Left Ventricular Ejection Fraction Trajectory and Long-Term Outcomes Following Percutaneous Coronary Intervention for Myocardial Infarction Authors: Revathy Sampath-Kumar, Sacharias von Koch, Moman A. Mohammad, Anna Fenerty, Ryan R. Reeves, Belal Al Khiami, Lawrence Ang, Anna Melendez, Ehtisham Mahmud, David Erlinge, Ori Ben-Yehuda Published: 2026 ABSTRACT BackgroundData are lacking on left ventricular ejection fraction (LVEF) trajectory following myocardial infarction (MI).ObjectivesThis study aimed to evaluate the impact of LVEF trajectory after percutaneous coronary intervention (PCI) for MI on long-term outcomes.MethodsPatients undergoing PCI for MI (2007-2022) with serial echocardiograms were identified in the Swedish Coronary Angiography and Angioplasty Registry and the UC San Diego internal National Cardiovascular Data Registry CathPCI Registry. All-cause mortality and heart failure hospitalization within 6 years were assessed with time-to-event analysis adjusted for baseline characteristics.ResultsOf 13,439 patients (mean age 64 ± 10 years, 24.5% female, 46.8% ST-segment elevation MI), 28.2% had consistently reduced LVEF (redEF), 14.8% recovered their LVEF (recEF), 53.0% had consistently preserved LVEF (pEF), and 3.9% decreased from preserved to reduced LVEF (decEF). Heart failure hospitalization rates were 40.0% for redEF, 30.4% for decEF, 16.1% for recEF, and 6.4% for pEF (P < 0.001). Compared to pEF, the heart failure hospitalization risk was significantly higher for redEF (adjusted HR [aHR]: 8.49; 95% CI: 7.49-9.64; P < 0.001), decEF (aHR: 5.74; 95% CI: 4.69-7.02; P < 0.001), and recEF (aHR: 3.18; 95% CI: 2.70-3.76; P < 0.001). Mortality rates were 14.0% for redEF, 14.9% for decEF, 7.8% for recEF, and 6.7% for pEF (P < 0.001). Compared to pEF, the mortality risk was significantly higher for redEF (aHR: 2.20; 95% CI: 1.86-2.59; P < 0.001), decEF (aHR: 2.10; 95% CI: 1.56-2.82; P < 0.001), and recEF (aHR: 1.30; 95% CI: 1.02-1.65; P = 0.032).ConclusionsLVEF trajectory after MI significantly impacts long-term outcomes. Even patients with LVEF recovery remain at increased risk for heart failure hospitalization and mortality compared to those with pEF. Left ventricular dysfunction, with or without clinical heart failure, following myocardial infarction (MI) is associated with an increased risk of mortality and major adverse cardiovascular events.1, 2, 3 Myocardial ischemia and injury can lead to either transient stunning or permanent myocardial dysfunction, with long-term reduction in left ventricular ejection fraction (LVEF) resulting from infarction, adverse remodeling, and cardiomyocyte immune responses.4 Conversely, reverse remodeling and myocardial recovery can occur, leading to an improvement in LVEF. LVEF assessed beyond 6 weeks post-MI has demonstrated greater predictive power for subsequent cardiovascular outcomes compared to LVEF at the time of MI.5,6 Changes in LVEF post-MI, hereafter referred to as LVEF trajectory, have not been fully characterized. Studies have identified a subset of patients who recover their LVEF after MI and others who have persistently reduced LVEF despite percutaneous coronary intervention (PCI). Lack of LVEF recovery has been associated with larger biochemical infarct size, anterior wall MI, severe angiographic disease, and lower baseline LVEF.7, 8, 9, 10, 11, 12, 13, 14, 15 While female sex and certain comorbidities have also been associated with lack of LVEF recovery, results have not been consistent.9,16, 17, 18 LVEF recovery has been associated with beta-blocker and renin-angiotensin-aldosterone system inhibitor prescription.15,16 However, existing studies did not follow the LVEF trajectory of patients with preserved LVEF at the time of MI. Few studies have evaluated the impact of LVEF trajectory after PCI on clinical outcomes5, 6, 7,9,11,12,15 reporting that patients who do not experience LVEF recovery are at higher risk of death and heart failure hospitalization independent of revascularization. Any incremental improvement in LVEF decreases the risk of adverse outcomes.15 These studies are limited by modest sample sizes with loss to echocardiographic follow-up and only one study evaluated outcomes beyond 5 years.11 Characterizing LVEF trajectory after MI and identifying patients at risk for left ventricular dysfunction may allow for targeted interventions to improve patient outcomes. We undertook this study to evaluate the impact of LVEF trajectory on long-term mortality and heart failure hospitalization in a large, international cohort of patients after primary PCI for MI in Sweden and the United States. Methods Data source The UC San Diego Health (UCSD) internal National Cardiovascular Data Registry CathPCI Registry was used to obtain data on patients who underwent PCI from January 2007 to September 2022 at UCSD, a university hospital system in the United States. The National Cardiovascular Data Registry CathPCI Registry contains comprehensive data on patients who underwent PCI at participating hospitals in the United States.19,20 All-cause mortality within 6 years post-PCI was obtained from the UCSD electronic medical record confirmed by the California Department of Public Health vital records and merged with registry data using unique patient medical record numbers. Heart failure hospitalization within 6 years post-PCI and LVEF on serial transthoracic echocardiograms were obtained from the UCSD electronic medical record and merged with registry data using unique patient medical record numbers. The Swedish Coronary Angiography and Angioplasty Registry (SCAAR) was used to obtain data on patients who underwent PCI in Sweden from June 2015 to December 2020. SCAAR is a population-based, mandated national registry that includes -coronary angiography and PCI procedures performed at all 29 hospitals in Sweden that provide PCI.21 All-cause mortality and heart failure hospitalization within 6 years post-PCI and LVEF on serial transthoracic echocardiograms were obtained from SCAAR. PCI procedures were performed by attending interventional cardiologists with or without interventional cardiology fellows. The Institutional Review Board of the University of California, San Diego, approved the study (#806572). Patient selection In the Swedish cohort, 13,484 patients underwent PCI for MI defined as ST-segment elevation MI (STEMI) or non–ST-segment elevation MI (NSTEMI) from 2015 to 2020 with available LVEF during the index admission and at follow-up. To minimize survivor bias and select an uncomplicated MI population, we excluded 28 patients over 80 years of age, 826 with cardiac arrest, and 179 with cardiogenic shock. The final Swedish cohort comprised 12,451 patients. In the U.S. cohort, 1,233 patients underwent PCI for MI from 2007 to 2022 with available LVEF during the index admission and at follow-up. Of these patients, 180 were over 80 years of age, 34 had cardiac arrest, and 31 had cardiogenic shock and were excluded. The final U.S. cohort comprised 988 patients. The total study cohort comprised 13,439 patients (Figure 1). An additional 37,522 patients met inclusion criteria but lacked an available LVEF trajectory. A comparison of baseline characteristics between the study cohort (n = 13,439) and those without an available LVEF trajectory (n = 37,522) is provided in Supplemental Table 1. Compared to the study cohort, patients with missing LVEF trajectory were older and had a higher prevalence of cardiovascular comorbidities.Figure 1Study FlowchartFlowchart illustrating inclusion and exclusion criteria. The final study population eligible for analysis consisted of 13,439 patients. decEF = decreased from preserved to reduced LVEF; LVEF = left ventricular ejection fraction; NSTEMI = non–ST-segment elevation myocardial infarction; PCI = percutaneous coronary intervention; pEF = consistently preserved LVEF; recEF = recovered LVEF; redEF = consistently reduced LVEF; SCAAR = Swedish Coronary Angiography and Angioplasty Registry; STEMI = ST-segment elevation myocardial infarction; UCSD = UC San Diego. Study Flowchart Flowchart illustrating inclusion and exclusion criteria. The final study population eligible for analysis consisted of 13,439 patients. decEF = decreased from preserved to reduced LVEF; LVEF = left ventricular ejection fraction; NSTEMI = non–ST-segment elevation myocardial infarction; PCI = percutaneous coronary intervention; pEF = consistently preserved LVEF; recEF = recovered LVEF; redEF = consistently reduced LVEF; SCAAR = Swedish Coronary Angiography and Angioplasty Registry; STEMI = ST-segment elevation myocardial infarction; UCSD = UC San Diego. Definitions Patients were categorized into LVEF groups based on their LVEF at the time of MI and at echocardiographic follow-up. LVEF was obtained during the MI admission, utilizing the measurement closest to the time of MI, and the method of LVEF assessment varied. Echocardiographic follow-up was variable, occurring between 30 days and 90 days post-MI. Consistently reduced LVEF (redEF) was defined as LVEF ≤50% at the time of MI and at follow-up. Recovered LVEF (recEF) was defined as LVEF ≤50% at the time of MI and LVEF >50% at follow-up. Consistently preserved LVEF (pEF) was defined as LVEF >50% at the time of MI and at follow-up. Decreased LVEF (decEF) was defined as LVEF >50% at the time of MI and LVEF ≤50% at follow-up. LVEF of 50% was chosen as a cutoff based on societal guidelines.22 Patients were also categorized based on the percent change in LVEF from the time of MI to echocardiographic follow-up. The predefined categories were no improvement, 10% to 20% improvement, >20% improvement, no reduction, 10% to 20% reduction, and >20% reduction in LVEF. Statistical methods The distribution of continuous variables was assessed using histograms. All continuous variables were normally distributed and presented as mean and SD. Group comparisons were performed using one-way analysis of variance or independent samples t-test. Categorical data are presented as counts with percentages and compared using the chi-square test. All-cause mortality and time to first heart failure hospitalization within 6 years post-PCI were evaluated with time-to-event analysis using Kaplan-Meier curves with log-rank statistic stratified by LVEF trajectory. Univariable cause-specific Cox proportional hazards regression model was used to estimate HRs with 95% CIs for all-cause mortality and heart failure hospitalization according to LVEF trajectory. Adjusted mortality and heart failure hospitalization analysis was conducted with a multivariable Cox proportional hazards regression model adjusting for age, sex, registry (UCSD or SCAAR), hypertension, previous MI, diabetes mellitus, PCI indication, angiographic findings, and previous coronary artery bypass graft surgery and presented as the adjusted HR (aHR) with 95% CI. The proportional hazards assumption was verified using Schoenfeld residuals. Subgroup analyses were conducted based on clinical presentation and sex. Analyses were performed using Stata Statistical Software: Release 18 (StataCorp LLC), a 2-sided P value <0.05 was considered statistically significant. The Sankey diagram was generated using open-source software (Sankeymatic.com). Results Baseline characteristics A total of 13,439 patients were included, with 12,451 in the Swedish cohort and 988 in the U.S. cohort. Baseline characteristics for the entire population are shown in Supplemental Table 2. The mean age was 64 ± 10 years, 24.5% were female, and 46.8% presented with STEMI while 53.2% presented with NSTEMI. Comorbidities included hypertension (31.5%), hyperlipidemia (31.1%), prior MI (13.8%), prior PCI (14.1%), and diabetes mellitus (24.4%). The majority of patients had radial access (85.2%) and non-left main one vessel disease (52.5%). Periprocedural and discharge medications are included in Supplemental Table 2. At follow-up, 28.2% had redEF, 14.8% had recEF, 53.0% had pEF, and 3.9% had decEF (Central Illustration). Patients with decEF had higher body mass index, a greater burden of comorbidities, and were more likely to have presented with STEMI compared to patients with pEF while there was no difference in age or sex between groups (Table 1). There was a greater proportion of patients with decEF in the U.S. cohort compared to the Swedish cohort. Patients with decEF were more likely to have received periprocedural bivalirudin, clopidogrel, or prasugrel and less likely to have received periprocedural aspirin and ticagrelor compared to patients with pEF. Patients with decEF were less likely to be prescribed ticagrelor or statins, and more likely to be prescribed beta-blockers at discharge compared to patients with pEF.Figure 2Unadjusted Heart Failure Hospitalization by Left Ventricular Ejection Fraction TrajectoryCumulative incidence curves of heart failure hospitalization during a 2000-day follow-up. Results are presented for (left) patients from SCAAR and (right) patients from UCSD. Outcome is stratified by LVEF trajectory. Abbreviations as in Figure 1.Figure 3Forest Plot of Unadjusted and Adjusted Mortality and Heart Failure HospitalizationAll-cause mortality and heart failure hospitalization were assessed as outcomes using univariable and multivariable Cox proportional hazards models at 2000-day follow-up. The multivariable model included the following variables: age, sex, registry (UCSD or SCAAR), hypertension, previous MI, diabetes mellitus, indication (NSTEMI or STEMI), angiographic findings, and previous CABG. Results are presented as HRs with 95% CIs and P values. CABG = coronary artery bypass graft surgery; other abbreviations as in Figure 1.Figure 4Unadjusted All-Cause Mortality by Left Ventricular Ejection Fraction TrajectoryCumulative incidence curves of all-cause mortality during a 2000-day follow-up. Results are presented for (left) patients from SCAAR and (right) patients from UCSD. Outcome is stratified by LVEF trajectory. Abbreviations as in Figure 1.Central IllustrationLeft Ventricular Ejection Fraction Trajectory and Long-Term Outcomes Following Percutaneous Coronary Intervention for Myocardial Infarction(Top) Sankey diagram illustrating left ventricular ejection fraction trajectory from baseline (left) to follow-up echocardiography (right). The width of each line reflects the proportion of patients in each group. (Bottom Left) Cumulative incidence curves of heart failure hospitalization during a 2000-day follow-up across LVEF trajectory groups. (Bottom Right) Cumulative incidence curves of all-cause mortality during a 2000-day follow-up across LVEF trajectory groups. MI = myocardial infarction; NCDR = National Cardiovascular Data Registry; other abbreviations as in Figure 1.Table 1Baseline Characteristics of Consistently Preserved LVEF vs Decreased From Preserved to Reduced LVEFpEFdecEFP Value(n = 7,127)(n = 529)SCAAR registry6,640 (93.2%)403 (76.2%)<0.001Patient characteristics Age, y63.9 (9.5)64.2 (10.0)0.46 Female1,835 (25.7%)128 (24.2%)0.43 BMI, kg/m228.2 (6.3)28.9 (5.3)0.014Past medical history Smoker1,746 (24.8%)152 (29.5%)0.019 Hypertension2,166 (30.4%)248 (46.9%)<0.001 Hyperlipidemia2,218 (31.3%)237 (44.9%)<0.001 Prior MI858 (12.0%)135 (25.5%)<0.001 Prior PCI931 (13.1%)128 (24.2%)<0.001 Prior CABG276 (3.9%)46 (8.7%)<0.001 Dialysis52 (0.7%)19 (3.6%)<0.001 Prior stroke304 (4.3%)41 (7.8%)<0.001 Peripheral artery disease226 (3.2%)42 (7.9%)<0.001 Chronic lung disease355 (5.0%)42 (7.9%)0.003 Diabetes mellitus1,543 (21.7%)176 (33.3%)<0.001 Previous heart failure192 (2.7%)66 (12.5%)<0.001Procedural characteristics Arterial access Radial6,197 (87.0%)395 (74.7%)<0.001 Femoral853 (12.0%)126 (23.8%) Other77 (1.1%)8 (1.5%) Angiographic findings 1 VD3,757 (54.4%)261 (51.5%)0.16 2 VD1,901 (27.5%)139 (27.4%) 3 VD972 (14.1%)84 (16.6%) 1 VD + LM58 (0.8%)1 (0.2%) 2 VD + LM95 (1.4%)11 (2.2%) 3 VD + LM110 (1.6%)11 (2.2%) LM13 (0.2%)0 (0.0%) STEMI indication2,617 (36.7%)233 (44.0%)<0.001Periprocedural medications Unfractionated heparin6,917 (97.5%)508 (96.2%)0.079 Aspirin6,975 (98.3%)510 (96.6%)0.005 Bivalirudin684 (9.6%)83 (15.7%)<0.001 Clopidogrel233 (3.3%)79 (14.9%)<0.001 Ticagrelor6,272 (88.4%)378 (71.6%)<0.001 Prasugrel45 (0.6%)16 (3.0%)<0.001Discharge medications Aspirin6,553 (91.9%)461 (87.1%)<0.001 Ticagrelor6,013 (84.4%)334 (63.1%)<0.001 Clopidogrel889 (12.5%)154 (29.1%)<0.001 Statin6,734 (94.5%)480 (90.7%)<0.001 Beta-blocker5,503 (77.2%)436 (82.4%)0.006 RAAS inhibitor5,507 (77.3%)395 (74.7%)0.17Values are mean ± SD or n (%).pEF = consistently preserved LVEF; decEF = decreased from preserved to reduced LVEF; LVEF = left ventricular ejection fraction; SCAAR = Swedish Coronary Angiography and Angioplasty Registry; BMI = body mass index; MI = myocardial infarction; PCI = percutaneous coronary intervention; CABG = coronary artery bypass graft surgery; VD = vessel disease; LM = left main; STEMI = ST-elevation myocardial infarction; RAAS = renin-angiotensin-aldosterone system. Unadjusted Heart Failure Hospitalization by Left Ventricular Ejection Fraction Trajectory Cumulative incidence curves of heart failure hospitalization during a 2000-day follow-up. Results are presented for (left) patients from SCAAR and (right) patients from UCSD. Outcome is stratified by LVEF trajectory. Abbreviations as in Figure 1. Forest Plot of Unadjusted and Adjusted Mortality and Heart Failure Hospitalization All-cause mortality and heart failure hospitalization were assessed as outcomes using univariable and multivariable Cox proportional hazards models at 2000-day follow-up. The multivariable model included the following variables: age, sex, registry (UCSD or SCAAR), hypertension, previous MI, diabetes mellitus, indication (NSTEMI or STEMI), angiographic findings, and previous CABG. Results are presented as HRs with 95% CIs and P values. CABG = coronary artery bypass graft surgery; other abbreviations as in Figure 1. Unadjusted All-Cause Mortality by Left Ventricular Ejection Fraction Trajectory Cumulative incidence curves of all-cause mortality during a 2000-day follow-up. Results are presented for (left) patients from SCAAR and (right) patients from UCSD. Outcome is stratified by LVEF trajectory. Abbreviations as in Figure 1. Left Ventricular Ejection Fraction Trajectory and Long-Term Outcomes Following Percutaneous Coronary Intervention for Myocardial Infarction (Top) Sankey diagram illustrating left ventricular ejection fraction trajectory from baseline (left) to follow-up echocardiography (right). The width of each line reflects the proportion of patients in each group. (Bottom Left) Cumulative incidence curves of heart failure hospitalization during a 2000-day follow-up across LVEF trajectory groups. (Bottom Right) Cumulative incidence curves of all-cause mortality during a 2000-day follow-up across LVEF trajectory groups. MI = myocardial infarction; NCDR = National Cardiovascular Data Registry; other abbreviations as in Figure 1. Table 1Baseline Characteristics of Consistently Preserved LVEF vs Decreased From Preserved to Reduced LVEFpEFdecEFP Value(n = 7,127)(n = 529)SCAAR registry6,640 (93.2%)403 (76.2%)<0.001Patient characteristics Age, y63.9 (9.5)64.2 (10.0)0.46 Female1,835 (25.7%)128 (24.2%)0.43 BMI, kg/m228.2 (6.3)28.9 (5.3)0.014Past medical history Smoker1,746 (24.8%)152 (29.5%)0.019 Hypertension2,166 (30.4%)248 (46.9%)<0.001 Hyperlipidemia2,218 (31.3%)237 (44.9%)<0.001 Prior MI858 (12.0%)135 (25.5%)<0.001 Prior PCI931 (13.1%)128 (24.2%)<0.001 Prior CABG276 (3.9%)46 (8.7%)<0.001 Dialysis52 (0.7%)19 (3.6%)<0.001 Prior stroke304 (4.3%)41 (7.8%)<0.001 Peripheral artery disease226 (3.2%)42 (7.9%)<0.001 Chronic lung disease355 (5.0%)42 (7.9%)0.003 Diabetes mellitus1,543 (21.7%)176 (33.3%)<0.001 Previous heart failure192 (2.7%)66 (12.5%)<0.001Procedural characteristics Arterial access Radial6,197 (87.0%)395 (74.7%)<0.001 Femoral853 (12.0%)126 (23.8%) Other77 (1.1%)8 (1.5%) Angiographic findings 1 VD3,757 (54.4%)261 (51.5%)0.16 2 VD1,901 (27.5%)139 (27.4%) 3 VD972 (14.1%)84 (16.6%) 1 VD + LM58 (0.8%)1 (0.2%) 2 VD + LM95 (1.4%)11 (2.2%) 3 VD + LM110 (1.6%)11 (2.2%) LM13 (0.2%)0 (0.0%) STEMI indication2,617 (36.7%)233 (44.0%)<0.001Periprocedural medications Unfractionated heparin6,917 (97.5%)508 (96.2%)0.079 Aspirin6,975 (98.3%)510 (96.6%)0.005 Bivalirudin684 (9.6%)83 (15.7%)<0.001 Clopidogrel233 (3.3%)79 (14.9%)<0.001 Ticagrelor6,272 (88.4%)378 (71.6%)<0.001 Prasugrel45 (0.6%)16 (3.0%)<0.001Discharge medications Aspirin6,553 (91.9%)461 (87.1%)<0.001 Ticagrelor6,013 (84.4%)334 (63.1%)<0.001 Clopidogrel889 (12.5%)154 (29.1%)<0.001 Statin6,734 (94.5%)480 (90.7%)<0.001 Beta-blocker5,503 (77.2%)436 (82.4%)0.006 RAAS inhibitor5,507 (77.3%)395 (74.7%)0.17Values are mean ± SD or n (%).pEF = consistently preserved LVEF; decEF = decreased from preserved to reduced LVEF; LVEF = left ventricular ejection fraction; SCAAR = Swedish Coronary Angiography and Angioplasty Registry; BMI = body mass index; MI = myocardial infarction; PCI = percutaneous coronary intervention; CABG = coronary artery bypass graft surgery; VD = vessel disease; LM = left main; STEMI = ST-elevation myocardial infarction; RAAS = renin-angiotensin-aldosterone system. Baseline Characteristics of Consistently Preserved LVEF vs Decreased From Preserved to Reduced LVEF Values are mean ± SD or n (%). pEF = consistently preserved LVEF; decEF = decreased from preserved to reduced LVEF; LVEF = left ventricular ejection fraction; SCAAR = Swedish Coronary Angiography and Angioplasty Registry; BMI = body mass index; MI = myocardial infarction; PCI = percutaneous coronary intervention; CABG = coronary artery bypass graft surgery; VD = vessel disease; LM = left main; STEMI = ST-elevation myocardial infarction; RAAS = renin-angiotensin-aldosterone system. Patients with redEF were older, had a greater proportion of men, more comorbidities, and were more likely to have presented with STEMI compared to patients with recEF (Table 2). They were also more likely to have had angiographic multivessel disease. There was a greater proportion of patients with redEF in the Swedish cohort compared to the U.S. cohort. Patients with redEF were less likely to have received periprocedural bivalirudin and ticagrelor compared to patients with recEF. Patients with redEF were less likely to be prescribed aspirin, ticagrelor, or statins compared to patients with recEF.Table 2Baseline Characteristics of Recovered LVEF vs Consistently Reduced LVEFrecEFredEFP Value(n = 1,995)(n = 3,788)SCAAR registry1,830 (91.7%)3,578 (94.5%)<0.001Patient characteristics Age, y63.6 (9.6)65.4 (9.4)<0.001 Female521 (26.1%)805 (21.3%)<0.001 BMI, kg/m227.9 (7.7)28.0 (5.2)0.93Past medical history Smoker491 (25.1%)1,100 (29.8%)<0.001 Hypertension546 (27.4%)1,278 (33.7%)<0.001 Hyperlipidemia499 (25.2%)1,200 (31.9%)<0.001 Prior MI173 (8.7%)693 (18.3%)<0.001 Prior PCI176 (8.8%)656 (17.3%)<0.001 Prior CABG51 (2.6%)248 (6.5%)<0.001 Dialysis22 (1.1%)51 (1.3%)0.43 Prior stroke86 (4.3%)232 (6.1%)0.004 Peripheral artery disease57 (2.9%)201 (5.3%)<0.001 Chronic lung disease96 (4.8%)248 (6.5%)0.008 Diabetes mellitus437 (21.9%)1,122 (29.6%)<0.001 Previous heart failure105 (5.3%)422 (11.1%)<0.001Procedural characteristics Arterial access Radial1,711 (85.8%)3,152 (83.2%)0.04 Femoral259 (13.0%)576 (15.2%) Other25 (1.3%)60 (1.6%) Angiographic findings 1 VD1,084 (55.5%)1,759 (47.4%)<0.001 2 VD496 (25.4%)1,005 (27.1%) 3 VD269 (13.8%)733 (19.8%) 1 VD + LM24 (1.2%)40 (1.1%) 2 VD + LM33 (1.7%)51 (1.4%) 3 VD + LM44 (2.3%)113 (3.0%) LM3 (0.2%)9 (0.2%) STEMI indication1,145 (57.4%)2,298 (60.7%)0.016Periprocedural medications Unfractionated heparin1,934 (97.2%)3,658 (97.1%)0.78 Aspirin1,957 (98.3%)3,706 (98.4%)0.97 Bivalirudin233 (11.7%)358 (9.5%)0.009 Clopidogrel75 (3.8%)127 (3.4%)0.42 Ticagrelor1,727 (86.8%)3,166 (84.0%)0.005 Prasugrel16 (0.8%)31 (0.8%)0.94Discharge medications Aspirin1,823 (91.4%)3,236 (85.4%)<0.001 Ticagrelor1,602 (80.3%)2,703 (71.4%)<0.001 Clopidogrel327 (16.4%)852 (22.5%)<0.001 Statin1,877 (94.1%)3,459 (91.3%)<0.001 Beta-blocker1,766 (88.5%)3,341 (88.2%)0.72 RAAS inhibitor1,758 (88.1%)3,272 (86.4%)0.06Values are mean ± SD or n (%).recEF = recovered LVEF; redEF = consistently reduced LVEF; other abbreviations as in Table 1. Table 2Baseline Characteristics of Recovered LVEF vs Consistently Reduced LVEFrecEFredEFP Value(n = 1,995)(n = 3,788)SCAAR registry1,830 (91.7%)3,578 (94.5%)<0.001Patient characteristics Age, y63.6 (9.6)65.4 (9.4)<0.001 Female521 (26.1%)805 (21.3%)<0.001 BMI, kg/m227.9 (7.7)28.0 (5.2)0.93Past medical history Smoker491 (25.1%)1,100 (29.8%)<0.001 Hypertension546 (27.4%)1,278 (33.7%)<0.001 Hyperlipidemia499 (25.2%)1,200 (31.9%)<0.001 Prior MI173 (8.7%)693 (18.3%)<0.001 Prior PCI176 (8.8%)656 (17.3%)<0.001 Prior CABG51 (2.6%)248 (6.5%)<0.001 Dialysis22 (1.1%)51 (1.3%)0.43 Prior stroke86 (4.3%)232 (6.1%)0.004 Peripheral artery disease57 (2.9%)201 (5.3%)<0.001 Chronic lung disease96 (4.8%)248 (6.5%)0.008 Diabetes mellitus437 (21.9%)1,122 (29.6%)<0.001 Previous heart failure105 (5.3%)422 (11.1%)<0.001Procedural characteristics Arterial access Radial1,711 (85.8%)3,152 (83.2%)0.04 Femoral259 (13.0%)576 (15.2%) Other25 (1.3%)60 (1.6%) Angiographic findings 1 VD1,084 (55.5%)1,759 (47.4%)<0.001 2 VD496 (25.4%)1,005 (27.1%) 3 VD269 (13.8%)733 (19.8%) 1 VD + LM24 (1.2%)40 (1.1%) 2 VD + LM33 (1.7%)51 (1.4%) 3 VD + LM44 (2.3%)113 (3.0%) LM3 (0.2%)9 (0.2%) STEMI indication1,145 (57.4%)2,298 (60.7%)0.016Periprocedural medications Unfractionated heparin1,934 (97.2%)3,658 (97.1%)0.78 Aspirin1,957 (98.3%)3,706 (98.4%)0.97 Bivalirudin233 (11.7%)358 (9.5%)0.009 Clopidogrel75 (3.8%)127 (3.4%)0.42 Ticagrelor1,727 (86.8%)3,166 (84.0%)0.005 Prasugrel16 (0.8%)31 (0.8%)0.94Discharge medications Aspirin1,823 (91.4%)3,236 (85.4%)<0.001 Ticagrelor1,602 (80.3%)2,703 (71.4%)<0.001 Clopidogrel327 (16.4%)852 (22.5%)<0.001 Statin1,877 (94.1%)3,459 (91.3%)<0.001 Beta-blocker1,766 (88.5%)3,341 (88.2%)0.72 RAAS inhibitor1,758 (88.1%)3,272 (86.4%)0.06Values are mean ± SD or n (%).recEF = recovered LVEF; redEF = consistently reduced LVEF; other abbreviations as in Table 1. Baseline Characteristics of Recovered LVEF vs Consistently Reduced LVEF Values are mean ± SD or n (%). recEF = recovered LVEF; redEF = consistently reduced LVEF; other abbreviations as in Table 1. Unadjusted and adjusted heart failure hospitalization post-PCI Patients with redEF had the highest cumulative incidence of heart failure hospitalization, followed by those with decEF, recEF, and pEF (40.0% vs 30.4% vs 16.1% vs 6.4%; P < 0.001) (Central Illustration). Findings were consistent in the United States (36.5% vs 24.1% vs 12.2% vs 7.3%; P < 0.001) and Swedish cohorts (40.3% vs 31.5% vs 17.4% vs 6.5%; P < 0.001) (Figure 2). The cumulative incidence of heart failure hospitalization increased in all groups during the first 2 years, and then plateaued. Similar findings were observed in the subgroup analysis of STEMI and NSTEMI (Supplemental Figure 1). In contrast to men, women with decEF had the highest cumulative incidence of heart failure hospitalization, followed by redEF, recEF, and pEF (44.0% vs 40.6% vs 20.3% vs 7.5%; P < 0.001) (Supplemental Figure 2). In the adjusted analysis using patients with pEF as the reference, patients with redEF had the highest HR for heart failure hospitalization (aHR: 8.55; 95% CI: 7.53-9.70; P < 0.001), followed by patients with decEF (aHR: 6.28; 95% CI: 5.12-7.69; P < 0.001) and recEF (aHR: 3.23; 95% CI: 2.74-3.82; P < 0.001) (Figure 3). Unadjusted and adjusted all-cause mortality post-PCI The cumulative incidence of all-cause mortality was higher in patients with decEF or redEF compared to those with recEF or pEF (14.9%, 14.0% vs 7.8%, 6.7%; P < 0.001) (Central Illustration). Findings were consistent in the United States (12.9%, 10.3% vs 7.9%, 6.5%; P < 0.001) and Swedish cohorts (29.6%, 23.3% vs 9.6%, 9.2%; P < 0.001) (Figure 4) in patients with NSTEMI and STEMI (Supplemental Figure 1). Similar findings were observed among men, but women with decEF had the highest cumulative incidence of mortality followed by redEF, recEF, and pEF (24.4% vs 15.9% vs 8.5% vs 8.3%; P < 0.001) (Supplemental Figure 2). In the adjusted analysis using patients with pEF as the reference, patients with redEF (aHR: 2.21; 95% CI: 1.87-2.61; P < 0.001) and decEF (aHR: 1.96; 95% CI: 1.45-2.64; P < 0.001) had higher HRs for all-cause mortality compared to patients with recEF (aHR: 1.28; 95% CI: 1.01-1.63; P = 0.043) (Figure 3). Outcomes by percent change in LVEF Among patients with preserved baseline LVEF, those with a >20% reduction in LVEF at follow-up had a significantly higher cumulative incidence of all-cause mortality (29.3% vs 9.8% vs 6.7%; P < 0.001) and heart failure hospitalization (50.1% vs 23.3% vs 6.4%; P < 0.001) compared to those with a 10% to 20% reduction or no reduction (Supplemental Figure 3). Among patients with reduced baseline LVEF, those with a >20% or 10% to 20% improvement had a lower cumulative incidence of all-cause mortality compared to those with no improvement (10.6% vs 9.0% vs 14.1%; P < 0.001). Patients with no improvement in LVEF had the highest cumulative incidence of heart failure hospitalization, followed by those with a >20% improvement and a 10% to 20% improvement (38.4% vs 29.9% vs 23.2%; P < 0.001). There was a higher proportion of patients with baseline LVEF <40% in the group with a >20% improvement compared to the group with a 10% to 20% improvement (96% vs 37%). Discussion In this large, multicenter international cohort of patients who underwent PCI for MI in Sweden and the United States, we found that LVEF trajectory has a significant impact on long-term mortality and heart failure hospitalization, with consistent prognostic value observed across both cohorts. To our knowledge, this is the largest study of its kind and the first to characterize patients with preserved LVEF at the time of MI who subsequently experience LVEF reduction despite PCI. The findings of this study suggest that LVEF trajectory is a powerful predictor of long-term outcomes in patients with MI. Despite the widespread use of timely primary PCI with drug-eluting stents, potent P2Y12 inhibitors, and guideline-directed medical therapy for heart failure, nearly one-third of patients in this study had persistently reduced LVEF after MI. Multivessel disease and STEMI presentation were associated with redEF, consistent with previous findings.9,11 This is likely due to a larger territory of ischemia and transmural infarction, resulting in the formation of unsalvageable scar tissue. A subset of these patients may have had pre-existing ischemic cardiomyopathy and PCI may not have conferred benefit in improving LVEF in this population.23 We observed a smaller proportion of patients with LVEF recovery compared to a previous study in a younger cohort.11 Patients who experienced LVEF recovery were more likely to have been prescribed ticagrelor at discharge. Ticagrelor has been shown to protect against reperfusion injury and prevent left ventricular remodeling unlike clopidogrel, which may explain the association with LVEF recovery.24,25 The importance of initiating guideline-directed medical therapy during the index hospitalization and titrating to target doses during outpatient follow-up to maximize the potential for LVEF recovery warrants emphasis. Renin-angiotensin-aldosterone system inhibitors, proven to attenuate left ventricular remodeling and reduce adverse cardiovascular events in patients with ischemic cardiomyopathy and anterior wall MI, were highly prescribed in this cohort.26, 27, 28 Data on contemporary guideline-directed medical therapy agents, such as angiotensin receptor neprilysin inhibitors and sodium glucose co-transporter 2 inhibitors, were unavailable. In the immediate post-MI setting, angiotensin receptor neprilysin inhibitor therapy has not demonstrated superiority over renin-angiotensin-aldosterone system inhibitors, and sodium glucose co-transporter 2 inhibitors have not consistently shown efficacy among patients with left ventricular dysfunction.29, 30, 31 Nevertheless, given their established safety and proven benefit in chronic heart failure with reduced ejection fraction, these agents should be initiated in patients with reduced LVEF following MI. In addition to optimizing guideline-directed medical therapy to promote LVEF recovery, continued therapy after recovery may help maintain cardiac function and prevent relapse.32 Interventions to improve guideline-directed medical therapy prescription and adherence, such as decision support tools within the electronic health record and patient and provider education, should be employed in this regard. Whether guideline-directed medical therapy agents can prevent LVEF reduction in acute MI patients with initially preserved LVEF remains a subject for future study. Conversely, we found that beta-blocker prescription at discharge among patients with preserved LVEF was associated with a reduction in LVEF at follow-up. Although beta-blockers are effective in post-MI patients with reduced LVEF, they provided no clinical benefit in patients with preserved LVEF undergoing early coronary angiography.33 It is unknown whether the association between beta-blocker prescription and LVEF decline in this study reflects confounding by indication or a direct deleterious effect of therapy. Nevertheless, these findings reinforce recent data questioning the utility of routine beta-blocker prescription in this population. A small but notable subset of patients experienced a deterioration from initially preserved LVEF to reduced LVEF at follow-up, despite undergoing PCI, contrary to previous studies.34,35 These patients had higher body mass index, more comorbidities, and were more likely to have presented with STEMI. There was a higher proportion of patients with decEF in the U.S. cohort. Managing comorbidities and promoting lifestyle modifications are crucial components of post-MI care. Given this observed LVEF deterioration in a subset of patients with initially preserved LVEF, our data suggest the potential prognostic utility of a single, well-timed surveillance echocardiogram for patients post-MI, irrespective of initial LVEF. Current U.S. and European practice guidelines only recommend repeat LVEF assessment in patients with reduced LVEF at the time of MI.36,37 In practice, however, even these patients often do not receive follow-up echocardiograms.38 LVEF trajectory had a significant impact on the long-term risk of heart failure hospitalization and mortality. These findings were consistent in both countries and persisted in the adjusted analysis. Patients with redEF had the highest rate of heart failure hospitalization. While LVEF recovery was associated with a reduction in risk by more than half, the risk remained over twice that of patients with pEF. This suggests that while LVEF trajectory provides vital information about changes in risk, the initial severity of myocardial injury, reflected by the baseline LVEF, continues to contribute to long-term outcomes. The seemingly paradoxical finding of greater LVEF improvement being associated with an increased risk of heart failure hospitalization is likely driven by the lower LVEF at baseline in these patients. Therefore, initial LVEF may serve as a baseline indicator of risk, while LVEF trajectory provides a dynamic assessment to refine risk stratification and inform ongoing management. These findings underscore that LVEF at a single time point alone may not fully capture the complexity of myocardial injury. Both redEF and deterioration in LVEF from preserved to reduced were associated with the highest mortality rates, with no difference in mortality between the 2 groups. The magnitude of LVEF reduction directly correlated with the risk of all-cause mortality and heart failure hospitalization. Conversely, patients with LVEF recovery had half the mortality risk which was no different than the mortality risk observed in patients with pEF. Women with decEF had the highest mortality rate, more than double that of men with decEF. They also had the highest heart failure hospitalization rates. These findings align with a previous study that demonstrated increased mortality and heart failure risk among women with acute coronary syndromes.39 Further research is needed to elucidate the underlying mechanisms driving sex differences in outcomes. Additionally, increasing the representation of women in coronary artery disease clinical trials may allow for the development of sex-specific treatment strategies. Patient and procedure-level factors including older age, comorbidities, and the severity of angiographic disease were associated with LVEF reduction. In addition to close follow-up, these patients may benefit from adjunctive therapies such as supersaturated oxygen which has been shown to reduce infarct size by mitigating ischemia-reperfusion injury and microvascular obstruction in anterior STEMI.40 Therapies targeting pro-inflammatory pathways during acute MI are also currently under investigation.41 To improve outcomes for MI patients, a comprehensive approach is needed that extends beyond PCI. While our study demonstrates the prognostic significance of LVEF trajectory, prospective studies are needed to determine the clinical utility of surveillance echocardiography after MI and whether implementing targeted therapies to modify this trajectory can improve clinical outcomes. Study limitations The limitations of this study include its retrospective nature using registry data and the inherent heterogeneity of the U.S. and Swedish cohorts. We compared all hospitals in Sweden to a large university hospital system in the United States, which may not fully capture outcomes in the entire United States. The duration of echocardiographic follow-up was not standardized and may have been provider-initiated, introducing potential selection bias. While mortality is a competing risk for heart failure hospitalization, a cause-specific approach was employed to evaluate the etiological association between LVEF trajectory and heart failure morbidity. In the SCAAR registry, LVEF and LVEF change were reported as categorical variables which precluded the analysis of continuous absolute values and may result in misclassification near the 50% threshold. Only the first follow-up echocardiogram was available within the registries, which precluded the assessment of additional LVEF changes over time. LVEF was evaluated as a trajectory rather than a time-dependent covariate to identify distinct clinical phenotypes of cardiac remodeling and their effect on outcomes. Furthermore, echocardiographic data were not evaluated by a core laboratory, and thus interobserver variability and image quality variance may impact the reliability of LVEF assessment. Nonetheless, sensitivity analyses using absolute LVEF change demonstrated a graded association between the magnitude of LVEF decline and adverse outcomes, suggesting these findings reflect clinically significant changes in cardiac function. Finally, data on prescriptions for mineralocorticoid receptor antagonists, sodium glucose co-transporter 2 inhibitors, and angiotensin receptor neprilysin inhibitors were unavailable. Conclusions LVEF trajectory after MI significantly impacts long-term outcomes. Even patients with LVEF recovery remain at increased risk for heart failure hospitalization and mortality compared to those with pEF.PerspectivesCOMPETENCY IN MEDICAL KNOWLEDGE: This international, multicenter study is among the largest to demonstrate the prognostic significance of LVEF trajectory following PCI for MI. The analysis identified a subset of patients with preserved LVEF at the time of MI who subsequently experience LVEF reduction despite successful revascularization. These patients exhibit a high risk for heart failure hospitalization and mortality similar to those with redEF. Patients with recEF remain at significantly higher risk for these outcomes compared to those with pEF. These findings illustrate that baseline LVEF may not fully capture long-term prognosis.TRANSLATIONAL OUTLOOK: Longitudinal assessment of LVEF trajectory may provide incremental prognostic value in patients following MI, particularly for identifying high-risk cohorts not captured by a single baseline measurement. Prospective trials are required to determine if a strategy of routine serial echocardiography and subsequent trajectory-guided management can improve clinical outcomes compared to the current standard of care. COMPETENCY IN MEDICAL KNOWLEDGE: This international, multicenter study is among the largest to demonstrate the prognostic significance of LVEF trajectory following PCI for MI. The analysis identified a subset of patients with preserved LVEF at the time of MI who subsequently experience LVEF reduction despite successful revascularization. These patients exhibit a high risk for heart failure hospitalization and mortality similar to those with redEF. Patients with recEF remain at significantly higher risk for these outcomes compared to those with pEF. These findings illustrate that baseline LVEF may not fully capture long-term prognosis. TRANSLATIONAL OUTLOOK: Longitudinal assessment of LVEF trajectory may provide incremental prognostic value in patients following MI, particularly for identifying high-risk cohorts not captured by a single baseline measurement. Prospective trials are required to determine if a strategy of routine serial echocardiography and subsequent trajectory-guided management can improve clinical outcomes compared to the current standard of care. Funding support and author disclosures This work was supported by an unrestricted educational grant from the Warren Family Foundation. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.