In this study, 82 patients from the Dankook University Hospital ECMO registry were analyzed. From July 2008 to March 2016, we analyzed 42 patients with AMI with cardiogenic shock who underwent CPR and who were treated using the ECMO system (Figure 1). The major inclusion criteria were AMI with cardiogenic shock patients who underwent CPR and required ECMO support as well as visited catheterization laboratory room to receive primary PCI. Exclusion criteria include unwitnessed cardiac arrest, definite non-cardiogenic arrest, no return of spontaneous circulation (ROSC) after CPR, irreversible organ failure when ECMO is not expected to be beneficial (hepatic failure, late stage of adult respiratory distress syndrome, etc.), and patients eligible for “do-not-resuscitate” status. This study was approved by the Institutional Review Board in Dankook University Hospital (approval number: 2018-05-004).

Based on the third universal definition of AMI implemented by a joint task force from the European Society of Cardiology (ESC), American College of Cardiology (ACC) Foundation, American Heart Association (AHA), and the World Heart Federation (WHF) in 2012, AMI was diagnosed. We included the type I AMI by clinical classification.13) Cardiogenic shock was defined by signs of tissue hypoperfusion, such as clammy and cool peripheries, prolonged capillary refill times, altered mental status, oliguria, pulmonary congestion, tachycardia, elevated lactate, or mixed venous saturation of <65%. Hemodynamic criteria for cardiogenic shock included a systolic pressure of <90 mmHg for more than 30 minutes, despite maximal medical treatments for the correction of hypovolemia, hypoxemia, and acidosis. CPR included both intermittent and persistent CPR. Total CPR time was defined as the interval between compression initiation and ROSC. CPR time summed the time of all actions related to CPR. CPR-ECMO time was defined as the interval between CPR initiation and ECMO application. Successful revascularization was defined when residual stenosis was <20% with thrombolysis in myocardial infarction (TIMI) III flow in all intervened lesions. The primary outcome was 30-day in-hospital mortality after primary PCI. To determine the predictors of 30-day in-hospital mortality, we collected clinical, angiographic, procedural, and outcome-related data from hospital medical records.

In this study, Capiox Emergency Bypass System (Capiox EBS™; Terumo Inc., Tokyo, Japan) was used as the ECMO system. This system consisted of an extracorporeal life controller including a back-up battery and a disposable bypass circuit integrated with a heparin-coated membrane oxygenator and a centrifugal pump. The device was implanted onto the femoral artery and vein by percutaneous cannulation using the Seldinger technique. The tip of the arterial cannula was positioned in the common iliac artery and the tip of the venous cannula was placed at the junction of the right atrium and the superior vena cava. Anticoagulation was achieved by administering a bolus injection of unfractionated heparin and maintaining an activated clotting time between 180 and 220 seconds with a continuous intravenous infusion of unfractionated heparin. The initial flow rate of ECMO was 2.2 L/min/m², which was subsequently regulated to maintain a mean arterial pressure of 65 mmHg. ECMO weaning was considered when the patient was hemodynamically stable and properly oxygenated and when ECMO flow rate was <1 L/min/m² for 4 hours. Successful weaning was defined as removal of ECMO support without reinsertion or death within 24 hours.

Continuous variables were presented as mean±standard deviation or median with interquartile ranges. Continuous variables were compared using Student's t-test or Mann-Whitney U test. Categorical variables were expressed as percentages and compared using the χ² test or Fisher's exact test. Event-free survival curves were plotted using the Kaplan-Meier method and compared using the log-rank test. The predictive value of CPR time to predict mortality was analyzed using receiver operating characteristic (ROC) curves, in which the best cut-off value corresponded with the value with the greatest accuracy (sensitivity + specificity). Univariate and multivariate Cox proportional hazard regression analysis was conducted to determine the predictors of in-hospital mortality. Covariates that were statistically significant in univariate analysis results and those considered clinically relevant were included in the multivariate models. All statistical analyses were two-tailed, with clinical significance defined at a p value of <0.05. All statistical analyses were conducted using statistical package for social science (SPSS 20.0 for Windows; SPSS Inc., Chicago, IL, USA) software.

The baseline characteristics for comparison between the survivor group (n=8) and the non-survivor group (n=34) are shown in Table 1. There were no significant differences in demographics or co-morbidities between the 2 groups.

The survivor group showed a significantly higher initial pH compared with the non-survivor group. There was no difference in cardiac enzyme levels between the 2 groups. Door-to-balloon time in the survivor group was shorter than that in the non-survivor group (74.63±29.8 vs. 132.09±109.32 minutes, p=0.03, respectively). Most patients had multi-vessel disease, but there was no significant difference between in the number of patients between the 2 groups. The major culprit artery was the left anterior descending artery (LAD). The revascularization rate of only the culprit artery was 64.3%, however, there was no difference in this rate between the 2 groups (Table 2).

CPR time was significantly longer in the non-survivor group than in the survivor group (43.06±38.72 vs. 11.25±12.08 minutes, p=0.028, respectively). Out-of-hospital cardiac arrest rate was higher in the non-survivor group than in the survivor group, but this difference was statistically insignificant. There was no patient who were applied ECMO before cardiac arrest in the survivor group. Only 1 patient (3.2%) was applied ECMO before cardiac arrest in non-survivor group. ECMO application before PCI was not significantly different between the 2 groups (3 [37.5%] vs. 8 [26.7%], p=0.55). During CPR, the presenting cardiac rhythm was asystole in 3 patients (7.1%), pulseless electrical activity (PEA) in 14 patients (33.3%), and ventricular tachycardia or fibrillation in 25 patients (33.3%) (Table 3).

ROC curve analysis was performed to define an optimal cut-off CPR time value to predict in-hospital mortality in patients with AMI and a complication of cardiogenic shock who were treated by ECMO-assisted primary PCI. The area under the curve was 0.83 (95% confidence interval [CI], 0.67–0.99; p=0.004), and the greatest accuracy in the differentiation of mortality outcome was obtained at a cut-off of 12.5 minutes (88.2% sensitivity and 75% specificity; Figure 2). There was no significant difference demographics, laboratory, and procedural findings between CPR ≤12.5 minutes group (n=10) and CPR >12.5 minutes group (n=32) (Supplementary Tables 1, 2, 3).

Thirty-day in-hospital mortality rate was observed for a total of 33 (78.6%) in our study. During the follow-up period (128.9±413.8 days), mortality rate was observed for total of 34 (81%). Seventeen (29.3%) patients could be weaned from ECMO. Multivariate analysis revealed that CPR time of >12.5 minutes was an independent predictor of 30-day in-hospital mortality outcome (adjusted hazard ratio, 4.66; 95% CI, 1.06–20.5; p=0.042; Table 4). Mortality rate was 2.4% in patients (n=1) who received CPR for <5 minutes, and they were 11.9%, 28.6%, and 52.4% in patients who received CPR for <15 (n=5), <20 (n=12), and <30 (n=22) minutes, respectively (Figure 3). Kaplan-Meier curves demonstrated that the mortality rate was significantly higher in patients who received CPR for a longer duration of >12.5 minutes (30-day mortality rate: 30% vs. 93.7%, p=0.001; Figure 4).