TYPE Brief Research Report PUBLISHED 03 February 2023 DOI 10.3389/fcvm.2023.991307 Cardiac magnetic resonance OPEN ACCESS outperforms echocardiography to EDITED BY Gabor Duray, Hungarian Defence Forces, Hungary predict subsequent implantable REVIEWED BY Maria Luce Caputo, cardioverter defibrillator therapies Ospedale Regionale di Lugano, Switzerland Giovanni Peretto, San Raffaele Hospital (IRCCS), Italy in ST-segment elevation *CORRESPONDENCE Vicente Bodi vicente.bodi@uv.es myocardial infarction patients † Theseauthors have contributed equally to this work and share first authorship Víctor Marcos-Garcés1,2† , Nerea Perez2† , Jose Gavara2,3 , SPECIALTY SECTION Maria P. Lopez-Lereu4 , Jose V. Monmeneu4 , Cesar Rios-Navarro2 , This article was submitted to Elena de Dios5 , Hector Merenciano-González1,2 , Cardiac Rhythmology, a section of the journal Ana Gabaldon-Pérez1,2 , Ángel Ferrero-De-Loma-Osorio1 , Frontiers in Cardiovascular Medicine Ángel Martínez-Brotons1 , Lourdes Bondanza1 , RECEIVED 12July 2022 ACCEPTED 16 January 2023 Juan Miguel Sánchez-Gómez1 , Cristina Albiach1 , Julio Nunez1,2,5,6 , PUBLISHED 03 February 2023 Antoni Bayés-Genís6,7,8 , Francisco J. Chorro1,2,5,6 , CITATION Marcos-Garcés V, Perez N, Gavara J, Ricardo Ruiz-Granell1 and Vicente Bodi1,2,5,6* Lopez-Lereu MP, Monmeneu JV, 1 Department of Cardiology, Hospital Clínico Universitario de Valencia, Valencia, Spain, 2 INCLIVA Health Rios-Navarro C, de Dios E, Research Institute, Valencia, Spain, 3 Center for Biomaterials and Tissue Engineering, Universitat Politècnica Merenciano-González H, Gabaldon-Pérez A, de València, Valencia, Spain, 4 Cardiovascular Magnetic Resonance Unit, ASCIRES Biomedical Group, Ferrero-De-Loma-Osorio Á, Valencia, Spain, 5 Faculty of Medicine and Odontology, University of Valencia, Valencia, Spain, 6 Centro de Martínez-Brotons Á, Bondanza L, Investigación Biomédica en Red de Enfermedades Cardiovasculares (CIBERCV), Madrid, Spain, 7 Cardiology Sánchez-Gómez JM, Albiach C, Nunez J, Department and Heart Failure Unit, Hospital Universitari Germans Trias i Pujol, Badalona, Spain, 8 Department Bayés-Genís A, Chorro FJ, Ruiz-Granell R and of Medicine, Universitat Autònoma de Barcelona, Barcelona, Spain Bodi V (2023) Cardiac magnetic resonance outperforms echocardiography to predict subsequent implantable cardioverter defibrillator therapies in ST-segment elevation myocardial infarction patients. Front. Cardiovasc. Med. 10:991307. Background: Implantable cardioverter defibrillators (ICD) are effective as a primary doi: 10.3389/fcvm.2023.991307 prevention measure of ventricular tachyarrhythmias in patients with ST-segment COPYRIGHT elevation myocardial infarction (STEMI) and depressed left ventricular ejection © 2023 Marcos-Garcés, Perez, Gavara, fraction (LVEF). The implications of using cardiac magnetic resonance (CMR) instead Lopez-Lereu, Monmeneu, Rios-Navarro, de Dios, Merenciano-González, Gabaldon-Pérez, of echocardiography (Echo) to assess LVEF prior to the indication of ICD in this setting Ferrero-De-Loma-Osorio, Martínez-Brotons, are unknown. Bondanza, Sánchez-Gómez, Albiach, Nunez, Bayés-Genís, Chorro, Ruiz-Granell and Bodi. Materials and methods: We evaluated 52 STEMI patients (56.6 ± 11 years, 88.5% This is an open-access article distributed under the terms of the Creative Commons Attribution male) treated with ICD in primary prevention who underwent echocardiography License (CC BY). The use, distribution or and CMR prior to ICD implantation. ICD implantation was indicated based on the reproduction in other forums is permitted, provided the original author(s) and the presence of heart failure and depressed LVEF (≤ 35%) by echocardiography, CMR, or copyright owner(s) are credited and that the both. Prediction of ICD therapies (ICD-T) during follow-up by echocardiography and original publication in this journal is cited, in CMR before ICD implantation was assessed. accordance with accepted academic practice. No use, distribution or reproduction is Results: Compared to echocardiography, LVEF was lower by cardiac CMR (30.2 ± 9% permitted which does not comply with these terms. vs. 37.4 ± 7.6%, p < 0.001). LVEF ≤ 35% was detected in 24 patients (46.2%) by Echo and in 42 (80.7%) by CMR. During a mean follow-up of 6.1 ± 4.2 years, 10 patients received appropriate ICD-T (3.16 ICD-T per 100 person-years): 5 direct shocks to treat very fast ventricular tachycardia or ventricular fibrillation, 3 effective antitachycardia pacing (ATP) for treatment of ventricular tachycardia, and 2 Frontiers in Cardiovascular Medicine 01 frontiersin.org Marcos-Garcés et al. 10.3389/fcvm.2023.991307 ineffective ATP followed by shock to treat ventricular tachycardia. Echo-LVEF ≤ 35% correctly predicted ICD-T in 4/10 (40%) patients and CMR-LVEF ≤ 35% in 10/10 (100%) patients. CMR-LVEF improved on Echo-LVEF for predicting ICD-T (area under the curve: 0.76 vs. 0.48, p = 0.04). Conclusion: In STEMI patients treated with ICD, assessment of LVEF by CMR outperforms Echo-LVEF to predict the subsequent use of appropriate ICD therapies. KEYWORDS myocardial infarction, implantable cardioverter-defibrillator, cardiac magnetic resonance, ventricular tachyarrhythmias, left ventricular ejection fraction 1. Introduction 2. Materials and methods Revolutionary advances in treatment of patients presenting with 2.1. Study group ST-segment elevation acute myocardial infarction (STEMI) during recent decades has led to a spectacular improvement in prognosis This study was derived from an ongoing, single-center (1). However, despite optimized medical therapy, and in most cases prospective registry including all patients discharged for a first revascularization, the risk of sudden cardiac death is substantial reperfused STEMI between 2004 and 2021 who were treated with (2, 3). primary percutaneous coronary intervention (pPCI) and followed in Left ventricular ejection fraction (LVEF) is the cornerstone for a specific outpatient clinic in our hospital. Clinical management was non-invasive risk stratification after STEMI and the main parameter as recommended in specific STEMI guidelines (5). Patient informed to select patients to undergo prophylactic implantable cardioverter- consent was obtained. The study was approved by the local Human defibrillator (ICD) implantation, which decreases the risk of sudden Research Ethics Committee and complied with the 1975 Declaration cardiac death by treating life-threatening arrhythmias as they arise of Helsinki guidelines. (4). Current guidelines recommend ICD implantation to reduce Patient characteristics including Global Registry of Acute sudden cardiac death in patients with New York Heart Association Coronary Events (GRACE) and TIMI (Thrombolysis in Myocardial (NYHA) class II–III and LVEF ≤ 35% despite optimal medical Infarction) scores, Killip class at admission, peak creatine kinase MB therapy for > 3 months and at least ≥ 6 weeks (≈40 days) after mass and TIMI flow grade in the culprit artery (before and after STEMI (4, 5). reperfusion) were recorded. Nevertheless, most patients treated with an ICD will never require appropriate ICD therapies (ICD-T). In a cohort including five landmark ICD trials, barely one in five patients (18%) were treated with ICD-T during follow-up (6), and the rate was even 2.2. ICD implantation lower (2.6% at 30 months) among primary prevention patients in a contemporary registry (7). Although ICD-T could be considered Selection criteria for this analysis were patients treated with ICD potentially lifesaving in those cases, there is an unmet need for in primary prevention in which pre-implantation Echo and CMR strategies aimed at better selection of patients who really benefit from had been performed. After at least 6 weeks of optimized medical ICD implantation. therapy, symptomatic (NYHA class II–III) patients with LVEF ≤ 35% Quantification of the scar zone after STEMI and more by Echo, CMR or both underwent ICD implantation. Median time precise measurement of LVEF could potentially improve from STEMI to ICD implantation was 40.93 (15.43–188.86) weeks. patient selection. LVEF by echocardiography (Echo) is routinely The flowchart of patients and reasons for exclusion can be consulted performed both at predischarge and during follow-up in most in Supplementary Figure 1. patients. However, cardiac magnetic resonance (CMR) imaging A single lead was positioned in the right ventricular apex using provides the gold standard measurement of LVEF (8, 9) and transvenous access. Five (9.6%) patients with expected need for major crossover between LVEF categories have been shown pacing due to atrioventricular block were fitted with an additional when Echo and CMR values are compared (10). Additionally, auricular lead, and an additional left ventricular lead through the CMR can characterize infarct size (IS), an important predictor coronary sinus was implanted as cardiac resynchronization therapy of arrhythmic risk (11). However, the utility of CMR before in three (5.8%) patients. The ICD generator was placed in a ICD implantation has neither been established in clinical prepectoral pocket in the left upper chest. practice nor studied in specific trials, so its usefulness in this Standard three-zone programming was initially provided, scenario is unknown. with approximate ranges of 175–210 bpm for slow Against this background, we aimed to compare LVEF by ventricular tachycardia, 210–250 bpm for fast ventricular Echo and CMR measured before ICD implantation for ability tachycardia, and > 250 bpm for ventricular fibrillation. to predict ICD-T in a STEMI population treated with ICD in Subsequent changes in programming were not systematically primary prevention. recorded in the registry. Remote patient monitoring was Frontiers in Cardiovascular Medicine 02 frontiersin.org Marcos-Garcés et al. 10.3389/fcvm.2023.991307 provided for most patients (n = 35, 67.3%) and in-person 2.5. LVEF and IS categorization medical visits were scheduled on a yearly basis. If remote monitoring was not available, in-person follow-up visits were Patients were categorized as LVEF ≤ 35% or LVEF > 35% scheduled every 6 months. by Echo and CMR following current recommendations for ICD implantation in primary prevention in ischemic cardiomyopathy (4, 5). Occurrence of the clinical endpoint was analyzed in these LVEF categories. 2.3. Echocardiography All patients underwent echocardiographic examination before ICD implantation. Median time from Echo to ICD implantation 2.6. Endpoint and follow-up was 13 (3.75–24.86) weeks. Local cardiologists carried out studies, The clinical endpoint of this study was occurrence of life- quantified parameters and prospectively included the data in threatening ventricular tachyarrhythmias requiring appropriate the local database. ICD-T [antitachycardia pacing (ATP), cardioversion, or LVEF (%), left ventricular (LV) end-diastolic volume (ml) both]. ICD therapies were considered appropriate when and LV end-systolic volume (ml) were assessed using the biplane ATP, cardioversion or both were used to treat ventricular method of disks (modified Simpson’s rule). Right ventricle function tachycardia or ventricular fibrillation, and inappropriate when was estimated by tricuspid annular plane systolic excursion ATP, cardioversion or both were used for treatment of heart (TAPSE, in mm), which was measured in the apical 4-chamber rhythms other than ventricular tachycardia or ventricular view using the M-mode. Diastolic mitral flow A and E wave fibrillation, such as fast atrial fibrillation. Events were prospectively velocities (m/s) were recorded. Left atrium diameter (mm) was adjudicated by clinical cardiologists via periodic review of also registered. regional electronic health records and remote home-monitoring systems if applicable. 2.4. CMR 2.7. Statistical analysis All patients were examined with a 1.5 T System (Sonata Magnetom, Siemens, Erlangen, Germany) using a standardized The one-sample Kolmogorov–Smirnov Test was used to test protocol (10, 12) before ICD implantation. Median time from CMR normal data distribution. For continuous parametric variables, data to ICD implantation was 10.43 (3.71–27.86) weeks (difference vs. are expressed as mean ± standard deviation and analyzed by Student’s time from Echo to ICD implantation: p = 0.51). t-test. Continuous non-parametric variables are shown as median Images were acquired by a phased-array body surface coil during plus interquartile range and compared with Mann–Whitney U test. breath-holds and were ECG-triggered. Local cardiologists specialized Qualitative variables are presented as percentage and compared by in CMR imaging with > 15 years of experience and accredited Chi-square test or Fisher’s exact test. by the European Society of Cardiology interpreted the studies. The association between variables and time to first ICD-T was Images were examined using customized software (Syngo, Siemens, assessed by multivariable Cox proportional hazard regression models. Erlangen, Germany). Variables with p-value < 0.1 in univariate analysis were included as Cine images were acquired in two-, three-, and four- cofactors in multivariate analysis. Results are presented as hazard chamber views, and in short-axis views using a steady-state free ratio (HR) and 95% confidence interval (CI). precession sequence (repetition time/echo time: 2.8/1.2 ms; flip Receiver operating characteristic curves were computed to angle: 58◦ ; matrix: 256 × 300; field of view: 320 × 270 mm; analyze the sensitivity, specificity, and positive and negative slice thickness: 7 mm). LVEF (%), LV end-diastolic volume predictive value of Echo- and CMR-derived LVEF (≤ 35% or > 35%) index (ml/m2 ), LV end-systolic volume index (ml/m2 ), categories to predict subsequent ICD-T. Areas under the curve and LV mass index (g/m2 ) were calculated by manual for continuous Echo- and CMR-derived LVEF were compared by planimetry of endocardial and epicardial borders in short-axis means of Z test. view cine images. Statistical significance was considered for two-tailed Late gadolinium enhancement (LGE) imaging was performed p-values < 0.05. The SPSS statistical package version 21.0 was used. 10 min after administering gadolinium-based contrast in the same locations as in the cine images, using a segmented inversion recovery steady-state free precession sequence (repetition time/echo time: 3. Results 750/1.26 ms; flip angle: 45◦ ; matrix: 256 × 184; field of view: 340 × 235 mm; slice thickness: 7 mm). Inversion time was adjusted 3.1. Cohort description to nullify normal myocardium. Areas showing LGE were visually quantified by manual In our cohort of 52 STEMI patients treated with ICD in primary planimetry. Infarct size (IS) was assessed as the percentage of LV prevention, mean age was 56.56 ± 11 years, most were male (n = 46, mass showing LGE. Microvascular obstruction (MVO) was defined 88.5%) and smoking was the most prevalent cardiovascular risk factor as the number of segments displaying a lack of contrast uptake in the (n = 36, 69.2%). Most patients in our cohort were Caucasian/White tissue core showing late gadolinium enhancement; the 17-segment (n = 49, 94.3%), while a minority were Asian (n = 1, 1.9%), North model was applied. African Black (n = 1, 1.9%), or Latin American (n = 1, 1.9%). Frontiers in Cardiovascular Medicine 03 frontiersin.org Marcos-Garcés et al. 10.3389/fcvm.2023.991307 TIMI flow grade 3 after pPCI was achieved in 39 (75%) patients. during follow-up. In our population of ICD carriers in primary Mean GRACE risk score was 137.5 ± 36.51 points, largely indicating prevention, therefore, Echo-LVEF ≤ 35% had 40% sensitivity, 52.4% moderate to high risk. Baseline characteristics of the cohort are specificity, 16.7% positive predictive value and 78.6% negative depicted in Table 1. predictive value for appropriate ICD-T, compared to the 100% sensitivity, 23.8% specificity, 23.8% positive predictive value, and 100% negative predictive value of CMR-LVEF ≤ 35% (Table 2). 3.2. Echo and CMR indices MACE per 100 person-years across the Echo and CMR LVEF categories is depicted in Figure 2. CMR-LVEF outperformed Echo- On pre-ICD Echo and CMR, patients displayed extensive LVEF for predicting ICD-T (area under the curve 0.76 vs. 0.48, infarction, LV dysfunction, and dilated LV volumes (Table 1). Mean p = 0.04). Echo-LVEF was 37.42 ± 7.61% compared to mean 30.19 ± 9% CMR- LVEF. Mean IS measured by CMR was 37.61 ± 12.7% of LV mass. The mean absolute difference of LVEF measured by Echo vs. CMR was -7.23 ± 11.51% (p < 0.001). In most patients (n = 41) LVEF was 4. Discussion lower by CMR than Echo; in these cases, the mean absolute difference was -11.81 ± 7.34% (p < 0.001). In a minority of patients (n = 11) In our population of STEMI patients treated with ICD for LVEF was higher by CMR than Echo; mean absolute difference in primary prevention, the main findings of our study are as follows: these cases was 9.8 ± 7.4% (p = 0.001). No interaction was found (1) LVEF measured by CMR outperformed Echo-LVEF and CMR-IS between patients undergoing Echo and CMR ≥ 12 or < 12 weeks to predict ICD-T; (2) CMR-LVEF ≤ 35% accurately identified 100% apart (p = 0.3, Supplementary Figure 2). of patients who would require ICD-T (100% sensitivity and negative predictive value), and (3) no ICD-T were registered during follow-up in patients with CMR-LVEF > 35%. 3.3. Predictors of ICD-T During a mean follow-up of 6.08 ± 4.16 years 4.1. LVEF and ICD in primary prevention (316.04 ± 216.12 weeks), 10 patients underwent appropriate ICD-T: 5 direct shocks to treat very fast ventricular tachycardia In conjunction with other clinical parameters, LVEF is the (n = 3) or ventricular fibrillation (n = 2), 3 effective ATP for cornerstone for early out-of-hospital cardiac arrest prediction in treatment of ventricular tachycardia, and 2 ineffective ATP followed STEMI patients (3). Patients with reduced (< 50%) LVEF have an by shock to treat ventricular tachycardia. A total of 6 patients increased risk of adverse events (10) and sudden cardiac death (3). received 17 additional recurrent ICD-T treatments: 8 direct shocks, However, LVEF can fluctuate widely either upward or downward, 1 effective ATP, and 8 ineffective ATP followed by shock. The rate of and together with the magnitude or absence of recovery, LVEF appropriate ICD-T during follow-up was 3.16 per 100 person-years. level several weeks after myocardial infarction is a relevant marker On univariate analysis, patients with ICD-T during follow-up of adverse events including sudden cardiac arrest (13, 14). Since a presented with higher heart rate on admission (102.8 ± 14.34 bpm trend toward recovery is usually seen in patients with reduced LVEF, vs. 85.83 ± 20.82 bpm, p = 0.019). No significant differences were remeasurement of this parameter is fundamental after the acute noted regarding Echo indices before ICD implantation. However, on phase (12, 15). Nonetheless, in clinical practice reassessment rates are preimplantation CMR, patients with ICD-T during follow-up had relatively low, even in patients with LVEF < 40% in acute phase (16), lower CMR-LVEF (23.7 ± 7.8 vs. 31.74 ± 8.64, p = 0.01) and more and this less closely monitored LVEF results in a lower likelihood of extensive IS (46.88 ± 13.24 vs. 35.41 ± 11.67, p = 0.009) than patients being treated with an ICD (17). without this adverse outcome. On multivariable analysis (Supplementary Table 1), first ICD- T could be predicted by CMR-LVEF [HR 0.9 (0.83–0.99) per %, p = 0.02] and heart rate on admission [HR 1.05 (1–1.1) per beat per 4.2. Echocardiography vs. CMR for risk min, p = 0.03]. The predictive power of IS was marginally significant stratification [HR 1.05 (0.99–1.12) per % of LV mass, p = 0.11]. Pre-ICD Echo- LVEF did not appear to accurately predict use of ICD-T in either Echocardiography has traditionally been the imaging technique univariate or multivariable analyses. of choice for patient follow-up and specifically for LVEF measurement. Indeed, clinical trials for ICD in primary prevention in ischemic heart disease have mostly relied on echocardiography 3.4. ICD-T stratification by LVEF categories for measuring LVEF and selecting the most appropriate cut-off for patient selection (18, 19). Using the recommended cutoff of LVEF ≤ 35% to select patients Nevertheless, CMR imaging is the current gold standard for eligible for ICD implantation in primary prevention, we stratified accurate and reproducible LVEF measurement (8, 9). Unfortunately, our cohort into two groups by both pre-ICD Echo and CMR agreement between LVEF values by echocardiography and CMR is (Figure 1). Using Echo-LVEF, LVEF ≤ 35% identified only 4 out of poor, especially if patients with LVEF of 35% or less are included 10 (40%) patients who received appropriate ICD-T during follow-up. (20). This has also been observed specifically in STEMI patients, an In contrast, CMR-LVEF ≤ 35% before ICD implantation identified effect probably exacerbated by the presence of segmental wall motion 10 out of 10 (100%) patients who underwent appropriate ICD-T abnormalities (10, 21). Frontiers in Cardiovascular Medicine 04 frontiersin.org Marcos-Garcés et al. 10.3389/fcvm.2023.991307 TABLE 1 Baseline, Echo, and CMR characteristics of the entire cohort and of patients with and without ICD-T. All patients ICD-T No ICD-T P-value (n = 52) (n = 10) (n = 42) Clinical variables Age (years) 56.56 ± 11 53.6 ± 12.55 57.38 ± 10.64 0.33 Male sex (%) 46 (88.5) 9 (90) 37 (88.1) 1 Diabetes mellitus (%) 14 (26.9) 3 (30) 11 (26.2) 1 Hypertension (%) 25 (48.1) 3 (30) 22 (52.4) 0.3 Hypercholesterolemia (%) 25 (48.1) 7 (70) 18 (42.9) 0.17 Smoker (%) 36 (69.2) 7 (70) 29 (69) 1 Heart rate on admission (bpm) 89.1 ± 20.74 102.8 ± 14.34 85.83 ± 20.82 0.019 Systolic pressure (mmHg) 127.33 ± 25.19 128.1 ± 20.7 127.14 ± 26.36 0.92 Killip class (%) 1 36 (69.2) 6 (60) 30 (71.4) 2 10 (19.2) 3 (30) 7 (16.7) 0.69 3 2 (3.8) 0 (0) 2 (4.8) 4 4 (7.7) 1 (10) 3 (7.1) Time to reperfusion (hours) 190 (135–432.5) 482 (194–559) 180 (120–420) 0.23 Peak creatine kinase MB mass (ng/ml) 300 (180–489) 482 (194–559) 300 (141.25–427.5) 0.17 Infarct location (%) Anterior 44 (84.6) 9 (90) 35 (83.3) Inferior 7 (13.5) 1 (10) 6 (14.3) 0.82 Lateral 1 (1.9) 0 (0) 1 (2.4) TIMI flow grade before pPCI (%) 0 32 (61.5) 4 (40) 28 (66.7) 1 3 (5.8) 1 (10) 2 (4.8) 0.16 2 9 (17.3) 4 (40) 5 (11.9) 3 8 (15.4) 1 (10) 7 (16.7) TIMI flow grade after pPCI (%) 0 2 (3.8) 0 (0) 2 (4.8) 1 0 (0) 0 (0) 0 (0) 0.61 2 11 (21.2) 3 (30) 8 (19) 3 39 (75) 7 (70) 32 (76.2) GRACE risk score 137.5 ± 36.51 143.5 ± 42.73 136.07 ± 35.3 0.57 TIMI risk score 3.38 ± 2.48 3.7 ± 1.95 3.31 ± 2.6 0.66 Residual ST-segment elevation (n of derivations) 2.76 ± 1.79 2.33 ± 1.41 2.89 ± 1.89 0.42 QRS duration (ms) 97.73 ± 19.07 99.78 ± 12.95 97.28 ± 20.3 0.73 Left bundle branch block (%) 2 (3.8) 1 (10) 1 (2.4) 0.35 Echo indices before ICD implantation Echo-LVEF (%) 37.42 ± 7.61 37.6 ± 5.72 37.38 ± 8.05 0.94 Echo-LV end-diastolic volume (ml) 146.78 ± 43.55 135.4 ± 50.61 148.61 ± 42.97 0.54 Echo-LV end-systolic volume (ml) 90.36 ± 30.48 77.4 ± 29.81 92.45 ± 30.54 0.31 TAPSE (mm) 20.83 ± 4.14 22 ± 3.61 20.67 ± 4.25 0.61 E wave velocity (m/s) 0.78 ± 0.31 1.21 ± 0.52 0.72 ± 0.22 0.15 A wave velocity (m/s) 0.65 ± 0.21 0.57 ± 0.35 0.66 ± 0.2 0.56 Left atrium diameter (mm) 38.31 ± 4.78 37.13 ± 4.16 38.57 ± 4.91 0.45 (Continued) Frontiers in Cardiovascular Medicine 05 frontiersin.org Marcos-Garcés et al. 10.3389/fcvm.2023.991307 TABLE 1 (Continued) All patients ICD-T No ICD-T P-value (n = 52) (n = 10) (n = 42) CMR indices before ICD implantation CMR-LVEF (%) 30.19 ± 9 23.7 ± 7.8 31.74 ± 8.64 0.01 CMR-LV end-diastolic volume index (ml/m2 ) 116.12 ± 32.9 123.6 ± 49.77 114.29 ± 27.9 0.58 CMR-LV end-systolic volume index (ml/m2 ) 82.52 ± 30.83 96.2 ± 46.31 79.1 ± 25.28 0.29 LV mass (g/m2 ) 94.17 ± 19.44 96.29 ± 23.32 93.66 ± 18.82 0.75 Infarct size (% of LV mass) 37.61 ± 12.7 46.88 ± 13.24 35.41 ± 11.67 0.009 bpm, beats per min; CMR, cardiovascular magnetic resonance; Echo, echocardiography; GRACE, Global Registry of Acute Coronary Events; ICD-T, implantable cardioverter-defibrillator therapies; IS, infarct size; LV, left ventricular; LVEF, left ventricular ejection fraction; pPCI, primary percutaneous coronary intervention; TAPSE, tricuspid annular plane systolic excursion; TIMI, Thrombolysis in Myocardial Infarction. In patients with atrial fibrillation at the time of echocardiography, E and A wave velocities were not considered for analyses. FIGURE 1 Stratification of ICD-T according to Echo and CMR LVEF categories before ICD implantation. Occurrence of life-threatening ventricular tachyarrhythmias (manifesting as appropriate ICD therapies) was analyzed in a population of STEMI patients treated with ICD for primary prevention. Patients were categorized by pre-ICD LVEF (≤ 35% vs. >35%) via Echo or CMR. Compared to Echo-LVEF, CMR-LVEF categories allowed for detection of 100% of ICD-T (when CMR-LVEF was ≤ 35%) and exclusion of ICD-T (when CMR-LVEF was > 35%). ATP, antitachycardia pacing; CMR, cardiac magnetic resonance; Echo, echocardiography; ICD, implantable cardioverter-defibrillator; ICD-T, appropriate implantable cardioverter-defibrillator therapies; LVEF, left ventricular ejection fraction; STEMI, ST-segment elevation acute myocardial infarction. 4.3. LVEF to predict ICD therapies predicting major arrhythmic events after myocardial infarction (22). In fact, most patients with LVEF ≤ 35% will never require ICD-T if Basing the decision to implant an ICD in primary prevention implanted with an ICD. Across five landmark ICD trials, only 18% solely on LVEF has several limitations as an approach, as well as a of patients were treated with ICD-T during follow-up (6). The rate is relatively low pooled sensitivity (59.1%) and specificity (77.8%) for even lower (2.6% at 30 months) among primary prevention patients Frontiers in Cardiovascular Medicine 06 frontiersin.org Marcos-Garcés et al. 10.3389/fcvm.2023.991307 TABLE 2 ROC curve characteristics of LVEF categories by Echo and CMR before ICD implantation to predict appropriate ICD-T. Variable Sensitivity Specificity Positive predictive value Negative predictive value Echo-LVEF ≤ 35% 40% 52.4% 16.7% 78.6% CMR-LVEF ≤ 35% 100% 23.8% 23.8% 100% CMR, cardiovascular magnetic resonance; Echo, echocardiography; ICD-T, implantable cardioverter-defibrillator therapies; IS, infarct size; LVEF, left ventricular ejection fraction; ROC, receiver operating characteristic. in a contemporary registry (7), probably due to improved preventive therapies. The rate of appropriate ICD-T during follow-up was 3.16 per 100 person-years in our cohort, similar to previously published cohorts. As an example, in the Danish ICD register patients with ischemic heart disease and ICD implanted for primary prevention showed an appropriate ICD-T rate of 4.12 per 100 person-years (23). Furthermore, as most contemporary STEMI patients maintain a relatively preserved LVEF after the acute event, the majority of sudden cardiac deaths and arrhythmic events occur in this subset regardless of their reduced risk of arrhythmic events (24, 25). There is a considerable risk of sudden cardiac arrest in STEMI patients with LVEF 35–50%, a population in which ICD is generally not indicated in primary prevention (26). Indeed, in our study most (60%) appropriate ICD-T occurred in patients with Echo-LVEF > 35%. As previously noted, therefore, Echo-derived LVEF is not a sensitive predictor of ICD-T. FIGURE 2 ICD-T per 100 person-years across Echo and CMR LVEF categories before ICD implantation. CMR, cardiac magnetic resonance; Echo, echocardiography; ICD, implantable cardioverter-defibrillator; ICD-T, 4.4. CMR to predict ICD therapies appropriate implantable cardioverter-defibrillator therapies; LVEF, left ventricular ejection fraction. The efficacy of CMR imaging to predict the use of appropriate therapies in ICD carriers is an understudied area, despite being a non-invasive and safe technique with proven predictive value in In the selection of STEMI patients who could benefit from an STEMI patients (10, 12). Studies have shown that CMR-derived ICD in primary prevention, clinicians should weigh up the risk of LVEF (27) and IS (27, 28) can predict adverse arrhythmic cardiac undertreatment, which would deprive certain patients of potentially events and ICD-T. In fact, CMR-LVEF outperforms Echo-LVEF life-saving therapy in case of fatal ventricular arrhythmias, against to predict a composite endpoint of all-cause death or ICD-T in the risk of overtreatment, which would increase provider costs and ICD carriers in primary prevention with ischemic and non-ischemic morbidity associated with complications and inappropriate therapies cardiomyopathy (29). Likewise, quantification of myocardial fibrosis in patients fitted with a device. Based on our data, a LVEF ≤ 35% and gray zone fibrosis in CMR can predict sudden cardiac death cut-off by CMR imaging could accurately identify all patients who and an arrhythmic endpoint in a mixed (ischemic and non-ischemic) would require ICD-T during follow-up, and most importantly, population of cardiac implantable electronic device receivers (30). safely exclude ventricular tachyarrhythmias in patients with CMR- Interestingly, in patients with no evidence of fibrosis in CMR, sudden LVEF > 35%, in whom theoretically an ICD would render no cardiac death can be virtually excluded. However, in our STEMI preventive benefit. cohort the presence of at least some degree of myocardial necrosis was universal, and CMR-LVEF appears more useful to predict (or exclude) ICD-T during follow-up. 4.5. Study limitations Occurrence of pre-ICD CMR-LVEF ≤ 35% identified all patients (100%) in this cohort who would undergo ICD-T during follow-up, Our cohort is limited in the number of recruited patients and clearly outperforming pre-ICD Echo-LVEF ≤ 35%, which had only compiled variables. Additionally, patients were included over a long 40% sensitivity to detect ICD-T. The presence of CMR-LVEF > 35% period (between 2004 and 2021) during which time there were thus indicates a low risk of adverse arrhythmic events and can variations in acute and chronic STEMI treatment. Only STEMI pinpoint a population in which ICD implantation should be withheld. patients who underwent both echocardiography and CMR before In contrast, the possibility of adverse arrhythmic events could not be ICD implantation were selected, so the cohort may not be entirely ruled out in the subgroup with Echo-LVEF > 35%. representative of the whole STEMI population. Another drawback is We hypothesize that these findings can be explained by the the time difference between Echo and CMR and ICD implantation increased accuracy of LVEF measurement by CMR (i.e., LVEF ≤ 35% in our cohort; performing these techniques sequentially over a by CMR represents a truly reduced LVEF) and the lower LVEF short period could have allowed better comparison. Patients not obtained by CMR compared to echocardiography (i.e., LVEF > 35% undergoing ICD implantation were not studied, which allowed us by CMR represents a truly not-severely reduced LVEF). to accurately analyze the occurrence of ICD-T but limited our Frontiers in Cardiovascular Medicine 07 frontiersin.org Marcos-Garcés et al. 10.3389/fcvm.2023.991307 results to ICD carriers. Changes in ICD programming were not NP, JG, ML-L, JM, and VB: investigation and methodology. VM-G, systematically recorded during follow-up. Furthermore, although NP, CR-N, and VB: project administration. VM-G, NP, JG, ML-L, most patients showed extensive infarction, underwent CMR study JM, CR-N, ED, HM-G, AG-P, ÁF-D-L-O, ÁM-B, LB, JS-G, CA, JN, and ischemic etiology could be inferred as the most probable cause, AB-G, FC, RR-G, and VB: resources. JN, AB-G, FC, RR-G, and VB: other underlying cardiomyopathies could not be ruled out with supervision and writing—review and editing. VM-G, NP, JG, and VB: absolute certainty. Lastly, the implications of ICD implantation validation and visualization. All authors contributed to the article and based on CMR-LVEF vs. Echo-LVEF should be explored in approved the submitted version. specifically designed, prospective, and randomized studies. Due to the observational nature of our study no formal recommendations regarding ICD indication can be inferred. Funding This work was supported by “Instituto de Salud Carlos III” and 5. Conclusion “Fondo Europeo de Desarrollo Regional (FEDER)” (grant numbers PI20/00637, CIBERCV16/11/00486, a postgraduate contract In STEMI patients treated with ICD in primary prevention, FI18/00320 to CR-N, and CM21/00175 to VM-G), Conselleria assessment of LVEF by CMR outperforms Echo-LVEF to predict de Educación—Generalitat Valenciana (PROMETEO/2021/008), subsequent appropriate ICD therapies. Occurrence of CMR- and Sociedad Española de Cardiología (Grant SEC/FEC-INV- LVEF ≤ 35% identified all patients who would undergo ICD therapies CLI 21/024). JG acknowledges financial support from the (100% sensitivity and negative predictive value). Strategies aimed at “Agencia Estatal de Investigación” (grant FJC2020-043981- selective ICD implantation based on CMR data should be further I/AEI/10.13039/501100011033). explored in properly designed studies. Data availability statement Conflict of interest The authors declare that the research was conducted in the The datasets presented in this article are not readily available absence of any commercial or financial relationships that could be because of privacy and ethical restrictions. Requests to access the construed as a potential conflict of interest. datasets should be directed to VB, vicente.bodi@uv.es. Ethics statement Publisher’s note The study was reviewed and approved by the Comité Ético All claims expressed in this article are solely those of the de Investigación con Medicamentos of the Hospital Clínico authors and do not necessarily represent those of their affiliated Universitario de Valencia. The patients/participants provided their organizations, or those of the publisher, the editors and the reviewers. written informed consent to participate in this study. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Author contributions VM-G, NP, and VB: conceptualization and writing—original Supplementary material draft. VM-G, NP, JG, ML-L, JM, CR-N, ED, HM-G, AG-P, ÁF-D-L-O, ÁM-B, LB, JS-G, and CA: data curation. VM-G, NP, JG, ML-L, JM, The Supplementary Material for this article can be found online CR-N, ED, HM-G, AG-P, ÁF-D-L-O, ÁM-B, LB, JS-G, CA, and VB: at: https://www.frontiersin.org/articles/10.3389/fcvm.2023.991307/ formal analysis. JN, AB-G, FC, and VB: funding acquisition. VM-G, full#supplementary-material References 1. Nabel E, Braunwald E. A tale of coronary artery disease and myocardial 4. Al-Khatib S, Stevenson W, Ackerman M, Bryant W, Callans D, Curtis A, et al. 2017 infarction. N Engl J Med. (2012) 366:54–63. doi: 10.1056/NEJMra11 AHA/ACC/HRS Guideline for Management of Patients With Ventricular Arrhythmias 12570 and the Prevention of Sudden Cardiac Death: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines 2. Fan X, Hua W, Xu Y, Ding L, Niu H, Chen K, et al. Incidence and predictors of sudden and the Heart Rhythm Society. Circulation. (2018) 138:e272–391. doi: 10.1161/CIR. cardiac death in patients with reduced left ventricular ejection fraction after myocardial 0000000000000614 infarction in an era of revascularisation. Heart. (2014) 100:1242–9. doi: 10.1136/heartjnl- 2013-305144 5. Ibanez B, James S, Agewall S, Antunes M, Bucciarelli-Ducci C, Bueno H, et al. 2017 ESC Guidelines for the management of acute myocardial infarction in patients presenting 3. Faxén J, Jernberg T, Hollenberg J, Gadler F, Herlitz J, Szummer K. Incidence and with ST-segment elevationThe Task Force for the management of acute myocardial predictors of out-of-hospital cardiac arrest within 90 days after myocardial infarction. J infarction in patients presenting with ST-segment elevation of the European Society of Am Coll Cardiol. (2020) 76:2926–36. doi: 10.1016/j.jacc.2020.10.033 Cardiology (ESC). Eur Heart J. (2018) 39:119–77. Frontiers in Cardiovascular Medicine 08 frontiersin.org Marcos-Garcés et al. 10.3389/fcvm.2023.991307 6. Aktaş M, Younis A, Zareba W, Kutyifa V, Klein H, Daubert J, et al. Survival after 18. Moss A, Zareba W, Hall W, Klein H, Wilber D, Cannom D, et al. Prophylactic implantable cardioverter-defibrillator shocks. J Am Coll Cardiol. (2021) 77:2453–62. doi: implantation of a defibrillator in patients with myocardial infarction and reduced ejection 10.1016/j.jacc.2021.03.329 fraction. N Engl J Med. (2002) 346:877–83. doi: 10.1056/NEJMoa013474 7. Sabbag A, Suleiman M, Laish-Farkash A, Samania N, Kazatsker M, Goldenberg I, et al. 19. Bardy G, Lee K, Mark D, Poole J, Packer D, Boineau R, et al. Amiodarone or an Contemporary rates of appropriate shock therapy in patients who receive implantable implantable cardioverter–defibrillator for congestive heart failure. N Engl J Med. (2005) device therapy in a real-world setting: From the Israeli ICD Registry. Heart Rhythm. 352:225–37. doi: 10.1056/NEJMoa043399 (2015) 12:2426–33. doi: 10.1016/j.hrthm.2015.08.020 20. Pellikka P, She L, Holly T, Lin G, Varadarajan P, Pai R, et al. Variability in 8. Grothues F, Smith G, Moon J, Bellenger N, Collins P, Klein H, et al. Comparison ejection fraction measured by echocardiography, gated single-photon emission computed of interstudy reproducibility of cardiovascular magnetic resonance with two-dimensional tomography, and cardiac magnetic resonance in patients with coronary artery disease echocardiography in normal subjects and in patients with heart failure or left ventricular and left ventricular dysfunction. JAMA Netw Open. (2018) 1:e181456. doi: 10.1001/ hypertrophy. Am J Cardiol. (2002) 90:29–34. doi: 10.1016/S0002-9149(02)02381-0 jamanetworkopen.2018.1456 9. Bulluck H, Dharmakumar R, Arai A, Berry C, Hausenloy D. Cardiovascular magnetic 21. Hendriks T, Al Ali L, Maagdenberg C, van Melle J, Hummel Y, Oudkerk M, resonance in acute st-segment-elevation myocardial infarction: Recent advances, et al. Agreement of 2D transthoracic echocardiography with cardiovascular magnetic controversies, and future directions. Circulation. (2018) 137:1949–64. doi: 10.1161/ resonance imaging after ST-elevation myocardial infarction. Eur J Radiol. (2019) 114:6–13. CIRCULATIONAHA.117.030693 doi: 10.1016/j.ejrad.2019.02.039 10. Marcos-Garces V, Gavara J, Lopez-Lereu M, Monmeneu JV, Rios-Navarro C, de 22. Bailey J, Berson A, Handelsman H, Hodges M. Utility of current risk stratification Dios E, et al. Ejection fraction by echocardiography for a selective use of magnetic tests for predicting major arrhythmic events after myocardial infarction. J Am Coll Cardiol. resonance after infarction. Circ Cardiovasc Imaging. (2020) 13:e011491. doi: 10.1161/ (2001) 38:1902–11. doi: 10.1016/S0735-1097(01)01667-9 CIRCIMAGING.120.011491 23. Weeke P, Johansen J, Jorgensen O, Nielsen J, Moller M, Videbaek R, et al. Mortality 11. Alzuhairi K, Lønborg J, Ahtarovski K, Nepper-Christensen L, Kyhl K, Lassen J, et al. and appropriate and inappropriate therapy in patients with ischaemic heart disease and Sub-acute cardiac magnetic resonance to predict irreversible reduction in left ventricular implanted cardioverter-defibrillators for primary prevention: data from the Danish ICD ejection fraction after ST-segment elevation myocardial infarction: A DANAMI-3 sub- Register. Europace. (2013) 15:1150–7. doi: 10.1093/europace/eut017 study. Int J Cardiol. (2020) 301:215–9. doi: 10.1016/j.ijcard.2019.10.034 24. Dagres N, Hindricks G. Risk stratification after myocardial infarction: is left 12. Gavara J, Marcos-Garces V, Lopez-Lereu M, Monmeneu J, Rios-Navarro C, de Dios ventricular ejection fraction enough to prevent sudden cardiac death? Eur Heart J. (2013) E, et al. Magnetic resonance assessment of left ventricular ejection fraction at any time 34:1964–71. doi: 10.1093/eurheartj/eht109 post-infarction for prediction of subsequent events in a large multicenter STEMI registry. J Magn Reson Imaging. (2021) 56:476–87. doi: 10.1002/jmri.27789 25. Bui A, Waks J. Risk stratification of sudden cardiac death after acute myocardial infarction. J Innov Cardiac Rhythm Manage. (2018) 9:3035–49. doi: 10.19102/icrm.2018. 13. Exner D, Kavanagh K, Slawnych M, Mitchell L, Ramadan D, Aggarwal S, et al. 090201 Noninvasive risk assessment early after a myocardial infarction. J Am Coll Cardiol. (2007) 50:2275–84. doi: 10.1016/j.jacc.2007.08.042 26. Adabag S, Zimmerman P, Lexcen D, Cheng A. Predictors of Sudden Cardiac Arrest Among Patients With Post-Myocardial Infarction Ejection Fraction Greater Than 35%. J 14. Chew D, Heikki H, Schmidt G, Kavanagh K, Dommasch M, Bloch Thomsen P, Am Heart Assoc (2021) 10:e020993. doi: 10.1161/JAHA.121.020993 et al. Change in left ventricular ejection fraction following first myocardial infarction and outcome. JACC Clin Electrophysiol. (2018) 4:672–82. doi: 10.1016/j.jacep.2017.12.015 27. Izquierdo M, Ruiz-Granell R, Bonanad C, Chaustre F, Gomez C, Ferrero A, et al. Value of early cardiovascular magnetic resonance for the prediction of adverse arrhythmic 15. Søholm H, Lønborg J, Andersen M, Vejlstrup N, Engstrøm T, Møller J, cardiac events after a first noncomplicated ST-segment-elevation myocardial infarction. et al. Repeated echocardiography after first ever ST-segment elevation myocardial Circ Cardiovasc Imaging. (2013) 6:755–61. doi: 10.1161/CIRCIMAGING.113.000702 infarction treated with primary percutaneous coronary intervention – is it necessary? Eur Heart J Acute Cardiovasc Care. (2015) 4:528–36. doi: 10.1177/2048872614 28. Alexandre J, Saloux E, Dugué A, Lebon A, Lemaitre A, Roule V, et al. Scar extent 556000 evaluated by late gadolinium enhancement CMR: a powerful predictor of long term appropriate ICD therapy in patients with coronary artery disease. J Cardiovasc Magn 16. Chew D, Wilton S, Kavanagh K, Southern D, Tan-Mesiatowsky L, Exner DV. Left Reson. (2013) 15:12. doi: 10.1186/1532-429X-15-12 ventricular ejection fraction reassessment post-myocardial infarction: Current clinical practice and determinants of adverse remodeling. Am Heart J. (2018) 198:91–6. doi: 29. Champ-Rigot L, Gay P, Seita F, Benouda L, Morello R, Pellissier A, et al. Clinical 10.1016/j.ahj.2017.11.014 outcomes after primary prevention defibrillator implantation are better predicted when the left ventricular ejection fraction is assessed by cardiovascular magnetic resonance. J 17. Pokorney S, Miller A, Chen A, Thomas L, Fonarow G, de Lemos J, et al. Cardiovasc Magn Reson. (2020) 22:48. doi: 10.1186/s12968-020-00640-0 Reassessment of cardiac function and implantable cardioverter-defibrillator use among medicare patients with low ejection fraction after myocardial 30. Leyva F, Zegard A, Okafor O, Foley P, Umar F, Taylor R, et al. Myocardial infarction. Circulation. (2017) 135:38–47. doi: 10.1161/CIRCULATIONAHA.116.02 fibrosis predicts ventricular arrhythmias and sudden death after cardiac electronic device 2359 implantation. J Am Coll Cardiol. (2022) 79:665–78. doi: 10.1016/j.jacc.2021.11.050 Frontiers in Cardiovascular Medicine 09 frontiersin.org