JACC: CARDIOVASCULAR IMAGING VOL. 12, NO. 12, 2019 ª 2019 THE AUTHORS. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION. THIS IS AN OPEN ACCESS ARTICLE UNDER THE CC BY-NC-ND LICENSE (http://creativecommons.org/licenses/by-nc-nd/4.0/). ORIGINAL RESEARCH Prognostic Value of Initial Left Ventricular Remodeling in Patients With Reperfused STEMI Jose F. Rodriguez-Palomares, MD, PHD,a,b,* Jose Gavara, MSC,c,* Ignacio Ferreira-González, MD, PHD,a,d Filipa Valente, MD,a,b César Rios, MSC,c Julián Rodríguez-García, MD,a,b Clara Bonanad, MD, PHD,c Bruno García del Blanco, MD, PHD,a,b Gema Miñana, MD, PHD,c Maria Mutuberria, MD,a,b Julio Nuñez, MD, PHD,c José Barrabés, MD, PHD,a,b Artur Evangelista, MD, PHD,a,b Vicente Bodí, MD, PHD,c David García-Dorado, MD, PHDa,b ABSTRACT OBJECTIVES This study sought to establish the best definition of left ventricular adverse remodeling (LVAR) to predict outcomes and determine whether its assessment adds prognostic information to that obtained by early cardiac magnetic resonance (CMR). BACKGROUND LVAR, usually defined as an increase in left ventricular end-diastolic volume (LVEDV) is the main cause of heart failure after an ST-segment elevated myocardial infarction; however, the role of assessment of LVAR in pre- dicting cardiovascular events remains controversial. METHODS Patients with ST-segment elevated myocardial infarction who received percutaneous coronary intervention within 6 h of symptom onset were included (n ¼ 498). CMR was performed during hospitalization (6.2  2.6 days) and after 6 months (6.1  1.8 months). The optimal threshold values of the LVEDV increase and the LV ejection fraction decrease associated with the primary endpoint were ascertained. Primary outcome was a composite of cardiovascular mortality, hospitalization for heart failure, or ventricular arrhythmia. RESULTS The study was completed by 374 patients. Forty-nine patients presented the primary endpoint during follow- up (72.9  42.8 months). Values that maximized the ability to identify patients with and without outcomes were a relative rise in LVEDV of 15% (hazard ratio [HR]: 2.1; p ¼ 0.007) and a relative fall in LV ejection fraction of 3% (HR: 2.5; p ¼ 0.001). However, the predictive model (using C-statistic analysis) failed to demonstrate that direct observation of LVAR at 6 months adds information to data from early CMR in predicting outcomes (C-statistic: 0.723 vs. 0.795). CONCLUSIONS The definition of LVAR that best predicts adverse cardiovascular events should consider both the increase in LVEDV and the reduction in LV ejection fraction. However, assessment of LVAR does not improve information provided by the early CMR. (J Am Coll Cardiol Img 2019;12:2445–56) © 2019 The Authors. Published by Elsevier on behalf of the American College of Cardiology Foundation. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). From the aDepartment of Cardiology, Hospital Universitari Vall d’Hebron, Vall d’Hebron Institut de Recerca, Universitat Autònoma de Barcelona, Barcelona, Spain; bCentro de Investigación Biomédica en Red-para enfermedades cardiovasculares, CIBERCV, Madrid, Spain; cDepartment of Cardiology, Hospital Clinico Universitario, CIBERCV, INCLIVA, University of Valencia, Valencia, Spain; and the dCIBER de Epidemiología y Salud Pública, CIBERESP, Madrid, Spain. *Drs. Rodriguez-Palomares and Gavara contributed equally to this study. This work was funded by the Instituto de Salud Carlos III and co-funded by the European Regional Development Fund grants EC07/90511, PI17/01836, PIE15/00013, CIBERCV16/11/00486, and CIBERCV16/11/ 00479. Dr. Ferreira-González served on advisory boards for Bayer, Bristol-Myers Squibb, and Sanofi; and received grants from the Spanish Health Ministry. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Manuscript received October 8, 2018; revised manuscript received January 17, 2019, accepted February 22, 2019. ISSN 1936-878X https://doi.org/10.1016/j.jcmg.2019.02.025 2446 Rodriguez-Palomares et al. JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Prognostic Value of LV Remodeling in Reperfused STEMI Patients DECEMBER 2019:2445–56 M ABBREVIATIONS ortality during the acute phase of are undergoing PCI. These changes were correlated AND ACRONYMS ST-segment elevation myocardial with the primary outcome of ventricular arrhythmia, infarction (STEMI) has steadily hospitalization for heart failure, or cardiovascular AMI = acute myocardial infarction decreased over the past 3 decades and ap- death during follow-up. We also sought to determine CI = confidence interval pears now to have reached a plateau at lower whether baseline CMR-derived parameters or LVAR values than those in the pre-reperfusion era constitutes the best predictor of cardiovascular CMR = cardiac magnetic resonance (1). However, the main impact of STEMI is events during follow-up. HR = hazard ratio shifting from acute mortality to progressive left ventricular (LV) dysfunction and chronic METHODS IS = infarct size heart failure (2). LV = left ventricle STUDY POPULATION. The present study compiled SEE PAGE 2457 LVAR = left ventricular the databases of 2 previous studies (14,15). The first adverse remodeling The main cause of heart failure after an was a large prospective STEMI registry (14) that LVEDV = left ventricular end- acute myocardial infarction (AMI) is left included consecutive patients admitted for a first diastolic volume ventricular adverse remodeling (LVAR) (3) in STEMI as defined by current definitions (16), treated LVEF = left ventricular ejection fraction response to the increase in wall stress caused with PCI, and undergoing CMR pre-discharge. A sub- LVESV = left ventricular end- by cardiomyocyte loss and distension in the sample of 234 patients underwent a second CMR systolic volume infarct area (4). Post-infarct LVAR is gener- 6 months post-discharge, and they constituted the MI = myocardial infarction ally defined as a 20% increase in LV end- target population. MVO = microvascular diastolic volume (LVEDV) (5,6) and has been The second database stemmed from a double-blind obstruction associated with poorer outcomes (5,7). randomized clinical trial in which 201 patients with a NRI = net reclassification The definition of LVAR is controversial. STEMI were randomized to receive 4.5 mg of adeno- improvement First, the increase in LV volume post- sine or saline intracoronary injection immediately PCI = percutaneous coronary before PCI (15). The primary endpoint of this study intervention infarction may be associated with a rise or fall in left ventricular ejection fraction was the infarct size (%LV) by CMR post-reperfusion. STEMI = ST-segment elevation (LVEF), which has also been considered a Per protocol, in both clinical studies, all patients myocardial infarction major predictor of outcomes after AMI (8). who met the inclusion criteria were scheduled for an Changes in LVEDV may reflect changes in LV volume, early CMR during hospitalization and at 6-month EF, or both, and its analysis provides less information follow-up. than the combined analysis of LV volumes and LVEF. Our target population were patients who survived Furthermore, LVAR is determined 3 to 6 months after at least 6 months after a STEMI and in whom LVAR a STEMI (5,9), and, it has been shown that LVEF im- could be assessed with a subsequent CMR after the proves in most patients after 1-month post-STEMI, acute phase and underwent a minimum of 1-year thereby implying that further delay in clinical de- follow-up (Figure 1). cisions may not be warranted (10). Exclusion criteria were death, previously docu- To overcome these limitations, early cardiac mag- mented MI, severe clinical instability during admis- netic resonance (CMR)-derived parameters such as sion, and any contraindications to CMR, including infarct size (IS), microvascular obstruction (MVO), claustrophobia, existing pacemaker, decision of the and LVEF have been considered excellent predictors patient, and a history of adverse reactions to contrast. of adverse cardiovascular events during follow-up Baseline characteristics were collected prospec- (11,12). A recent meta-analysis showed IS measured tively in all cases. The PCI technique was left to the by CMR or technetium-99m sestamibi single-photon discretion of the interventional cardiologists. Patients emission computed tomography within 1 month af- were managed both in-hospital and after discharge ter the primary percutaneous coronary intervention following specific STEMI guidelines (16). (PCI) to be strongly associated with all-cause mortal- Both study protocols were approved by the hospi- ity and hospitalization for heart failure within 1 year tal Ethics Committee on Human Research and com- (13). However, although several studies support the plied with the 1975 Declaration of Helsinki guidelines. prognostic value of baseline CMR data, no studies to CMR. All CMR studies were performed with a 1.5-T date have shown that they are superior to the direct clinical scanner (Sonata or Avanto scanner, Siemens, assessment of early LVAR for predicting outcomes. Erlangen, Germany). Further details on the technical The present study first tested the hypothesis that aspects of CMR acquisition, sequences, and quantifi- changes in both LV volumes and LVEF better define cation can be found in the Supplemental Appendix. the presence of LVAR because they provide additional LVEF (%), LVEDV index (ml/m 2), left ventricular end- prognostic information in patients with STEMI who systolic volume (LVESV) index (ml/m 2), LV mass JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Rodriguez-Palomares et al. 2447 DECEMBER 2019:2445–56 Prognostic Value of LV Remodeling in Reperfused STEMI Patients F I G U R E 1 Study Flow Chart First STEMI revascularized with PCI: N = 498 6-month CMR not performed: N = 94 (11 of whom died before the CMR was scheduled) 6-month CMR: N = 404 Lost to follow-up: N = 25 Patients with an event (heart failure) before the 2nd CMR was performed: N=5 Completed the study: N = 374 Primary endpoint: N = 49 No primary endpoint: N = 325 Mean follow-up: 72.9 ± 42.8 months Flow chart of participants through the study. Graph showing selection and attrition of participants through the study. Primary endpoint is cardiovascular mortality, hospitalization for heart failure, or ventricular arrhythmia. CMR ¼ cardiac magnetic resonance; PCI ¼ percutaneous coronary intervention; STEMI ¼ ST-segment elevation myocardial infarction. index (g/m 2), IS (% of LV mass), microvascular primary endpoint was cardiovascular mortality, hospi- obstruction (% of LV mass), myocardial edema (% of talization for heart failure, or ventricular arrhythmia. LV mass), and myocardial salvage index (% of LV The optimal threshold values of the LVEDV in- mass) were calculated. crease and the LVEF decrease associated with the CMR studies were analyzed offline by an experi- primary endpoint were ascertained using a graphic enced observer blinded to all patient data using method for biomarker cutoff optimization (17). The customized software (QMASS MR 6.15, Medis, Leiden, hazard ratios (HRs) including 95% confidence in- the Netherlands; or Cvi42, Circle Cardiovascular Im- tervals (CIs) were plotted regardless of the cutoffs, aging Inc., Calgary, Alberta, Canada). and vertical lines designated the dichotomization STATISTICAL ANALYSIS. Continuous demographic showing the most significant correlation with sur- variables were expressed as mean  SD. Normality vival. To avoid absurd estimations, 10% of extreme distribution was assessed by the Kolmogorov-Smirnov outliers of LVEDV and LVEF values were excluded test. Differences between groups for continuous pa- as potential candidates to obtain the cutoff points. rameters were assessed by Student’s t-test if they Finally, the values obtained were rounded up to a presented a normal distribution or analysis of variance superior whole for pragmatic purposes. A scale of 1 with Bonferroni correction for multiple comparisons, percentage unit was used for the decrease in LVEF and Mann-Whitney U test if they did not present a and a scale of 5 percentage units for LVEDV. normal distribution. For categorical variables, general Survival free from the primary endpoint was characteristics of the sample were assessed by per- compared using Kaplan-Meier curves, and log-rank centages (chi-square test or Fisher exact test, test across the 4 subgroups of patients considered accordingly). the cutoff points selected: 1) those whose LVEDV and Our main hypothesis was that both LVEDV and LVEF did not change; 2) those whose LVEDV LVEF provide complementary prognostic informa- increased but LVEF did not decrease; 3) those whose tion for STEMI after primary PCI and thus both LVEDV did not increase but the LVEF decreased; should be considered in the definition of LVAR. The and 4) those whose LVEDV increased and LVEF 2448 Rodriguez-Palomares et al. JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Prognostic Value of LV Remodeling in Reperfused STEMI Patients DECEMBER 2019:2445–56 T A B L E 1 Baseline Characteristics of the Entire Study Group and of Patients With or Without Cardiovascular Events All Patients Patients With No Events Patients With Events (N ¼ 374) (n ¼ 325) (n ¼ 49) N Mean  SD or n (%) n Mean  SD or n (%) n Mean  SD or n (%) p Value Demographics Age, yrs 374 59.2  12 325 58.2  11.8 49 65.5  11.2 <0.001 Female 374 62 (16.6) 325 56 (17.2) 49 6 (12.2) 0.536 Hypertension 374 168 (44.9) 325 144 (44.3) 49 24 (49) 0.542 Hypercholesterolemia 374 144 (38.5) 325 128 (39.4) 49 16 (32.7) 0.432 Diabetes 374 63 (16.8) 325 56 (17.2) 49 7 (14.3) 0.687 Smoking 374 235 (62.8) 325 205 (63.1) 49 30 (61.2) 0.874 Prior coronary heart disease 374 34 (9.1) 325 31 (9.5) 49 3 (6.1) 0.597 Index episode AMI location 374 325 49 0.340 Anterior 211 (56.4) 179 (55.1) 32 (65.3) Inferior 146 (39) 130 (40) 16 (32.7) Lateral 17 (4.5) 16 (4.9) 1 (2) Heart rate 374 77.6  18.8 325 76.3  17.8 49 86.3  22.4 0.003 Systolic blood pressure 374 129.1  28.8 325 128.7  28.8 49 131.8  29 0.541 Killip class >I 374 51 (13.6) 325 43 (13.2) 49 8 (16.3) 0.510 PCI <12 h 368 247 (67.1) 319 215 (67.4) 49 32 (65.3) 0.747 Coronary artery 371 323 48 0.149 LAD 210 (56.6) 178 (55.1) 32 (66.7) LCX 130 (35) 115 (35.6) 15 (31.3) RCA 31 (8.4) 30 (9.3) 1 (2.1) Multivessel disease 366 116 (31.7) 317 103 (32.5) 49 13 (26.5) 0.510 Initial TIMI 369 321 48 0.052 Occlusion 236 (64) 210 (65.4) 26 (54.2) Low 24 (6.5) 22 (6.9) 2 (4.2) Intermediate 27 (7.3) 23 (7.2) 4 (8.3) Normal 82 (22.2) 66 (20.6) 16 (33.3) Final TIMI 374 325 49 0.850 Occlusion 6 (1.6) 6 (1.8) 0 (0) Low 2 (0.5) 2 (0.6) 0 (0) Intermediate 27 (7.2) 20 (6.2) 7 (14.3) Normal 339 (90.6) 297 (91.4) 42 (85.7) Drug-eluting stent 241 0.4  0.5 219 0.4  0.5 22 0.1  0.4 0.009 Stents, total number 229 1.2  0.6 208 1.2  0.6 21 1.3  0.9 0.544 CK-MB mass 319 267.4  382.2 270 258.8  404.1 49 314.9  223.6 0.022 Continued on the next page decreased. The association between LVEDV and LVEF included information related to the change between changes were estimated in multivariable crude and the early and 6-month CMR; and 4) models that adjusted Cox models. All baseline variables were included information from the early CMR and the analyzed for their association with the primary change in LVEDV and LVEF. All models were adjusted outcome (Supplemental Table 1). Specifically, the for age and sex. baseline variables of age, sex, heart rate, LVEF, IS, Finally, the discrimination ability of each model to and MVO were selected for statistical adjustment. classify patients at a higher risk for the primary The estimated best model to identify patients with outcome was additionally assessed by the net the primary outcome during follow-up was based on reclassification improvement (NRI). Considering a the C-statistic. The objective of this analysis was to baseline model including only the variables age and compare the predictive ability of 4 different models: sex, the NRI was estimated for each new model. NRI 1) models including information exclusively from the was estimated at the median follow-up (297 weeks) early CMR; 2) models including information exclu- and considered a threshold of 20% to define a high- sively from the 6-month CMR; 3) models that risk patient. JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Rodriguez-Palomares et al. 2449 DECEMBER 2019:2445–56 Prognostic Value of LV Remodeling in Reperfused STEMI Patients T A B L E 1 Continued All Patients Patients With No Events Patients With Events (N ¼ 374) (n ¼ 325) (n ¼ 49) N Mean  SD or n (%) n Mean  SD or n (%) n Mean  SD or n (%) p Value CMR baseline data LVEF, % 374 51.6  12 325 52.5  11.5 49 45.8  13.3 0.001 LVEDV, ml 374 79.7  21.9 325 79.1  20.8 49 83.9  28.2 0.422 LVESV, ml 374 39.6  18.8 325 38.4  17.6 49 47.6  23.9 0.018 Edema, % of LV mass 360 31.3  16 312 30.6  15.8 48 36  16.5 0.038 Myocardial salvage index, % 359 22  22.9 311 23.2  23.8 48 14.5  14.4 0.039 IS, % of LV mass 373 21.9  14.3 324 20.8  13.8 49 28.7  15.6 0.001 MVO, % of LV mass 373 1.8  3.8 324 1.6  3.6 49 2.9  4.3 0.004 Global CS, % 229 16.5  4 198 16.8  3.8 31 14.3  4.2 0.001 Global LS, % 229 10.4  3.1 198 10.7  3.1 31 8.5  2.5 <0.001 Global RS, % 229 38.5  11.7 198 39  11.6 31 34.8  11.9 0.066 Medication at follow-up Antiplatelet agent 374 374 (100) 325 325 (100) 49 49 (100) NA Anticoagulation 374 43 (11.5) 325 41 (12.6) 49 2 (4.1) 0.094 Beta-blockers 373 306 (82) 324 271 (83.6) 49 35 (71.4) 0.046 ACE Inhibitors 374 251 (67.1) 325 220 (67.7) 49 31 (63.3) 0.625 Angiotensin II receptor antagonist 374 42 (11.2) 325 38 (11.7) 49 4 (8.2) 0.629 Statin 374 354 (94.7) 325 307 (94.5) 49 47 (95.9) 1.000 Values are n, mean  SD, or n (%). ACE ¼ angiotensin-converting enzyme inhibitor; AMI ¼ acute myocardial infarction; CK-MB ¼ creatine kinase-myocardial band; CMR ¼ cardiac magnetic resonance; CS ¼ circumferential strain; IS ¼ infarct size; LAD ¼ left anterior descending coronary artery; LCX ¼ left circumflex coronary artery; LS ¼ longitudinal strain; LV ¼ left ventricle; LVEDV ¼ left ventricular end-diastolic volume; LVEF ¼ left ventricular ejection fraction; LVESV¼ left ventricular end-systolic volume; MVO ¼ microvascular obstruction; NA ¼ not applicable; PCI ¼ percutaneous coronary intervention; RCA ¼ right coronary artery; RS ¼ radial strain; TIMI ¼ Thrombolysis In Myocardial Infarction flow grade. T A B L E 2 Differences Between Baseline and 6-Month LVEDV, LVESV, LVEF, and All analyses were made with SPSS statistical pack- IS (%LV Mass) in Patients With and Without the Cardiovascular Event age (SPSS Statistics 23.0, IBM, Armonk, New York) Global Event No Event p and R (version 3.4.3, R Foundation for Statistical (n ¼ 374) (n ¼ 49) (n ¼ 325) Value Computing, Vienna, Austria). LVEDV Baseline 80  22 84  28 79  21 0.422 6-month 82  26 90  37 80  24 0.278 RESULTS Difference 2  17.8 6  23.1 1.4  16.8 0.255 Relative difference, % 4  23.2 8.3  26.6 3.4  22.7 0.224 BASELINE CHARACTERISTICS. Ninety-four of the p Value 0.031 0.076 0.136 498 patients with a first STEMI revascularized with LVESV Baseline 40  19 48  24 38  18 0.018 PCI were excluded because a second CMR was not 6-month 39  23 52  32 37  20 0.005 performed at 6 months. Twenty-five patients (6.2%) Difference 0.3  12.8 4.4  18.4 1  11.6 0.052 were lost to follow-up, and 5 (1.2%) had an event Relative difference, % 1.6  34.5 12.1  41.8 0.1  33 0.032 (heart failure) before the second CMR was scheduled; p Value 0.662 0.099 0.122 thus, 374 patients were finally included in the study. LVEF The mean time between the PCI and early CMR was Baseline 52  12 46  13 53  12 0.001 6.2  2.6 days. Baseline data of the study population 6-month 55  12 46  14 56  11 <0.001 are shown in Table 1. Most patients had an anterior Difference 3  8.4 0.6  10 3.4  8.1 0.018 Relative difference, % 7.6  20 3.4  27.2 8.3  18.6 0.017 STEMI (n ¼ 211; 56.4%). The culprit artery was the left p Value <0.001 0.689 <0.001 anterior descending coronary artery (n ¼ 210; 56.6%), IS (% of LV mass) which in most cases was completely occluded Baseline 25.4  13 31.7  14.4 24.2  12.4 <0.001 (n ¼ 236; 64%). Multivessel coronary artery disease 6-month 22.9  13.7 28.7  15.8 21.7  13.1 0.001 was present in 116 patients (31.7%). PCI was per- Difference 2.5  6 3.1  7.9 2.4  5.6 0.499 formed in the first 12 h after initial chest pain in the Relative difference, % 10.9  19.9 10.9  21 10.9  18.7 0.985 majority of patients (n ¼ 247; 67.1%), in most of whom p Value <0.001 0.009 <0.001 (n ¼ 339; 90.6%) it achieved normal coronary flow. Values are mean  SD unless otherwise indicated. Event is considered a cardiovascular mortality, hospitalization Early CMR showed a baseline EF of 51.6  12% and for heart failure, or ventricular arrhythmia. mean IS (%LV) was 21.9 14.3%. No differences Abbreviations as in Table 1. 2450 Rodriguez-Palomares et al. JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Prognostic Value of LV Remodeling in Reperfused STEMI Patients DECEMBER 2019:2445–56 be greater in patients who had the primary outcome F I G U R E 2 HR Curves during follow-up (6  23.1 ml; 8.3  26.6%; p ¼ 0.076) Change in LVEDV (%) than in those without an event (1.4  16.8 ml; 3.4  5.0 22.7%; p ¼ 0.136), although these differences were not statistically significant (p ¼ 0.224). Also, LVEF pre- sented a relative increase of 7.6  20% (p < 0.001). 2.0 This increase was greater (p ¼ 0.017) in patients HR without the primary outcome during follow-up 1.0 (relative increase 8.3  18.6%; p < 0.001) than in HR Cutoff = 12.5 those with events (relative 3.4  27.2%; p ¼ 0.689). A Significance Cutoff = 12.5 decrease in infarct size of 10.9  19.9% was observed 0.5 in the overall population with no differences between –20 –10 0 10 Cutoff patients with or without major cardiovascular events (p ¼ 0.985). These results show that patients with and without Change in LVEF (%) events during follow-up can present similar changes 5.0 in LVEDV and IS during the first 6 months after STEMI; however, they do present differences in the change in LVEF during this period. 2.0 CLINICAL EVENTS AND LV REMODELING. In the HR univariate analysis, age, heart rate, Killip class >I, 1.0 HR Cutoff = –1.3 infarction in the left anterior descending artery, basal Significance Cutoff = 2.9 LVEF (%), LVESV (ml), myocardial edema (% of LV 0.5 mass), myocardial salvage index (%), MVO, and –40 –30 –20 –10 0 infarct size (% of LV mass) constituted the main pre- Cutoff dictors of cardiovascular events during follow-up (Supplemental Table 1). After analysis of the associ- Hazard ratio (HR) curves to ascertain the best cutoff point in left ventricular end-diastolic ation with the primary outcome, the following base- volume (LVEDV) and left ventricular ejection fraction (LVEF) to predict the event. Values line variables were selected for adjustment: age, sex, that maximized the ability to distinguish between patients with and without a subsequent heart rate, IS, MVO, and LVEF. event were a relative increase of 12.5% in LVEDV and a relative decrease in LVEF of 2.9%. HR for the primary outcome across the whole range of values of relative change in LVEDV and relative change in LVEF at 6 months after the index episode existed in medical treatment at discharge between are shown in Figure 2. Values that maximized the groups. ability to distinguish between patients with and FOLLOW-UP AND CMR CHANGES. After discharge, a without a subsequent event were a relative increase second CMR was performed at a mean time of 6.1  of 12.5% in LVEDV (HR: 1.98; 95% CI: 1.16 to 3.4; 1.8 months, and patients were followed up for a mean p ¼ 0.013) and a relative decrease in LVEF of 2.9% of 72.9  42.8 months. Forty-nine patients (13.1%) (HR: 2.38; 95% CI: 1.39 to 4.06; p ¼ 0.001). Rounding presented with the primary outcome: 3 ventricular up to the superior whole corresponded to a relative arrhythmias (0.8%); 28 hospitalization for heart fail- increase in LVEDV of 15% (HR: 2.10; 95% CI: 1.22 to ure (7.5%); and 26 cardiovascular death (7%). In 325 3.61; p ¼ 0.007) and a relative decrease in LVEF of 3% patients (86.9%), none of the primary outcomes (HR: 2.50; 95% CI: 1.47 to 4.27; p ¼ 0.001). occurred. All-cause mortality also occurred in 26 pa- Classification of the population according to the tients (7%). Differences between patients with and relative changes in LVEDV and LVEF resulted in 4 without the primary outcome are presented in groups with different prognoses (Figure 3). Thus, the Table 1. Patients with cardiovascular events were survival rate of patients with neither an increase in older; had a higher heart rate at admission; lower LVEDV (<15%) nor a decrease in LVEF (<3%) was the LVEF and myocardial salvage index at the baseline highest, whereas the presence of both an increase in CMR; and higher LVESV, area at risk, IS, and MVO LVEDV >15% and a decrease in LVEF >3% identified a (p < 0.005 in all cases). subgroup of patients with the poorest prognosis. After 6 months, LVEDV presented a relative in- The association between the changes in LVEDV crease of 4  23.2% (Table 2). This increase tended to and LVEF at 6 months and the primary outcome JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Rodriguez-Palomares et al. 2451 DECEMBER 2019:2445–56 Prognostic Value of LV Remodeling in Reperfused STEMI Patients was also present on crude Cox regression analyses F I G U R E 3 Kaplan-Meier Cumulative Event Curves (Table 3), which showed HR for the primary outcome of 2.8 (95% CI: 1.2 to 6.4; p ¼ 0.015) in patients with no decrease in LVEF (<3%) but an 0.6 increase in LVEDV >15%, an HR of 3.5 (95% CI: 1.7 to 7.2; p ¼ 0.001) in those with a decrease in LVEF 0.5 >3% without an increase in LVEF (<15%), and an HR of 3.9 (95% CI: 1.7 to 8.8; p ¼ 0.001) in those Cumulative Event Rate 0.4 with both conditions, an increase in LVEDV, and a decrease in LVEF. In the Cox-adjusted regression analysis, the 0.3 poorest prognosis was observed in patients with both conditions (HR: 6.4; 95% CI: 2.6 to 15.4; 0.2 p < 0.001). Patients in this group were predomi- nantly men (p ¼ 0.022), who at admission presented 0.1 lower initial TIMI (Thrombolysis In Myocardial Infarction) flow grade (p ¼ 0.026), a higher area at risk (p ¼ 0.023), larger IS (p ¼ 0.039), and a higher 0.0 MVO (p ¼ 0.044) (Table 4). 0 100 200 300 400 500 CMR PREDICTORS OF CLINICAL EVENTS. The ability Time (Weeks) of different models, assessed by the C-statistic, to Risk Group Neither Decrease EF Nor Increase LVEDV detect patients with a higher risk for the primary Not Decrease EF but Increase of LVEDV outcome is shown in Table 5. We observed that the Decrease of EF and NOT Increase of LVEDV combination of a decrease in LVEF associated with Both Decrease of EF and Increase of LVEDV an increase in LVEDV exceeded the predictive ability of the individual change for these parameters. Kaplan-Meier cumulative event curves for the presence of the primary endpoint (car- However, this predictive ability was lower than that diovascular death, hospitalization for heart failure, and ventricular arrhythmias) in the 4 of the model including exclusively the baseline CMR- different groups based on the variations in LVEF and LVEDV during follow-up (n ¼ 379). derived parameters LVEDV, LVEF, and IS: C-statistic: Abbreviations as in Figure 2. 0.714 (95% CI: 0.661 to 0.766) versus 0.770 (95% CI: 0.707 to 0.823). In addition, complementing the in- formation from the baseline CMR-derived parameters suggest that the information provided by a second with the occurrence of LVAR did not significantly CMR at 6 months does not increase the prognostic increase the prediction capacity for cardiovascular implications of the baseline CMR. The same data were events: C-statistic: 0.775 (95% CI: 0.707 to 0.823) obtained using the NRI that showed a higher versus 0.798 (95% CI: 0.741 to 0.848). These results improvement in the reclassification based on the T A B L E 3 Association Between LV Remodeling and the Primary Outcome Crude Model Adjusted Model HR (95% CI) p Value HR (95% CI) p Value Risk group (reference no decrease LVEF nor increase LVEDV) 1.00 0.002 1.00 0.000 No decrease LVEF and Increase LVEDV >15% 2.80 (1.22–6.4) 0.015 1.89 (0.80–4.50) 0.149 Decrease LVEF >3% and not increase LVEDV 3.48 (1.67–7.24) 0.001 4.02 (1.88–8.62) 0.000 Decrease LVEF >3% and increase of LVEDV >15% 3.86 (1.69–8.8) 0.001 6.38 (2.64–15.42) 0.000 Age 1.07 (1.05–1.10) 0.000 Female 0.65 (0.26–1.60) 0.348 Heart rate, beats/min 1.01 (1.00–1.03) 0.079 Baseline IS, % of LV mass 1.00 (0.99–1.02) 0.605 Baseline LVEF, % 0.95 (0.92–0.97) <0.000 MVO, % of LV mass 0.99 (0.93–1.06) 0.832 Association between LV remodeling defined by LVEDV and LVEF changes and the primary outcome when adjusted for sex (female), heart rate (beats/min), baseline IS (% of LV mass), baseline LVEF (%), and MVO (% of LV mass). CI ¼ confidence interval; HR ¼ hazard ratio; other abbreviations as in Table 1. 2452 Rodriguez-Palomares et al. JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Prognostic Value of LV Remodeling in Reperfused STEMI Patients DECEMBER 2019:2445–56 T A B L E 4 Baseline Characteristics of the 4 Different Groups of Patients No Changes LVEF (n ¼ 265) Decrease LVEF >3% (n ¼ 109) No Changes LVEDV Increase LVEDV >15% No Changes LVEDV Increase LVEDV >15% n Mean  SD or n (%) N Mean  SD or n (%) p Value n Mean  SD or n (%) n Mean  SD or n (%) p Value p Value Age, yrs 200 59.2  11.7 65 59.4  12 0.992 69 60.3 12.7 40 56.5  12 0.121 0.436 Female 200 43 (21.5) 65 6 (9.2) 0.028 69 9 (13) 40 4 (10) 0.765 0.022 Hypertension 200 82 (41) 65 31 (47.7) 0.387 69 34 (49.3) 40 21 (52.5) 0.843 0.098 Hypercholesterolemia 200 82 (41) 65 24 (36.9) 0.662 69 21 (30.4) 40 17 (42.5) 0.218 0.462 Diabetes 200 32 (16) 65 11 (16.9) 0.848 69 12 (17.4) 40 8 (20) 0.800 0.551 Smoking 200 125 (62.5) 65 41 (63.1) 1.000 69 43 (62.3) 40 26 (65) 0.839 0.837 Prior coronary heart disease 200 22 (11) 65 3 (4.6) 0.149 69 6 (8.7) 40 3 (7.5) 1.000 0.364 Index episode AMI location 200 65 0.830 69 40 0.649 0.964 Anterior 112 (56) 35 (53.8) 39 (56.5) 25 (62.5) Inferior 81 (40.5) 28 (43.1) 25 (36.2) 12 (30) Lateral 7 (3.5) 2 (3.1) 5 (7.2) 3 (7.5) Heart rate 200 76.8  17.1 65 78.6  23 0.942 69 79.3  17 40 77  22.3 0.558 0.759 Systolic blood pressure 200 130.5  29.1 65 121.1  19.5 0.019 69 134.1  34.7 40 127  27.2 0.444 0.086 Killip class >I 200 25 (12.5) 65 9 (13.8) 0.831 69 11 (15.9) 40 6 (15) 1.000 0.483 PCI <12 h 197 126 (64) 63 48 (76.2) 0.090 68 42 (61.8) 40 31 (77.5) 0.136 0.258 Coronary artery 198 65 0.930 68 40 0.940 0.947 LAD 111 (56.1) 35 (53.8) 40 (58.8) 24 (60) RCA 72 (36.4) 26 (40) 21 (30.9) 11 (27.5) LCX 15 (7.6) 4 (6.2) 7 (10.3) 5 (12.5) Multivessel disease 194 63 (32.5) 65 15 (23.1) 0.164 68 24 (35.3) 39 14 (35.9) 1.000 0.656 Initial TIMI 196 0.9 (1.3) 65 0.6 (1.1) 0.045 68 1.1 (1.4) 40 0.6 (1.1) 0.026 0.026 Final TIMI 200 2.9 (0.5) 65 2.9 (0.2) 0.443 69 2.9 (0.3) 40 2.8 (0.6) 0.154 0.252 Drug-eluting stent 128 0.4  0.5 47 0.4  0.5 0.519 42 0.3  0.5 24 0.3  0.5 1.000 0.568 Stents, total number 121 1.2  0.6 45 1.2  0.5 0.948 40 1.3  0.7 23 1.2  0.5 0.949 0.995 CMR data LVEF, % 200 51.8  10.9 65 46.1  12.4 0.001 69 55.3  12.1 40 53.7  13.2 0.571 <0.001 LVEDV 200 82  19.8 65 67.7  19.8 <0.001 69 84.5  22.5 40 79.7  27.6 0.162 <0.001 LVESV 200 40.6  17.5 65 37.2  17 0.204 69 39.2  20.1 40 39.7  24.6 0.602 0.302 Edema, % of LV mass 190 29.8  15.6 63 36.6  16 0.003 68 29.9  15.8 39 33  16.6 0.485 0.023 Myocardial salvage index, % 190 23.4  22.9 63 19.3  17.4 0.521 68 19.2  24.1 38 24.9  28.4 0.323 0.339 IS, % of LV mass 200 19.9  12.3 65 25.7  13.3 0.003 69 22.7  17.3 39 23.7  17.8 0.778 0.039 MVO, % of LV mass 200 1.1  2.1 65 2.4  3.5 0.005 69 1.7  3 39 4.1  8.5 0.335 0.044 Values are n, mean  SD, or n (%). Baseline characteristics in the 4 different groups based on the variations in LVEF and LVEDV during follow-up. Abbreviations as in Table 1. baseline CMR variables (12.9%) compared with the of “antiremodeling” medications (such as angiotensin- presence of LVAR (4.7%) (Table 5). converting enzyme inhibitors and beta-blockers). However, these improvements in AMI management DISCUSSION have not abolished LVAR, which remains a relatively frequent event after an anterior MI (21,22). During the weeks following a STEMI, the LV may un- The present study shows that the incidence of dergo changes in volume, geometry, and function major cardiovascular events during a mean follow-up associated with the development of heart failure and of 72.9 months in patients treated with PCI after a worse prognosis, a process known as adverse post- STEMI is not trivial (13.1%). Our data also show that infarction remodeling (5,18,19). Although there is no the definition of LVAR should necessarily consider universally accepted definition of adverse post- not only changes in LVEDV but also changes in LVEF infarction remodeling, a 15% to 20% increase in to increase its prognostic significance. Moreover, af- LVEDV is the most widely used criterion (5,20). The ter adjustment for baseline LVEF and other variables, incidence and extent of LVAR after an AMI have the isolated increase in LVEDV was not statistically declined in the era of PCI and the almost systematic use associated with an adverse outcome. However, the JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Rodriguez-Palomares et al. 2453 DECEMBER 2019:2445–56 Prognostic Value of LV Remodeling in Reperfused STEMI Patients increase in LVEDV in the presence of a decrease in T A B L E 5 C-Statistic and NRI for Evaluation of Added Benefit of Different CMR Variables LVEF implied worse prognosis. The cutoff analysis Alone and in Combined Models in Predicting the Primary Endpoint identified an increase in LVEDV of 15% or more and an C-Statistic* (95% CI) NRI† (95% CI) (%) absolute reduction of 3% or more in LVEF during the Decrease LVEF >3% and increase 0.714 (0.661 to 0.766) 4.7 (19.8 to 14.5) 6 months following a STEMI as the criteria that best of LVEDV >15% at 6 months identify patients according to their risk of cardiovas- Change in LVEF at 6 months, % 0.694 (0.625 to 0.757) 3.3 (12.5 to 23.4) cular death, readmission for heart failure, or ven- Change in LVEDV at 6 months, ml 0.708 (0.658 to 0.762) 9.6 (22.4 to 12.0) tricular arrhythmias in subsequent years. Patients not Baseline IS, % of LV mass 0.701 (0.637 to 0.759) 2.0 (13.6 to 13.9) fulfilling either of the 2 criteria had the best prog- Baseline LVEF, % 0.768 (0.707 to 0.820) 10.2 (10.9 to 33.1) Baseline LVEDV, ml 0.705 (0.645 to 0.762) 2.1 (11.7 to 22.3) nosis, and those fulfilling both criteria had the worst. 6-month IS, % of LV mass 0.700 (0.636 to 0.762) 0.3 (14.7 to 19.1) However, analysis of changes in LVEDV and LVEF 6-month LVEF, % 0.792 (0.732 to 0.846) 10.8 (9.9 to 32.9) over the first 6 months after STEMI did not increase 6-month LVEDV, ml 0.746 (0.694 to 0.796) 1.3 (21.2 to 28.7) the prognostic value of the principal CMR-derived Model 1: Baseline LVEF (%) þ IS (%) 0.770 (0.707 to 0.823) 12.9 (8.4 to 33.7) variables obtained during hospitalization. These re- þ LVEDV (ml) sults underline the importance of considering both Model 2: Model 1 þ decrease LVEF >3% and 0.798 (0.741 to 0.848) 8.0 (4.1 to 12.1) increase of LVEDV >15% at 6 months LVEDV and LVEF to define LVAR but do not favor the Model 3: 6-month LVEF (%) þ IS (%) 0.795 (0.737 to 0.848) 8.7 (7.7 to 39.9) routine analysis of early LVAR with CMR to identify þ LVEDV (ml) patients at high risk of cardiovascular events, because The primary endpoint is cardiovascular death, heart failure, or ventricular arrhythmias. *Adjusted for sex and age. this strategy provides delayed prognostic information †NRI assessed at median follow-up. Every value represents the NRI of patients with a higher risk (>20%) for the and does not improve that provided by CMR-derived primary outcome, of each model over a baseline model including the variables sex and age. NRI ¼ net reclassification improvement; other abbreviations as in Tables 1 and 3. variables determined in the early setting. LVEF AND LVEDV IN POST-INFARCTION REMODELING. LVEF is a key prognostic factor in coronary heart fibroblasts into myofibroblasts that play an essential disease and should be assessed in all patients after a role in the healing process (28). The reparative phase STEMI (23). Patients with reduced LVEF have a is associated with reductions in proinflammatory cells greater likelihood of developing progressive LVEDV and an increase in anti-inflammatory Ly6C low, and LVESV dilation during follow-up. However, a mononuclear cell, and M2 macrophages. Increased subgroup of patients with normal baseline LVEF can wall stress in the healthy distant myocardium may also increase the ventricular volumes over time (24). lead to progressive eccentric hypertrophy, LV dila- Wu et al. (24) showed that 15% of patients with a tion, heart failure, and LVAR (29). It is important to smaller infarct size (<18.5% of LV mass) developed note that salvaged and distant myocardia are also LVAR (defined as an increase in LVEDV >10 ml/m 2) infiltrated by inflammatory cells that modulate hy- and 60% of patients with larger IS ($18.5% of LV pertrophy and fibrosis (30,31). mass) did not present LVAR. Thus, IS and LV function Our results showed that patients with major car- in the acute phase do not permit accurate prediction diovascular events presented a trend toward a higher of LVAR occurrence. This could be explained in part increase in LVEDV at 6 months when compared with by subtle abnormalities of segmental LV function that those without events; however, these differences did are not amenable to quantification with a global not reach statistical significance (p ¼ 0.224). endpoint such as LVEF. Thus, LVAR has been pro- Furthermore, the predictive model (using C-statistic posed as a surrogate marker in clinical trials (21,25) analysis and the NRI) of the combination of increase and as a parameter to predict outcomes (26,27). Var- in LVEDV and reduction in LVEF was superior iables measuring LVAR are expected to be more (C-statistic: 0.714; 95% CI: 0.661 to 0.766) to that closely correlated with clinical outcomes because estimated by the LVEDV individually. Although a 3% they integrate different aspects of post-infarction reduction in LVEF may be subtle, these variations can pathophysiology. Some are global (e.g., LV volumes) be identified (32) because of the accuracy and repro- and some regional (e.g., LVEF or strain). ducibility of CMR. A recent study showed that treat- After a STEMI, the loss of contractile activity in the ment with intravenous beta-blockers in STEMI infarct segments and its expansion may increase wall patients can induce changes in LVEF of the same tension in distant LV wall segments. In the infarcted magnitude (an improvement in LVEF of 3.49%) (33). segments, early neutrophil infiltration and proin- Our results, however, failed to demonstrate that flammatory cytokine liberation recruit other inflam- direct observation of LVAR at 6 months is superior to matory cells that are involved in removing necrotic data from CMR in the acute phase for predicting cardiomyocytes and in the differentiation of outcomes. 2454 Rodriguez-Palomares et al. JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Prognostic Value of LV Remodeling in Reperfused STEMI Patients DECEMBER 2019:2445–56 C E NT R AL IL L U STR AT IO N Best Definition of LV Adverse Remodeling and its Prognostic Implications 379 patients with STEMI receiving PCI CMR at baseline and at 6 months follow-up Mean follow-up: 72.9 ± 42.8 months Primary endpoint: • Cardiovascular death • Hospitalization for heart failure • Ventricular arrhythmias 0.6 Risk Group Neither Decrease EF Nor Increase EDLW C-Statistic Predictive Model Not Decrease EF but Increase of EDLW 0.5 Decrease of EF and NOT Increase of EDLW Both Decrease of EF and Increase of EDLW 6 Months LVEF (%) + infarct size (%) + Cumulative Event Rate 0.4 LVEDV (mL) 6 Months ↓ LVEF >3% and ↑ LVEDV 0.3 >15% Baseline LVEF (%) + infarct size (%) + 0.2 LVEDV (mL) 0.1 Baseline infarct size (% of LV mass) 0.0 64 66 68 70 72 74 76 78 80 0. 0. 0. 0. 0. 0. 0. 0. 0. 0 100 200 300 400 500 Time (Weeks) LV remodeling definition with best prognostic LV adverse remodeling does not add prognostic implications : ↓ LVEF >3% and ↑ LVEDV >15% at information over basal CMR 6 months Rodriguez-Palomares, J.F. et al. J Am Coll Cardiol Img. 2019;12(12):2445–56. LV adverse remodeling definition with best prognostic implications includes an increase in LVEDV >15% and a concomitant reduction of LVEF >3%. However, the occurrence of LV adverse remodeling during the first 6 months of evolution post-STEMI does not add prognostic information over that provided by basal CMR. CMR ¼ cardiac magnetic resonance; LVEDV ¼ left ventricular end-diastolic volume; LVEF ¼ left ventricular ejection fraction; PCI ¼ percutaneous coronary intervention; STEMI ¼ ST-segment elevation myocardial infarction. IS AND LV CHANGES IN THE ACUTE PHASE OF onset improves the prognosis of patients with STEMI, STEMI. CMR has been considered the gold standard and that this effect is mediated somehow in part technique in the assessment of both acute and through the beneficial effect on infarct healing and chronic MI (34). Several studies in STEMI patients scar formation (35,37). Our results show that a model have shown IS and LV dysfunction severity to be based on the information provided by early CMR closely related (5,24). Hence, the main objective of (LVEDV, IS, and LVEF) is superior to a model based on myocardial reperfusion is to reduce of IS (35), and PCI, the presence of LVAR (C-statistic: 0.770 [95% CI: when performed early, may limit adverse post- 0.707 to 0.823] vs. 0.714 [95% CI: 0.661 to 0.766]) and infarction remodeling (9). However, different as good as a model based on the information provided studies have shown that PCI results in little myocar- by CMR at 6 months (Central Illustration). These re- dial salvage in most patients when performed beyond sults are consistent with previous studies showing the 4 h after symptom onset and, in a substantial number prognostic significance of acute IS to predict out- of patients, myocardial salvage is slight after shorter comes after a STEMI (11,13). In a recent multicenter ischemic times (13,23,35,36). However, there is solid study, Eitel et al. (11) demonstrated that after an AMI, evidence that reperfusion within 12 h of symptom an LVEF #47%, IS $19% LV, and the presence of MVO JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Rodriguez-Palomares et al. 2455 DECEMBER 2019:2445–56 Prognostic Value of LV Remodeling in Reperfused STEMI Patients predicted the occurrence of major adverse cardiac this regard, an increase in LVEDV (>15%) in the events. presence of a decrease in LVEF (>3%) implies A recent meta-analysis showed IS measured by CMR worse prognosis. However, assessment of LVAR does or single-photon emission computed tomography not increase the prognostic value of the principal 1 month after PCI to be strongly associated with 1-year CMR-derived variables provided by the early CMR. hospitalization for heart failure and all-cause mortality ACKNOWLEDGMENT The authors would like to (13). Furthermore, both microvascular obstruction and thank Christine O’Hara for English revisions. intramyocardial hemorrhage are associated with larger IS, adverse LV remodeling, and worse clinical out- comes. Given the prognostic significance of the infor- ADDRESS FOR CORRESPONDENCE: Dr. David García- mation obtained from the early CMR, it is not Dorado, Department of Cardiology, Hospital Uni- surprising that the information obtained from the versitari Vall d’Hebron, Vall d’Hebron Institut de follow-up CMR does not increase its prognostic value. Recerca (VHIR) Paseo Vall d’Hebron 119-129, 08035 Also, the reduction in the IS at follow-up usually occurs Barcelona, Spain. E-mail: dgdorado@vhebron.net. OR in the first 4 months after a STEMI with few subsequent Dr. Vicente Bodi, Department of Cardiology, Hospital changes, and thus it is not justified to delay the eval- Clinico Universitario, INCLIVA, University of Valen- uation of the patient 6 or more months (38). Finally, in cia, Blasco Ibañez 17, 46010 Valencia, Spain. E-mail: our study, the early CMR was performed at a mean of vicente.bodi@uv.es. 6.2 days after the STEMI when most of the dynamic changes in the infarcted area (e.g., edema, microvas- PERSPECTIVES cular obstruction) had occurred and the necrosis was more stable (39). This could also potentially influence COMPETENCY IN MEDICAL KNOWLEDGE: Our results show the absence of an increase in the prognostic value of that the criteria for defining LVAR should include not only changes the follow-up CMR. in LVEDV but also in LVEF to provide better prognostic information Overall, these results are consistent with the and yield optimal cutoff values for LVEDV and LVEF in the case of notion that early changes in the LV in the acute phase CMR. The present study also helps to establish which patients may of STEMI, as evaluated by CMR, summarize both ef- benefit most from this technique and at what time after PCI it fects of reperfusion and are important in determining should be performed to give the most useful clinical information. LVAR and outcomes. Our data show that CMR during the acute phase of STEMI may STUDY LIMITATIONS. CMR at 6 months could not be identify patients at higher risk of developing adverse outcomes performed in 94 of 498 patients, which could imply a during the following months and years, whereas a routine second risk of selection bias. However, baseline data of these examination at 6 months does not add significant prognostic in- patients were like those who had the second CMR formation and has the major limitation of not permitting prediction performed (Supplemental Table 2). Additionally, 25 of events during the first 6 months of follow-up. patients were lost to follow-up; however, their base- line characteristics did not differ from those of pa- TRANSLATIONAL OUTLOOK: LV remodeling after STEMI is a tients followed up and with 2 CMR studies long-term process; however, most studies evaluate LVAR be- (Supplemental Table 3). tween the third and sixth months post-AMI. Further studies should determine the best time to evaluate its presence for prognostic purposes. Furthermore, LVAR defined considering CONCLUSIONS changes in LVEF in addition to LVEDV better predicts adverse events and has more potential value as a surrogate endpoint in The definition of LV remodeling that best predicts clinical studies in patients with STEMI receiving PCI. adverse cardiovascular events should consider both the increase in LVEDV and the reduction in LVEF. In REFERENCES 1. Mehta RH, Harjai KJ, Cox D, et al., for the PAMI 2. Roger VL, Go AS, Lloyd-Jones DM, et al. Heart inflammation, remote myocardium, and prognosis. Investigators. Clinical and angiographic correlates disease and stroke statistics—2011 update: a report J Am Coll Cardiol Img 2015;8:779–89. and outcomes of suboptimal coronary flow in pa- from the American Heart Association. Circulation 4. Pfeffer MA, Braunwald E. Ventricular remodel- tients with acute myocardial infarction undergoing 2011;123:e18–209. ing after myocardial infarction: experimental ob- primary percutaneous coronary intervention. J Am 3. Carrick D, Haig C, Rauhalammi S, et al. Patho- servations and clinical implications. Circulation Coll Cardiol 2003;42:1739–46. physiology of LV remodeling in survivors of STEMI 1990;81:1161–72. 2456 Rodriguez-Palomares et al. JACC: CARDIOVASCULAR IMAGING, VOL. 12, NO. 12, 2019 Prognostic Value of LV Remodeling in Reperfused STEMI Patients DECEMBER 2019:2445–56 5. Bolognese L, Neskovic AN, Parodi G, et al. Left Society of Cardiology (ESC). Eur Heart J 2018;39: 30. Inserte J, Hernando V, Garcia-Dorado D. ventricular remodeling after primary coronary 119–77. Contribution of calpains to myocardial angioplasty: patterns of left ventricular dilation ischaemia/reperfusion injury. Cardiovasc Res 17. Budczies J, Klauschen F, Sinn BV, et al. Cutoff and long-term prognostic implications. Circulation 2012;96:23–31. finder: a comprehensive and straightforward Web 2002;106:2351–7. application enabling rapid biomarker cutoff opti- 31. Frangogiannis NG. Regulation of the inflam- 6. Bulluck H, Go YY, Crimi G, et al. Defining left mization. PLoS One 2012;7:e51862. matory response in cardiac repair. Circ Res 2012; ventricular remodeling following acute ST- 110:159–73. 18. Gheorghiade M, Bonow RO. Chronic heart segment elevation myocardial infarction using failure in the United States: a manifestation of 32. Grothues F, Smith GC, Moon JC, et al. Com- cardiovascular magnetic resonance. J Cardiovasc coronary artery disease. Circulation 1998;97: parison of interstudy reproducibility of cardiovas- Magn Reson 2017;19:26. 282–9. cular magnetic resonance with two-dimensional 7. St. John Sutton M, Pfeffer MA, Moye L, et al. Car- echocardiography in normal subjects and in pa- 19. Goldberg RJ, Konstam MA. Assessing the diovascular death and left ventricular remodeling two tients with heart failure or left ventricular hyper- population burden from heart failure: need for years after myocardial infarction: baseline predictors trophy. Am J Cardiol 2002;90:29–34. sentinel population-based surveillance systems. and impact of long-term use of captopril: information Arch Intern Med 1999;159:15–7. 33. Pizarro G, Fernández-Friera L, Fuster V, from the Survival and Ventricular Enlargement et al. Long-term benefit of early pre- (SAVE) trial. Circulation 1997;96:3294–9. 20. Gaudron P, Eilles C, Kugler I, Ertl G. Pro- reperfusion metoprolol administration in pa- gressive left ventricular dysfunction and remod- 8. Brooks GC, Lee BK, Rao R, et al. Predicting tients with acute myocardial infarction: results eling after myocardial infarction: potential persistent left ventricular dysfunction following from the METOCARD-CNIC trial (Effect of mechanisms and early predictors. Circulation myocardial infarction. J Am Coll Cardiol 2016;67: Metoprolol in Cardioprotection During an Acute 1993;87:755–63. 1186–96. Myocardial Infarction). J Am Coll Cardiol 2014; 21. Cung TT, Morel O, Cayla G, et al. Cyclosporine 63:2356–62. 9. Sheiban I, Fragasso G, Rosano GMC, et al. Time before PCI in patients with acute myocardial course and determinants of left ventricular func- 34. Kim RJ, Fieno DS, Parrish TB, et al. Rela- infarction. N Engl J Med 2015;373:1021–31. tion recovery after primary angioplasty in patients tionship of MRI delayed contrast enhancement with acute myocardial infarction. J Am Coll Cardiol 22. Savoye C, Equine O, Tricot O, et al. Left to irreversible injury, infarct age, and con- 2001;38:464–71. ventricular remodeling after anterior wall acute tractile function. Circulation 1999;100: myocardial infarction in modern clinical prac- 10. Sjöblom J, Muhrbeck J, Witt N, Alam M, Fryk- 1992–2002. tice (from the REmodelage VEntriculaire man-Kull V. Evolution of left ventricular ejection 35. Keeley EC, Boura JA, Grines CL. Primary an- [REVE] study group). Am J Cardiol 2006;98: fraction after acute myocardial infarction implica- gioplasty versus intravenous thrombolytic therapy 1144–9. tions for implantable cardioverter-defibrillator for acute myocardial infarction: a quantitative re- eligibility. Circulation 2014;130:743–8. 23. Ndrepepa G, Mehilli J, Martinoff S, view of 23 randomised trials. Lancet 2003;361: Schwaiger M, Schömig A, Kastrati A. Evolution of 13–20. 11. Eitel I, de Waha S, Wöhrle J, et al. Compre- left ventricular ejection fraction and its relation- 36. Fieno DS, Hillenbrand HB, Rehwald WG, et al. hensive prognosis assessment by CMR imaging ship to infarct size after acute myocardial infarc- Infarct resorption, compensatory hypertrophy, after ST-segment elevation myocardial infarction. tion. J Am Coll Cardiol 2007;50:149–56. and differing patterns of ventricular remodeling J Am Coll Cardiol 2014;64:1217–26. 24. Wu E, Ortiz JT, Tejedor P, et al. Infarct size by following myocardial infarctions of varying size. 12. Izquierdo M, Ruiz-Granell R, Bonanad C, et al. J Am Coll Cardiol 2004;43:2124–31. contrast enhanced cardiac magnetic resonance is Value of early cardiovascular magnetic resonance a stronger predictor of outcomes than left ven- 37. Chan W, Duffy SJ, White DA, et al. Acute left for the prediction of adverse arrhythmic cardiac tricular ejection fraction or end-systolic volume ventricular remodeling following myocardial events after a first noncomplicated ST-segment- index: prospective cohort study. Heart 2008;94: infarction: coupling of regional healing with elevation myocardial infarction. Circ Cardiovasc 730–6. remote extracellular matrix expansion. J Am Coll Imaging 2013;6:755–61. 25. Perin EC, Willerson JT, Pepine CJ, et al., for Cardiol Img 2012;5:884–93. 13. Stone GW, Selker HP, Thiele H, et al. Rela- the CCTRN Investigators. Effect of trans- 38. Pokorney SD, Rodriguez JF, Ortiz JT, Lee DC, tionship between infarct size and outcomes endocardial delivery of autologous bone marrow Bonow RO, Wu E. Infarct healing is a dynamic following primary PCI: patient-level analysis from mononuclear cells on functional capacity, left process following acute myocardial infarction. 10 randomized trials. J Am Coll Cardiol 2016;67: ventricular function, and perfusion in chronic J Cardiovasc Magn Reson 2012;14:62. 1674–83. heart failure: the FOCUS-CCTRN trial. JAMA 39. Bulluck H, Hammond-Haley M, 14. Bodí V, Sanchis J, López-Lereu MP, et al. 2012;307:1717–26. Weinmann S, Martinez-Macias R, Hausenloy DJ. Usefulness of a comprehensive cardiovascular 26. Torabi A, Cleland JGF, Khan NK, et al. The Myocardial infarct size by CMR in clinical car- magnetic resonance imaging assessment for pre- timing of development and subsequent clinical dioprotection studies: insights from randomized dicting recovery of left ventricular wall motion in course of heart failure after a myocardial infarc- controlled trials. J Am Coll Cardiol Img 2017; the setting of myocardial stunning. J Am Coll tion. Eur Heart J 2008;29:859–70. 10:230–40. Cardiol 2005;46:1747–52. 27. Kramer DG, Trikalinos TA, Kent DM, 15. Garcia-Dorado D, García-del-Blanco B, Antonopoulos GV, Konstam MA, Udelson JE. Otaegui I, et al. Intracoronary injection of adeno- Quantitative evaluation of drug or device effects KEY WORDS cardiac magnetic resonance, sine before reperfusion in patients with on ventricular remodeling as predictors of thera- infarct size, left ventricular ejection fraction, ST-segment elevation myocardial infarction: a peutic effects on mortality in patients with heart left ventricular end-diastolic volume, left randomized controlled clinical trial. Int J Cardiol failure and reduced ejection fraction. J Am Coll ventricular end-systolic volume, left 2014;177:935–41. Cardiol 2010;56:392–406. ventricular remodeling, microvascular 16. Ibanez B, James S, Agewall S, et al., for the obstruction, prognosis, ST-segment 28. Frangogiannis NG. The inflammatory response elevation myocardial infarction ESC Scientific Document Group. 2017 ESC guide- in myocardial injury, repair, and remodelling. Nat lines for the management of acute myocardial Rev Cardiol 2014;11:255–65. infarction in patients presenting with ST-segment elevation: the Task Force for the management of 29. Shah AM, Solomon SD. A unified view of A PPE NDI X For a supplemental appendix acute myocardial infarction in patients presenting ventricular remodelling. Eur J Heart Fail 2010;12: including tables, please see the online version with ST-segment elevation of the European 779–81. of this paper.