and adjudicated by 2 electrophysiologists (D.C. and F.M.) blinded to other clinical data including CMR analysis. The study outcome was the index combined end point of major arrhythmic events such as SCD, cardiac arrest due to VF, sustained VT, or appropriate ICD intervention. Sudden death was defined as any natural death occurring instantaneously or within 1 hour from the onset of symptoms. Sustained VT was defined as tachycardia originating in the ventricle with rate 4100 beats/min and lasting 430 seconds or requiring an intervention for termination. Nonsustained VT was defined as Z3 consecutive ventricular premature beats with a rate of 4100 beats/min that lasted o30 seconds during 24-hour Holter monitoring. Appropriate ICD intervention was defined as a device shock or antitachycardia overdrive pacing delivered in response to a ventricular tachyarrhythmia and documented by stored intracardiac ECG data. VF or ventricular flutter was defined as irregular or regular tachycardia with regard to polarity, amplitude, morphology, and sequence of intracardiac electrograms, with a mean cycle length of r240 ms (ie, Z250 beats/min). VT was defined as regular tachycardia with a mean cycle length of 4240 ms. For each patient, rate
858 cutoff criteria for antitachycardia pacing or shock discharge were programmed to avoid inappropriate ICD interventions and to guarantee successful therapy for ventricular tachyarrhythmias. ICD was routinely programmed to include a monitoring zone that identified VT with a rate of 4160 beats/min. Patients received a conventional device programming, with a 2.5-second delay at 170–199 beats/min and a 1.0-second delay at Z200 beats/min.
CE-CMR CMR was performed by using a comprehensive dedicated protocol including postcontrast sequences (see Online Supplemental Methods). The presence, location, and extent of LGE were independently assessed by 2 experienced observers (M.P.M. and F.C.) who were blinded to patient clinical data and outcome; ambiguous cases were reviewed by a third expert. To exclude artifact, LGE was deemed present only if visible in 2 orthogonal views (short-axis and long-axis views). Myocardial LGE was quantified by a semiautomatic detection and LGE mass (in grams) expressed as a percentage of total LV mass, according to previously validated methods.15,16 LGE was quantified by using a signal intensity threshold of 42 SD above a remote reference region.10–12,17–19 The pattern of LGE distribution was characterized as either epicardial, mid-wall, or patchy/junctional.12,20 If more than 1 pattern was present, the distribution was characterized on the basis of the predominant pattern. Patients diagnosed as having NIDC on the basis of coronary angiography, who showed at CE-CMR an “ischemic LGE pattern,” either subendocardial or transmural (most likely due to a transient occlusive coronary event, thrombosis or vasospasm, or embolization from nonstenotic but unstable coronary plaque), were excluded from the final analysis.
Heart Rhythm, Vol 11, No 5, May 2014 correlation of traditional clinical predictors of arrhythmic risk in NIDC with the index arrhythmic end point during follow-up was determined by means of the Cox regression analysis. Variables with a P value of o.15 in the univariate analysis were integrated into multivariable analysis to estimate the hazard ratio (HR) and to identify independent predictors of major arrhythmic events. The Cox model was used to calculate the relation between the amount of LGE and hazard ratios. HRs and confidence intervals (CIs) are presented both in univariate and in multivariable analysis. Linear regression was used to assess the correlation between the 2 independent observers performing LGE planimetry. A P value of o.05 was considered statistically significant. Statistics were analyzed with SPSS version 19 (SPSS Inc, Chicago, IL).
Results Baseline characteristics Of a total of 276 patients initially referred to our department for clinical evaluation of LV dilatation and dysfunction, 27 refused or had contraindication to CE-CMR. One hundred eight patients were excluded from the study because of a diagnosis of “ischemic” dilated cardiomyopathy on the basis of angiographic demonstration of flow-limiting coronary artery disease (n = 102), CE-CMR findings of “ischemic pattern” of LGE in the absence of coronary artery stenosis (n = 4), or an LV dilatation/dysfunction “out of proportion” to concomitant single vessel coronary artery disease (n = 2). Four patients were lost over the follow-up owing to no contact. The remaining 137 patients (median age 49 years; 108 men) fulfilled the criteria for NIDC and constituted the final study population. Baselines clinical characteristics of the enrolled patients are reported in Online Supplemental Table 1.
LV-LGE Statistical analysis Data are expressed as mean ⫾ SD and as median with 25–75 percentiles for normally distributed and skewed variables, respectively. Normal distribution was assessed by using the Shapiro-Wilk test. Categorical differences between groups were evaluated by using the χ2 test or the Fisher exact test, as appropriate. Paired and unpaired t tests were used to compare normally distributed continuous variables obtained, respectively, from the same patient and different patients; paired and unpaired rank sum tests were used for skewed continuous variables. Kaplan-Meier analysis was used to estimate the survival distributions of the arrhythmic end point and to show the differences in survival between patients with vs without LGE. The start of follow-up was defined as the date of the initial CMR. Patients were censored at the time of their first event or the time of their last clinical follow-up. The mean event rate per year was evaluated by the number of events occurring during the follow-up divided by the number of patients multiplied by the average duration of follow-up. The
In postcontrast sequences, LV-LGE was identified in 76 (55.5%) patients. The LGE pattern of distribution was epicardial in 23 (30.3%) patients, mid-wall in 46 (60.5%) patients, or patchy/junctional in 7 (9.2%) patients. In 8 patients, the predominant epicardial pattern was associated with either septal mid-wall (n ¼ 5) or junctional (n ¼ 3) LGE patterns. The median extent of LGE was 9% (4%–15%) of LV mass. The representative cases of LV-LGE regional distribution patterns and histopathologic correlations are, respectively, shown in Figures 1A–1C and Online Supplemental Figure 1. The presence of LV-LGE was unrelated to sex, age, New York Heart Association functional class, LVEF, or any other clinical baseline characteristic (see Online Supplemental Table 1). No significant correlation between the presence of LV-LGE and any of the CMR functional parameters was found. Linear regression analysis showed a high correlation between the 2 observers for planimetry of %LGE (r ¼ .94; P o .01).
Perazzolo Marra et al
LV-LGE by CE-CMR in NIDC
859
Figure 1 Representative cases of myocardial fibrosis on CE-CMR underlying major arrhythmic events during follow-up. A: Postcontrast short-axis view of a patient with focal mid-wall anteroseptal LGE who experienced aborted SCD owing to ventricular fibrillation, which was interrupted by ICD shock (D). B: Circumferential mid-wall LGE distribution on postcontrast short-axis view in a patient who was hospitalized for syncope, without ICD, and had in-hospital VF that required external defibrillation (E). C: Epicardial LGE distributing to anterior, septal, and inferior LV walls on postcontrast short-axis view in a patient who experienced a sustained ventricular tachycardia (F). CE-CMR ¼ contrast-enhanced cardiac magnetic resonance; ICD ¼ implantable cardioveter-defibrillator; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular.
Arrhythmic outcome During a median follow-up of 3 years (range 31 days to 9.6 years), a total of 69 (50.3%) patients underwent ICD implantation, either for primary (N ¼ 62) or secondary (N ¼ 7) prevention, which was combined with cardiac resynchronization therapy in 14 (20%) patients. The primary arrhythmic end point occurred in 22 (16.1%) patients, with a 5.4% annual rate of major arrhythmic events (see Online Supplemental Table 2). Eight (5.8%) patients had an episode of sustained VT, 9 (6.6%) had Z1 appropriate ICD interventions, either against VF (n ¼ 5) or VT (n ¼ 4), and 3 (2.2%) experienced aborted SCD and 2 (1.5%) died suddenly. Representative examples of major arrhythmic events that occurred in patients with LV-LGE are shown in Figures 1D–1F. Table 1 summarizes the clinical characteristics of patients with or without major arrhythmic events during follow-up. Patients who experienced arrhythmic events significantly more often showed a left bundle branch block ECG pattern (54.5% vs 28.9%; P ¼ .02) and the presence of LV-LGE on CE-CMR (77% vs 51%; P ¼ .04). Figures 2A and 2B show Kaplan-Meier analysis of survival from the index combined arrhythmic end point stratified by the presence of LGE and LVEF. Overall, the annual event rate was 7% in patients with LV-LGE vs 3.3% in those without LGE on CE-CMR (log-rank P ¼ .002)
(Figure 2A). The annual event rate was 5.1% in patients with reduced EF (r35%) vs 2.4% in those with EF 435% (logrank, P ¼ .467; Figure 2B). The arrhythmic event-free survival did not differ between patients whose %LV-LGE mass measured below the median and patients with values above the median (log-rank, P ¼ .295; Figure 3). The prevalence of major arrhythmic events was not associated with a specific LGE pattern. During follow-up, 27 of 137 (19.7%) patients experienced an episode of heart failure, requiring intravenous administration of diuretics and inotropes (13 in the LGE-negative group and 14 in the LGE-positive group) and 12 of 137 (8.8%) had non-SCD or heart transplantation (6 in the LGEnegative group and 6 in the LGE-positive group). Of the 68 patients undergoing serial 24-hour ECG Holter monitoring during follow-up, 30 (44%) had Z1 episode of nonsustained VT. There was an association of borderline statistical significance between nonsustained VT and LVLGE by CMR (P ¼ .071), while patients with (n ¼ 19) and without (n ¼ 11) nonsustained VT did not differ significantly with regard to LGE amount and pattern.
Predictors of arrhythmic events The results of univariate and multivariate analyses for predictors of major arrhythmic events during follow-up are summarized in Table 2. Univariate predictors of events were the presence of left bundle branch block on ECG and the
860
Heart Rhythm, Vol 11, No 5, May 2014
Table 1 Characteristics of patients with and without arrhythmic events during follow-up Characteristic Age (y) Sex: male, n (%) NYHA class, n (%) I II III IV Medications, n (%) ACEI or ARB therapy β-Blockers Spironolactone Diuretic Amiodarone Digoxin Statin LBBB, n (%) EDV (mL/m2) EF (%) EF o35%, n (%) LGE, n (%)
No arrhythmic events Arrhythmic events (n ¼ 115 [84%]) (n ¼ 22 [16%]) 47 (37–60) 90 (78.3)
59 (43–70) 18 (81.8)
29 37 45 4
(25) (32) (39) (4)
4 (18) 11 (50) 7 (32) 0 (0)
102 (87)
18 (82)
88 51 87 22 14 28 33 109 33 67 59
(76) (44) (76) (19) (12) (24) (28.9) (87–140) (28–40) (58.3) (51)
19 (86) 11 (50) 20 (91) 3 (14) 4 (18) 9 (41) 12 (54.5) 123 (105–143) 30 (29–40) 15 (68.2) 17 (77)
Continuous values are expressed as median (25th-75th percentiles). ACEI ¼ angiotensin-converting enzyme inhibitor; ARB ¼ angiotensin receptor blocker; EF ¼ echocardiographic ejection fraction; EDV ¼ echocardiographic end-diastolic volume; LBBB ¼ left bundle branch block; LGE ¼ late gadolinium enhancement; NIDC ¼ nonischemic dilated cardiomyopathy; NYHA ¼ New York Heart Association.
presence of LV-LGE on CE-CMR. Neither LVEF (HR 0.99; 95% CI 0.94–1.04; P ¼ .62) nor the percentage of LV-LGE extent (HR 1.04; 95% CI 0.98–1.09; P ¼ .18) significantly predicted arrhythmic outcome (Table 2). In multivariable analysis, the presence of LV-LGE remained the only independent predictor of arrhythmic events (HR 3.8; 95% CI 1.3–10.4; P ¼ .01).
Figure 3 Kaplan-Meier analysis of freedom from major arrhythmic events stratified by the extent of LGE (median value). LGE ¼ late gadolinium enhancement.
Discussion Our study was designed to evaluate the value of the presence and extent of LV-LGE by CE-CMR for predicting major arrhythmic events and SCD in a large population of patients with NIDC diagnosed according to the WHO criteria.14 By study design, the hemodynamic patient profile encompassed the entire disease spectrum including patients with severe LV dysfunction who meet criteria for ICD implantation and those with mild or moderate disease forms who did not. Dilated cardiomyopathy was classified as “nonischemic” by coronary angiography findings in association with the LGE distribution pattern, which is consistent with current clinical practice and allows the most accurate etiology definition for clinical research purposes. The major study results are as follows: (1) the presence of LV-LGE is a powerful and independent predictor of arrhythmic
Figure 2 Kaplan-Meier analysis of freedom from major arrhythmic events stratified by the presence of LGE (A) and reduced LV ejection fraction (B) on CECMR. CE-CMR ¼ contrast-enhanced cardiac magnetic resonance; EF ¼ ejection fraction; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular.
Perazzolo Marra et al Table 2
LV-LGE by CE-CMR in NIDC
861
Predictors of arrhythmic events during follow-up Univariate analysis
Multivariable analysis
Characteristic
HR
CI
P
Age Sex: male Family history of DCM NYHA class: I-II vs III-IV CRT LBBB LVEVD (mL/m2) LVEF (%) LGE presence LGE extent (% of LV mass)
1.02 0.89 0.39 1.34 1.06 2.34 1.01 0.98 4.17 1.04
0.99–1.04 0.3–2.6 0.11–1.41 0.55–3.29 0.98–1.08 1.01–5.41 0.99–1.02 0.94–1.08 1.56–11.2 0.98–1.09
0.22 .80 .38 .524 .51 .048 .48 .63 .005 .18
HR
CI
P
2.3
0.9–5.1
.06
3.8
1.3–10.4
.01
CI ¼ confidence interval; CRT ¼ cardiac resynchronization therapy; DCM ¼ dilated cardiomyopathy; EF ¼ ejection fraction; HR ¼ hazard ratio; LBBB ¼ left bundle branch block; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular; LVEDV ¼ left ventricular end-diastolic volume; NYHA ¼ New York Heart Association.
outcome during a median follow-up period of 3 years; (2) the presence of LV-LGE appears to be superior to traditional LVEF r35% in predicting major arrhythmic events and SCD; (3) the absence of LGE characterizes a low arrhythmic subgroup of patients with no SCD events during follow-up.
Previous studies See Online Supplemental Discussion.
Arrhythmic prognostic value of LGE In the present study, LGE was found in 76 of 137 (55.5%) patients with NIDC, a figure comparable to that of previous CMR studies which reported a prevalence of 30%–65% and similar to that of pathologic studies on NIDC hearts in which macroscopic LV scarring was detected in up to half of the cases.21,22 The present CMR study is in agreement with previous investigations, which consistently demonstrated the prognostic value of LGE in patients with NIDC.10–13 Our study results significantly showed that LGE on CE-CMR identifies patients at increased risk of major arrhythmic events and SCD during a long-term follow-up. Although in univariate Cox regression analysis, both abnormal CE-CMR and left bundle branch block were significant predictors for the composite arrhythmic end point, yielding an HR of 3.8 and 3.4, respectively, the presence of LGE remained the only independent predictor of adverse arrhythmic outcome in multivariable analysis (HR 3.8). Among patients with LGE, we found a relatively small amount of LGE with a median percentage of 9%. Moreover, in our study a greater extent of LGE did not appear to confer a higher arrhythmic risk. The annual rates of arrhythmic events did not differ significantly between patients whose % LV-LGE mass measured below the median (5.7%/y) compared with those with values above the median (8.6%/y). This finding diverged from that of previous studies, which reported a prognostic significance of LV-LGE size, with the arrhythmic risk being proportional to the increasing percentage of LV scar.10,16 In the Gulati study,23 the LGE size was
an independent predictor of malignant arrhythmic events with an HR of 1.08 per 1% increase. This discrepancy may be explained by the difference in study population, considering that we included only patients with angiographically proven NIDC while in the Gulati study coronary artery disease was not rigorously ruled out by coronary angiography. The fact that in the latter study a proportion of enrolled patients could have an ischemic myocardial scar is relevant if one considers that previous CMR studies actually showed a direct association between the amount of postinfarction scar and arrhythmic risk.24–28 In this regard, Klem et al24 found a significant increase in the risk when the LGE size exceeded 5% of LV mass, while Kwong et al25 reported a significant risk increase even for the lowest tertile of LVLGE amount (1.4%). Thus, we can speculate that the lack of a dose-response relationship between LGE and arrhythmic risk in our study may be explained by the different pathophysiologic significance of myocardial fibrosis in ischemic dilated cardiomyopathy vs NIDC. While ventricular tachyarrhythmias are mechanistically (ie, scar-related macroreentry) and, thus, quantitatively related to postinfarction scar, LV-LGE in NIDC may not be the sole source of ventricular arrhythmias. We can speculate that compact myocardial fibrosis, as evidenced by LGE, is part of a more complex fibrotic process also involving patchy and interstitial fibrosis, which may fall under the resolution power of traditional LGE (Online Supplemental Figure 1). Interstitial fibrosis, in which short collagen septa are interspersed with myocardial fibers, may play an important arrhythmogenic role in the setting of nonischemic heart disease by predisposing to slow conduction and leading to unidirectional block.26 The emerging T1 mapping sequence technique will offer the potential to improve the clinical assessment of interstitial myocardial fibrosis and to refine the prognostic value of LGE. Although we observed significantly higher rates of major arrhythmic events in LV-LGE– positive patients with NIDC, the absence of LV-LGE did not guarantee an uneventful arrhythmic outcome, given that 5 (8%) LV-LGE–negative patients experienced arrhythmic events such as VT (n ¼ 3
862 [4.9%]) and appropriate ICD intervention (n ¼ 2 [3.2%]). These findings may be explained by various mechanisms of ventricular tachyarrhythmias in NIDC, partly unrelated to myocardial fibrosis, which include functional bundle branch/ interfascicular macroreentry, anisotropic conduction, or spatial dispersion of ventricular repolarization.27–29 It is noteworthy that in our NIDC cohort, both aborted SCD and SCD occurred exclusively in the subgroup of patients with LGE, suggesting that VF mostly occurs in the context of scarring myocardium.
Study limitations Because endomyocardial biopsy was not performed systematically, an infiltrative disease could not be excluded definitively at the time of the study enrollment; however, no patients had a CMR pattern consistent with either infiltrative or storage myocardial disease. In the setting of ischemic heart disease, the “infarct gray zone” on CE-CMR has been demonstrated to reflect the infarct tissue heterogeneity and predict the occurrence of spontaneous ventricular arrhythmias.30 The prevalence and prognostic meaning of scar heterogeneity in patients with NIDC was not investigated by both the previous and the present studies because of methodological limitations related to the smaller amount of myocardial fibrosis and the specific scar pattern, not comparable with that of post-infarction scar. Moreover, the quantification of hyperenhanced signals of nonischemic myocardial scar is not yet standardized.31 We quantified LGE by using a signal intensity threshold of 42 SD above a remote reference region, according to the current guidelines17 and most of the previous studies on CE-CMR in NIDC. Although relatively large areas of scarring can be identified by using the traditional LGE technique, small patches of fibrosis as well as interstitial fibrosis, which may have a potential role in arrhythmia susceptibility, cannot be detected. In the near future, the use of T1 mapping system will offer the potential to improve clinical assessment of diffuse interstitial myocardial fibrosis and to further refine the prognostic value of LGE. Finally, our finding that scar amount does not provide additional prognostic value for malignant arrhythmic events and SCD needs to be confirmed by further studies showing a broader range of LV scar.
Clinical implications An ICD implantation has been proven to significantly reduce mortality in NIDC. The current guidelines recommend ICD implantation for primary prevention in the subgroup of patients with reduced LVEF (r35%).32 However, risk assessment based on LVEF alone is not accurate enough, given that the majority of patients receiving an ICD for primary prevention never experience appropriate therapy. In this regard, EF represents a global assessment of LV systolic function and does not necessarily correlate with myocardial lesions underlying ventricular electrical instability.
Heart Rhythm, Vol 11, No 5, May 2014
Conclusions Our study showed that LV-LGE by CE-CMR is independently associated with an adverse arrhythmic prognosis in patients with NIDC. This suggests that risk stratification based on the demonstration of the myocardial scar substrate has the potential to better identify patients with NIDC at risk of SCD and to improve accuracy of indication of ICD implantation. Unlike traditional imaging techniques such echocardiography that disclose LV mechanical dysfunction (either regional or global), CE-CMR allows the detection of myocardial fibrosis that may act as a potentially arrhythmogenic myocardial substrate. This explains why the presence of LV-LGE is a stronger predictor of adverse arrhythmic outcome rather than traditional LVEF. There is a high level of concern among physicians about stratification of the risk of SCD in patients with mild to moderate NIDC who do not meet ICD implantation criteria by EF. Our study population included such patients, and the results remained statistically significant after adjustment for LVEF. Thus, the detection of myocardial scar may identify a subgroup of patients with NIDC and with a high arrhythmic risk and less severe LV dysfunction, not currently fulfilling criteria for ICD implantation. In our study, a subanalysis of the predictive value of LGE for arrhythmic outcome in the subgroup of patients with LVEF 435% was not performed because of the limited number of events. Further studies on larger population with longer follow-up are needed to confirm the arrhythmic prognostic value of CE-CMR in patients with NIDC and with early disease stage and to substantiate the indication of prophylactic ICD implantation based on LV-LGE demonstration.
Appendix Supplementary data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.hrthm. 2014.01.014.
References 1. Jefferies JL, Towbin JA. Dilated cardiomyopathy. Lancet 2010;375:752–762. 2. Maron BJ, Towbin JA, Thiene G, Antzelevitch C, Corrado D, Arnett D, Moss AJ, Seidman CE, Young JB. Contemporary definitions and classification of the cardiomyopathies. Circulation 2006;113:1807–1816. 3. Felker GM, Thompson RE, Hare JM, Hruban RH, Clemetson DE, Howard DL, Baughman KL, Kasper EK. Underlying causes and long-term survival in patients with initially unexplained cardiomyopathy. New Engl J Med 2000;342: 1077–1084. 4. Goldberger JJ, Buxton AE, Cain M, et al. Risk stratification for arrhythmic sudden cardiac death: identifying the roadblocks. Circulation 2011;123: 2423–2430. 5. Passman R, Goldberger JJ. Predicting the future: risk stratification for sudden cardiac death in patients with left ventricular dysfunction. Circulation 2012;125: 3031–3037. 6. Ezekowitz JA, Rowe BH, Dryden DM, Hooton N, Vandermeer B, Spooner C, McAlister FA. Systematic review: implantable cardioverter-defibrillators for adults with left ventricular systolic dysfunction. Ann Intern Med 2007;147: 251–262. 7. Wu TJ, Ong JJ, Hwang C, Lee JJ, Fishbein MC, Czer L, Trento A, Blanche C, Kass RM, Mandel WJ, Karagueuzian HS, Chen PS. Characteristics of wave fronts during ventricular fibrillation in human hearts with dilated
Perazzolo Marra et al
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
LV-LGE by CE-CMR in NIDC
cardiomyopathy: role of increased fibrosis in the generation of reentry. J Am Coll Cardiol 1998;32:187–196. Pogwizd SM, McKenzie JP, Cain ME. Mechanisms underlying spontaneous and induced ventricular arrhythmias in patients with idiopathic dilated cardiomyopathy. Circulation 1998;98:2404–2014 Mewton N, Liu CY, Croisille P, Bluemke D, Lima JA. Assessment of myocardial fibrosis with cardiovascular magnetic resonance. J Am Coll Cardiol 2011;57: 891–903. Assomull RG, Prasad SK, Lyne J, Smith G, Burman ED, Khan M, Sheppard MN, Poole-Wilson PA, Pennell DJ. Cardiovascular magnetic resonance, fibrosis, and prognosis in dilated cardiomyopathy. J Am Coll Cardiol 2006;48:1977–1985. Wu KC, Weiss RG, Thiemann DR, Kitagawa K, Schmidt A, Dalal D, Lai S, Bluemke DA, Gerstenblith G, Marbán E, Tomaselli GF, Lima JA. Late gadolinium enhancement by cardiovascular magnetic resonance heralds an adverse prognosis in nonischemic cardiomyopathy. J Am Coll Cardiol 2008;51: 2414–2421. Lehrke S, Lossnitzer D, Schöb M, Steen H, Merten C, Kemmling H, Pribe R, Ehlermann P, Zugck C, Korosoglou G, Giannitsis E, Katus HA. Use of cardiovascular magnetic resonance for risk stratification in chronic heart failure: prognostic value of late gadolinium enhancement in patients with non-ischaemic dilated cardiomyopathy. Heart 2011;97:727–732. Masci PG, Barison A, Aquaro GD, Pingitore A, Mariotti R, Balbarini A, Passino C, Lombardi M, Emdin M. Myocardial delayed enhancement in paucisymptomatic nonischemic dilated cardiomyopathy. Int J Cardiol 2012;157:43–47. Richardson P, McKenna W, Bristow M, Maisch B, Mautner B, O'Connell J, Olsen E, Thiene G, Goodwin J, Gyarfas I, Martin I, Nordet P. Report of the 1995 World Health Organization/International Society and Federation of Cardiology Task Force on the Definition and Classification of Cardiomyopathies. Circulation 1996;93:841–842. Amado LC, Gerber BL, Gupta SN, Rettmann DW, Szarf G, Schock R, Nasir K, Kraitchman DL, Lima JA. Accurate and objective infarct sizing by contrastenhanced magnetic resonance imaging in a canine myocardial infarction model. J Am Coll Cardiol 2004;44:2383–2389. Kim RJ, Fieno DS, Parrish TB, Harris K, Chen EL, Simonetti O, Bundy J, Finn JP, Klocke FJ, Judd RM. Relationship of MRI delayed contrast enhancement to irreversible injury, infarct age, and contractile function. Circulation 1999;100: 1992–2002. Friedrich MG, Sechtem U, Schulz-Menger J, et al. Cardiovascular magnetic resonance in myocarditis: a JACC White Paper. J Am Coll Cardiol 2009;53: 1475–1487. Iles L, Pfluger H, Lefkovits L, Butler MJ, Kistler PM, Kaye DM, Taylor AJ. Myocardial fibrosis predicts appropriate device therapy in patients with implantable cardioverter-defibrillators for primary prevention of sudden cardiac death. J Am Coll Cardiol 2011;57:821–828. Neilan TG, Coelho-Filho OR, Danik SB, et al. CMR quantification of myocardial scar provides additive prognostic information in nonischemic cardiomyopathy. J Am Coll Cardiol Imaging 2013;6:944–954.
863 20. McCrohon JA, Moon JC, Prasad SK, McKenna WJ, Lorenz CH, Coats AJ, Pennell DJ. Differentiation of heart failure related to dilated cardiomyopathy and coronary artery disease using gadolinium-enhanced cardiovascular magnetic resonance. Circulation 2003;108:54–59. 21. Roberts WC, Siegel RJ, McManus BM. Idiopathic dilated cardiomyopathy: analysis of 152 necropsy patients. Am J Cardiol 1987;60:1340–1355. 22. Waller TA, Hiser WL, Capehart JE, Roberts WC. Comparison of clinical and morphologic cardiac findings in patients having cardiac transplantation for ischemic cardiomyopathy, idiopathic dilated cardiomyopathy, and dilated hypertrophic cardiomyopathy. Am J Cardiol 1998;81:884–894. 23. Gulati A, Jabbour A, Ismail TF, et al. Association of fibrosis with mortality and sudden cardiac death in patients with nonischemic dilated cardiomyopathy. JAMA 2013;309:896–908. 24. Klem I, Weinsaft JW, Bahnson TD, Hegland D, Kim HW, Hayes B, Parker MA, Judd RM, Kim RJ. Assessment of myocardial scarring improves risk stratification in patients evaluated for cardiac defibrillator implantation. J Am Coll Cardiol 2012;60:408–420. 25. Kwong RY, Chan AK, Brown KA, Chan CW, Reynolds HG, Tsang S, Davis RB. Impact of unrecognized myocardial scar detected by cardiac magnetic resonance imaging on event-free survival in patients presenting with signs or symptoms of coronary artery disease. Circulation 2006;113:2733–2743. 26. Nguyen TP, Qu Z, Weiss JN. Cardiac fibrosis and arrhythmogenesis: the road to repair is paved with perils. J Mol Cell Cardiol 2013. Epub ahead of print. 27. Gao P, Yee R, Gula L, et al. 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. Circ Cardiovasc Imaging 2012;5:448–456. 28. Anderson KP, Walker R, Urie P, Ershler PR, Lux RL, Karwandee SV. Myocardial electrical propagation in patients with idiopathic dilated cardiomyopathy. J Clin Invest 1993;92:122–140. 29. Fenoglio JJ Jr, Pham TD, Harken AH, Horowitz LN, Josephson ME, Wit AL. Recurrent sustained ventricular tachycardia: structure and ultrastructure of subendocardial regions in which tachycardia originates. Circulation 1983;68: 518–533. 30. Epstein AE, DiMarco JP, Ellenbogen KA, et al. ACC/AHA/HRS 2008 Guidelines for Device-Based Therapy of Cardiac Rhythm Abnormalities. J Am Coll Cardiol 2008;51:e1–e62. 31. Roes SD, Borleffs CJ, van der Geest RJ, Westenberg JJ, Marsan NA, Kaandorp TA, Reiber JH, Zeppenfeld K, Lamb HJ, de Roos A, Schalij MJ, Bax JJ. Infarct tissue heterogeneity assessed with contrast-enhanced MRI predicts spontaneous ventricular arrhythmia in patients with ischemic cardiomyopathy and implantable cardioverter-defibrillator. Circ Cardiovasc Imaging 2009;2:183–190. 32. AlJaroudi WA, Flamm SD, Saliba W, Wilkoff BL, Kwon D. Role of CMR imaging in risk stratification for sudden cardiac death. JACC Cardiovasc Imaging 2013;6:392–406.