We conducted a retrospective study of patients with ischemic cardiomyopathy who received ECMO-assisted PCI at our institution from January 2013 to December 2022. The institutional review board (Ethics approval number: 2023ZH135) approved the study protocol, and the need for informed consent was waived due to the retrospective nature of the study. We included patients who (1) were aged 18 years or older, (2) had a left ventricular ejection fraction (LVEF) $\le$35% on pre-procedure transthoracic echocardiography (TTE), and (3) had hemodynamic instability requiring ECMO support before or during PCI, including patients with STEMI who were in cardiogenic shock and patients with complex, high-risk anatomy. We excluded patients who (1) had other types of cardiomyopathy (e.g., dilated cardiomyopathy, hypertrophic cardiomyopathy), (2) had a previous cardiac transplantation, and (3) had incomplete medical records.

We collected data on demographic characteristics, medical history, laboratory tests, angiographic findings, procedural details, ECMO parameters, and in-hospital complications from the electronic medical records. The SYNTAX score was ascertained using the online SYNTAX score calculator (http://www.syntaxscore.com/calculator/start.htm). The primary outcome was the change in LV functional remodeling from baseline to 12 months post-procedure, measured by transthoracic echocardiography. The secondary outcomes were (1) the incidence of major adverse cardiac and cerebrovascular events (MACCEs), which was a composite outcome of death, reinfarction, cerebrovascular accident, or repeat revascularization. MACCEs were identified using the International Classification of Diseases, Clinical Modification, Tenth Revision (ICD–10-CM) diagnosis codes [7]; (2) the incidence of ECMO-related complications, such as bleeding, infection, limb ischemia, or hemolysis; (3) the length of hospital stay and the intensive care unit stay. Follow-up data were obtained from outpatient visits, telephone interviews, or the national death registry.

Our institution is a high-volume tertiary center with extensive experience in advanced heart failure therapies and a dedicated ECMO team comprising cardiovascular surgeons, interventional cardiologists, perfusionists, and intensive care specialists. The decision to utilize ECMO, LVAD, or other therapies, including surgical revascularization (e.g., CABG), is based on a multidisciplinary evaluation of each patient’s clinical status, comorbidities, coronary anatomy, risk profile, and overall prognosis. In patients with ischemic cardiomyopathy, ECMO support was initiated based on hemodynamic instability, cardiogenic shock, or significant left ventricular dysfunction as judged by the treating cardiologist and cardiac surgeon. We excluded patients with severe peripheral vascular disease, active bleeding, or mechanical complications of myocardial infarction. The patients underwent general anesthesia before ECMO implantation. Vascular access, selection of revascularization strategy and devices, use of hemodynamic support devices, and periprocedural pharmacotherapy were at the discretion of the operator.

Venoarterial ECMO was instituted percutaneously or surgically with cannulation of the common femoral vein and artery. Anticoagulation was achieved with heparin to maintain an activated clotting time of 180–220 seconds. The ECMO circuit consisted of a centrifugal pump and hollow-fiber microporous membrane oxygenator. The ECMO flow rate was adjusted to maintain a mean arterial pressure of 60–80 mmHg and a mixed venous oxygen saturation of $>$70%. Distal limb perfusion was maintained during femoral cannulation. PCI was performed via the femoral or radial artery using contemporary techniques. Thrombus aspiration, predilatation, and drug-eluting stent implantation were performed at the operator’s discretion. An intra-aortic balloon pump was used selectively. The operator determined the PCI strategy to achieve complete revascularization whenever feasible. Completeness of revascularization was defined angiographically as no residual stenosis $\ge$70% in major epicardial vessels $\ge$2.5 mm in diameter and all stenotic lesions in branch vessels supplying viable myocardium. ECMO weaning was attempted after hemodynamic stabilization, defined as a cardiac index of $>$2.2 L/min/m², a central venous pressure of 15 mmHg, and a lactate level of 2 mmol/L. After surgery, patients were admitted to the intensive care unit and received routine postoperative care. Mechanical ventilation was discontinued when patients were normothermic, hemodynamically stable, and awake.

The echocardiographic measurements were conducted in accordance with the guidelines established by the American Society of Echocardiography and the European Society of Echocardiography [8]. LVEF was measured via the biplane Simpson’s method by combining apical four- and two-chamber views. The wall motion score index (WMSI) was assessed using a 16-segment model of the LV. Each segment with a clear endocardial border was scored as follows: 1 for normal, 2 for hypokinesis, 3 for akinesis, and 4 for dyskinesis. Global WMSI was calculated by summating the scores divided by the number of analyzed segments. Additionally, end-diastolic volume index (EDVI) and end-systolic volume index (ESVI) were calculated using the modified Simpson’s rule and indexed to body surface area.

After the ECMO-assisted PCI procedure, patients were prescribed guideline-directed medical therapy (GDMT) for heart failure with reduced ejection fraction, according to the current guidelines from the American College of Cardiology/American Heart Association and the European Society of Cardiology. The GDMT included angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs), beta-blockers, mineralocorticoid receptor antagonists (MRAs), and sodium–glucose co-transporter 2 inhibitors (SGLT2i) when indicated. Diuretics were prescribed as needed to manage volume overload. Antiplatelet therapy, statins, and other secondary prevention medications were also prescribed as appropriate.

During the follow-up period, patients were closely monitored in the outpatient clinic, with regular assessments of clinical status, laboratory tests, and echocardiographic evaluations. Medication adjustments were made based on the patient’s response, tolerability, and any adverse effects. Cardiac rehabilitation and lifestyle modifications, including exercise, dietary changes, and smoking cessation, were strongly encouraged.

We expressed continuous variables as mean $\pm$ standard deviation or median (interquartile range (IQR)), depending on their distribution, and compared them using Student’s t-test or Wilcoxon signed-rank test, respectively. We presented categorical variables as counts or percentages and compared them using the appropriate chi-squared or Fisher’s exact test. A generalized linear mixed-effects model (GLMM) was used to account for repeated measures within patients and to model the longitudinal changes in LVEF, WMSI, EDVI, and ESVI over time. To investigate potential differences in the trends of LV functional remodeling between subgroups, we performed additional GLMM analyses. We compared the LVEF, WMSI, EDVI, and ESVI trends between survivors and non-survivors and between patients who achieved complete revascularization and those who did not. The GLMM included fixed effects for time, subgroup (survivors vs. non-survivors or complete vs. incomplete revascularization), and the interaction between time and subgroup while also accounting for repeated measurements per patient over time. The GLMM for LVED, WMSI, EDVI, and ESVI included fixed effects for time and the aforementioned covariates (clinical indication, baseline values, multivessel disease, collateral circulation, previous MI, postoperative re-infarction, and complete revascularization). These covariates were included as fixed effects in the GLMM to adjust for their potential influence on the LV functional parameters over time. Statistical significance was defined as a two-sided p-value of 0.05. The statistical software used throughout the analysis was SPSS v.26.0 (IBM SPSS Inc., Armonk, NY, USA) and R 4.0.1 (R Foundation for Statistical Computing, Vienna, Austria).

The study comprised 76 patients with ischemic cardiomyopathy who underwent ECMO-assisted PCI. Table 1 shows the baseline characteristics of the patients. The mean age was 62.67 $\pm$ 8.34 years and 66% was male. The majority of patients had a history of hypertension (70%), 33% of the patients had a previous myocardial infarction and 17% had received coronary artery bypass grafting. The mean baseline LVEF was 29.98 $\pm$ 2.65%. The majority of patients (n = 68, 89%) presented with cardiogenic shock, while the remaining patients were considered at high risk for anatomy during an elective PCI (n = 3, 4%), or had hemodynamic instability during an elective PCI procedure (n = 5, 7%). Most patients (88%) had a Killip classification of III or IV. Thirteen patients (17%) had previously received coronary artery bypass grafting. The mean SYNTAX score was 29.1 $\pm$ 6.4 in patients without previous coronary artery bypass grafting (n = 63).

Table 2 shows the details and outcomes of the ECMO-assisted PCI procedure. Emergent PCI was performed in 62% of patients. The median time of the procedure was 48.14 $\pm$ 10.53 minutes. Multivessel disease was present in 82% of the patients, and chronic total occlusion (CTO) was present in 13%. The rates of complete revascularization and CTO treatment were 58% and 29%, respectively. The number of stents implanted was 2 (IQR: 2–3). The median length of stay was 11 days (IQR: 9–13), and the median duration of ECMO support was 38.99 hours (IQR: 26.78–51.2). Most patients (86%) were connected to ECMO before PCI, while 15% were connected during PCI. The most common ECMO-related complications were bleeding (8%) and lower extremity ischemia (5%).

The Kaplan–Meier curve for freedom from MACCEs is shown in Fig. 1. The overall freedom from MACCEs at 12 months was 59% (95% confidence interval (CI): 49–71%). In the subgroup with complete revascularization, the freedom from MACCEs at 12 months was 73% (95% CI: 61–87%). Conversely, in the subgroup with incomplete revascularization, the freedom from MACCEs at 12 months was 41% (95% CI: 27–62%). The difference between these two subgroups was statistically significant (log-rank test: p-value = 0.018). The most common MACCE within 12 months was death (30%), followed by cerebrovascular accident (21%), reinfarction (20%), and repeat revascularization (5%). The one-year mortality rate was 30%. The observed in-hospital mortality rate was 24%, with the cardiovascular mortality rate being 18%. Cardiovascular mortality was attributed to various causes, including cardiac arrest (two cases), refractory cardiogenic shock (five cases), ventricular arrhythmia (three cases), and other cardiovascular-related causes (four cases). The overall survival rate after 12 months was recorded at 70%, while the cardiovascular-specific survival rate in the same period was 82%.

Fig. 1.

Kaplan–Meier curve showing MACCE-free survival with a standard error of 10% during or before the 16-month point. MACCE, major adverse cardiac and cerebrovascular event.

Before the ECMO-assisted PCI procedure, 60 patients (78%) were receiving inotropic support, including dobutamine, dopamine, or milrinone. Vasodilators, such as nitrates or sodium nitroprusside, were administered to 40 patients (53%). Other vasoactive medications, such as norepinephrine or vasopressin, were used in 28 patients (37%) to maintain adequate perfusion pressure. At the 12-month follow-up, among the 53 surviving patients, 50 patients (94%) were on aspirin, 53 patients (100%) were on a P2Y12 inhibitor (clopidogrel), 42 patients (79%) were on beta-blockers, 49 patients (92%) were on an angiotensin-converting enzyme inhibitor (ACEI) or angiotensin receptor blocker (ARB), 53 patients (100%) were on a statin, and 53 patients (100%) were on a mineralocorticoid receptor antagonist (MRA). These medications were prescribed in accordance with the current guidelines for the management of heart failure and coronary artery disease.

The changes in LV functional remodeling from baseline to 12 months after the procedure are shown in Table 3 and Fig. 2. A GLMM was used to account for repeated measures within patients and to model the longitudinal changes in these variables over time. There was a significant improvement in LVEF, WMSI, EDVI, and ESVI over time (p 0.001 for all) after adjusting for clinical indication, baseline values of LVEF, WMSI, EDVI, and ESVI, as well as multiple vessel disease, collateral circulation, previous myocardial infarction, postoperative re-infarction, and complete revascularization. The results showed that LVEF increased significantly after the procedure from baseline to 6 months, with a weekly change rate of 0.0007/week (95% CI: 0.0005 to 0.0010/week; p 0.001). However, LVEF decreased slightly at 1 year, although it was still higher than the baseline value. WMSI, EDVI, and ESVI decreased significantly from baseline to 1 year, indicating improved LV function and reduced LV size. The weekly change rates for WMSI, EDVI, and ESVI were –0.005/week (95% CI: –0.006 to 0.004/week), –0.085 mL/m²/week (95% CI: –0.098 to 0.072 mL/m²/week), and –0.042 mL/m²/week (95% CI: –0.063 to 0.021 mL/m²/week), respectively (p 0.001 for all). These results suggest that ECMO-assisted PCI positively impacted LV functional remodeling in patients with ICM.

Fig. 2.

Changes in LVEF (A), WMSI (B), EDVI (C), and ESVI (D) in the patients at baseline and follow-up. LVEF, left ventricular ejection fraction; WMSI, wall motion score index; EDVI, end-diastolic volume index; ESVI, end-systolic volume index.

The GLMM analysis did not reveal a significant difference between survivors and non-survivors in LVEF, WMSI, EDVI, and ESVI improvement trends over time. The adjusted $\beta$ for the group effect was –0.001 (95% CI: –0.015 to 0.012, p = 0.884) for LVEF; 0.036 (95% CI: –0.040 to 0.111, p = 0.382) for WMSI; 0.410 (95% CI: –0.663 to 1.494, p = 0.484) for EDVI; –0.179 (95% CI: –1.393 to 0.999, p = 0.782) for ESVI. Furthermore, the analysis did not demonstrate a significant difference in LVEF, WMSI, EDVI, and ESVI improvement between the subgroups with incomplete and complete revascularizations. The adjusted $\beta$ for the group effect was –0.002 (95% CI: –0.013 to 0.010, p = 0.788) for LVEF; 0.045 (95% CI: –0.018 to 0.108, p = 0.191) for WMSI; –0.171 (95% CI: –1.057 to 0.712, p = 0.721) for EDVI; 0.406 (95% CI: –0.573 to 1.390, p = 0.445) for ESVI. These findings are illustrated in Figs. 3,4.

Fig. 3.

Comparison of changes in LVEF (A), WMSI (B), EDVI (C), and ESVI (D) between survivor and non-survivor patient groups at baseline and follow-up. LVEF, left ventricular ejection fraction; WMSI, wall motion score index; EDVI, end-diastolic volume index; ESVI, end-systolic volume index.

Fig. 4.

Comparison of changes in LVEF (A), WMSI (B), EDVI (C), and ESVI (D) between patients with incomplete and complete revascularization at baseline and follow-up. LVEF, left ventricular ejection fraction; WMSI, wall motion score index; EDVI, end-diastolic volume index; ESVI, end-systolic volume index.