Improvement in the survival rates of extracorporeal membrane oxygenation-supported respiratory failure patients: a multicenter retrospective study in Korean patients

Background Although the utilization of extracorporeal membrane oxygenation (ECMO) is increasing and its technology is evolving, only a few epidemiologic reports have described the uses and outcomes of ECMO. The aim of this study was to investigate the changes in utilization and survival rate in patients supported with ECMO for severe respiratory failure in Korea. Methods This was a multicenter study on consecutive patients who underwent ECMO across 16 hospitals in Korea. The records of all patients who required ECMO for acute respiratory failure between 2012 and 2015 were retrospectively reviewed, and the utilization of ECMO was analyzed over time. Results During the study period, 5552 patients received ECMO in Korea as a whole, and a total of 2472 patients received ECMO at the participating 16 hospitals. We analyzed 487 (19.7%) patients who received ECMO for respiratory failure. The number of ECMO procedures provided for respiratory failure increased from 104 to 153 during the study period. The in-hospital survival rate increased from 30.8% to 35.9%. The use of prone positioning increased from 6.8% to 49.0% (p < 0.001), and the use of neuromuscular blockers also increased from 28.2% to 58.2% (p < 0.001). Multiple regression analysis showed that old age (OR 1.038 (95% CI 1.022, 1.054)), use of corticosteroid (OR 2.251 (95% CI 1.153, 4.397)), continuous renal replacement therapy (OR 2.196 (95% CI 1.135, 4.247)), driving pressure (OR 1.072 (95% CI 1.031, 1.114)), and prolonged ECMO duration (OR 1.020 (95% CI 1.003, 1.038)) were associated with increased odds of mortality. Conclusions Utilization of ECMO and survival rates of patients who received ECMO for respiratory failure increased over time in Korea. The use of pre-ECMO prone positioning and neuromuscular blockers also increased during the same period.


Background
Extracorporeal membrane oxygenation (ECMO), which provides respiratory and/or cardiac support, allows treatment of patients with refractory gas-exchange abnormalities [1]. The use of ECMO to support patients with respiratory failure is increasing worldwide following the use of ECMO for severe acute respiratory failure during the 2009 influenza A pandemic [2][3][4][5]. Recently, the EOLIA trial reported that in patients with severe acute respiratory distress syndrome (ARDS) there was no significant difference in 60-day mortality between patients who received early ECMO and those who received conventional mechanical ventilation that included ECMO as rescue therapy [6]. However, crossover to ECMO occurred in 28% of patients in the conventional group, who showed a high mortality rate of 57%. This suggests that ECMO can be used in severe ARDS patients who do not benefit from conventional treatment.
Survival of patients who received ECMO is also gradually increasing over time [7]. A recent epidemiologic report in Germany showed that ECMO utilization for severe respiratory failure significantly increased from 2007 until 2012, and in-hospital survival increased over time as well [8]. Sauer et al. [9] reported that the annual rates of ECMO cases increased by 433% from 2006 to 2011 in the United States, and that, albeit not statistically significant, there was an improving trend in the survival rate as well. In a single-center study in Korea, the survival rates associated with the ECMO procedure increased between 2009 and 2011 [10]. However, as we have previously reported, there was a discrepancy in the survival rate between those of the Extracorporeal Life Support Organization (ELSO) registry and Korean ECMO patients [11]. The in-hospital survival rate of ECMOtreated patients with acute respiratory failure was 46% from 2014 to 2015 in Korea, whereas the survival rate was 58% in the ELSO registry patients [7]. Also, we have suggested that age is an important factor in the survival of patients who received ECMO. Therefore, we sought to determine whether there has been an improvement in the survival rate of patients who received ECMO support for acute respiratory failure in Korea. Specifically, we evaluated the changes over time in the survival rates of patients supported with ECMO for severe respiratory failure and the factors associated with the survival rate.

Study design
This was a multicenter study of consecutive patients who received ECMO at 16 hospitals in Korea. The records of all patients who required ECMO for acute respiratory failure between 2012 and 2015 were retrospectively reviewed and the utilization of ECMO was analyzed over time. The decision to use ECMO was made at the discretion of the attending physicians at each center without standardization. The study protocol was approved by the institutional review board of Asan Medical Center, and by the local institutional review boards of all other participating centers. The requirement for informed consent was waived due to the retrospective design of the study.

Data collection
Data were collected from electronic medical records of patients older than 19 years who received ECMO support. Included variables were as follows: demographic information, Acute Physiology and Chronic Health Evaluation (APACHE) II and Sequential Organ Failure Assessment (SOFA) scores at intensive care unit (ICU) admission, etiology of respiratory failure, cardiac arrest, immunocompromised status, central nervous system (CNS) dysfunction, pre-ECMO hemodynamic data, mechanical ventilation parameters, and arterial blood gas data. Immunocompromised status and CNS dysfunction were defined according to the RESP study [12]. Immunocompromised status included hematological malignancies, solid tumors, solid-organ transplantation, high-dose or long-term corticosteroid and/or immunosuppressant use, and human immunodeficiency virus infection. CNS dysfunction included diagnoses of neurotrauma, stroke, encephalopathy, cerebral embolism, seizure, and epileptic syndrome. We collected information on adjunctive therapy such as the use of vasopressors, steroids, continuous renal replacement therapy (CRRT), prone positioning, nitric oxide, bicarbonate infusion, and neuromuscular blockers. We also collected data such as the ECMO mode, ECMO duration, duration of mechanical ventilation to ECMO initiation, hospital stay, and tracheotomy. The ECMO mode was categorized as veno-venous, veno-arterial, and veno-arteriovenous. Outcome variables of the study were survival at discharge and ECMO weaning (survival within 48 h after weaning from ECMO).

Statistical analysis
Demographics, pre-ECMO parameters, and outcomes were compared between 2012 and 2015. Differences with p < 0.05 were considered statistically significant. Categorical variables are expressed as the number (percentage). Continuous variables are expressed as the median (interquartile range). Pearson's chi-square test or Fisher's exact test was used to compare categorical data. The Kruskal-Wallis test was used to compare medians between groups.
Multiple logistic regression analysis using the backward elimination method was performed to identify the factors associated with survival at discharge. Candidate variables for inclusion in the multiple logistic regression model were chosen from the univariate analysis; variables with p < 0.1 in the univariate analyses were included in the multivariate analysis, and collinearity was assessed before the multivariate analysis. Calibrations of the models were evaluated with the Hosmer-Lemeshow goodness-of-fit test. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 22.0 (IBM Corporation, Armonk, NY, USA).

Baseline characteristics of the study population
During the study period (2012-2015), 5552 patients received ECMO support in Korea. ECMO support was given to 2472 patients in the participating 16 hospitals. We analyzed 487 (19.7%) patients who received ECMO specifically for respiratory failure. The annual number of ECMO cases at 16 institutions varied widely: eight centers had fewer than 20 cases per year and the other eight  Survival Prediction, ARF acute respiratory failure, ARDS acute respiratory distress syndrome, COPD chronic obstructive pulmonary disease, ILD interstitial lung disease, CNS central nervous system, CRRT continuous renal replacement therapy, MAP mean arterial pressure, PaO 2 partial pressure of arterial oxygen, PaCO 2 partial pressure of arterial carbon dioxide, HCO 3 − bicarbonate, SaO 2 oxygen saturation, FiO 2 fraction of inspired oxygen, PEEP positive endexpiratory pressure, PIP peak inspiratory pressure, MV mechanical ventilation a "Immunocompromised" included hematological malignancies, solid tumors, solid-organ transplantation, high-dose or long-term corticosteroid and/or immunosuppressant use, and human immunodeficiency virus infection b "CNS dysfunction" included diagnoses of neurotrauma, stroke, encephalopathy, cerebral embolism, seizure, and epileptic syndrome centers had more than 30 cases per year, with two of those centers having had more than 120 cases per year.
The patients' median age was 58 years (range 45-66 years), and the median body mass index was 22.2 kg/m 2 (range 20.6-23.2 kg/m 2 ). Pre-ECMO mechanical ventilation was provided in 92.2% of patients and corticosteroid therapy was used in 16.8% of patients. Prone positioning was applied in 29.5% of patients and neuromuscular blockers were used in 45.4% of patients. The majority of patients were initially supported with veno-venous ECMO (88.1%), and the median duration of support was 8 days (interquartile range (IQR) 4, 18 days). Survival and weaning rates were 38.8% and 57.1%, respectively (Table 1).

Demographics, pre-ECMO parameters, and outcomes over time
The number of ECMO procedures for respiratory failure increased from 104 to 153 during the study period ( Fig. 1). There were no significant differences in age, sex, APACHE II score, SOFA score, immunocompromised status, CNS dysfunction, cardiac arrest, CRRT, use of nitric oxide and bicarbonate infusion, PaO 2 /FiO 2 ratio, ECMO duration, and duration of mechanical ventilation to ECMO initiation between groups. Use of prone positioning increased from 6.8% to 49.0% (p < 0.001) and the use of neuromuscular blockers also increased from 28.2% to 58.2% (p < 0.001; Table 2). Although the survival rate remained relatively low, it increased over time from 30.8% to 35.9% (p = 0.005; Table 3). Post-hoc analysis showed that the survival rate in 2014 was significantly higher than the rates in 2012 and 2015.
The median age was older in the nonsurvivors (61 years; IQR 52, 69 years) than in survivors (51 years; IQR 37, 62 years) (p < 0.001). The survival rate decreased with age, with patients older than 60 years having a survival rate of 30.8% (Fig. 2). ECMO duration was significantly longer in the nonsurvivors (9 days; interquartile range (IQR) 4, 22 days) than in survivors (7 days; IQR 3, 13 days) (p = 0.002). Compared with the survival rate within 2 weeks of ECMO support, the overall survival rate after 2 weeks of ECMO support showed a significant decrease from 43.4% to 27.8% (p = 0.001).

Discussion
This multicenter study was conducted to evaluate the change in survival rates of patients who received ECMO support for acute respiratory failure in Korea. Utilization of ECMO for respiratory failure increased over time, and the survival rate was improved with increasing use of adjunctive management. Also, patient age and the duration of ECMO were significantly associated with survival.
A notable change during the study period was that the administration of neuromuscular blockades and use of prone positioning before ECMO had significantly increased from 28.2% to 58.2% and from 6.8% to 49.0%, respectively. Papazian et al. [13] reported that early use of neuromuscular blockades in patients with severe ARDS may improve survival. In the ELSO registry-based RESP study, neuromuscular blockade agents before ECMO were independently associated with hospital survival [12]. In addition, in patients with severe ARDS, early application of prolonged prone positioning was significantly associated with improved survival [14]. Schmidt et al. [15] demonstrated that use of prone positioning before ECMO was also associated with survival. These results are in accordance with those in a recent systematic review and meta-analysis [16]. Moreover, for patients with severe ARDS, prone positioning before and during ECMO may be helpful for weaning from ECMO [17,18]. Another distinctive finding was the change in pre-ECMO ventilator parameters. In recent years, the driving pressure was lower and minute ventilation was decreased. Therefore, improvement in hospital survival of ECMO-supported patients with respiratory failure might be the result of increasing experience with ECMO over time, including evolving adjuvant therapies and improved management of mechanical ventilation.
The results of this study showed that the number of ECMOs carried out for respiratory failure increased from 104 to 153 from 2012 to 2015, and that the in-hospital survival rate increased from 30.8% to 35.9% during the same period. The overall survival rate of 39% in "CNS dysfunction" included diagnoses of neurotrauma, stroke, encephalopathy, cerebral embolism, seizure, and epileptic syndrome ECMO-supported respiratory failure patients in Korea is lower than the reported rate of 58% in the ELSO registry [7]. Meanwhile, an ECMO epidemiologic study performed in Germany reported that from 2012 to 2014 the in-hospital survival had steadily increased and the rate of survival was approximately 40%, which is similar to our findings [8]. In addition, Sauer et al. [9] reported that in the United States the survival rate of the patients who received ECMO was approximately 40%. In the German study, approximately 80% of patients were older than 40 years and increasing numbers of older patients had received ECMO. In the US study, the mean age of the patients who received ECMO was 50 years, which is higher than that of the patients included in the ELSO registry. Taken together, the discrepancies in demographics between the patients of ECMO centers not included in the ELSO and those in the ELSO registry may explain the difference in survival rates. Also, another explanation for the relatively low survival rate of Korean ECMO patients could be the infrequent use of prone positioning. The use of prone positioning and use of neuromuscular blockers were low compared with those in the EOLIA trial [6], in which prone positioning was applied in 90% of patients in the conventional ventilator support group, who showed a 54% survival rate. The relatively low survival rate in Korean ECMO patients may be due to excessive use of ECMO in patients who may have shown good response to  prone positioning. Accordingly, the use of prone positioning is gradually increasing in Korea.
Another interesting finding of our study was that the survival rate was associated with the ECMO duration. The survival rate of patients who required prolonged ECMO (longer than 14 days) was significantly lower than that of patients who had shorter ECMO duration (28% vs 43%, respectively, p = 0.001). Recently, Posluszny et al. [19] reported that ECMO duration was inversely correlated with the survival rate in ECMO-supported patients with respiratory failure; the survival rate in patients who had longer ECMO duration was 10% lower than that in those with shorter ECMO duration. Nonetheless, the investigators suggested that prolonged ECMO was not futile because there was a significant improvement in survival from 37% to 49% in recent years. On the other hand, the aforementioned German epidemiologic study reported that prolonged ECMO was associated with poorer outcome; that the survival rate rapidly declined to 20% within 10 days after ECMO initiation [8]. Therefore, further studies are needed to provide a more solid association between ECMO duration and the survival rate.
Our study has several limitations. This study was retrospective and had a relatively short study period. Because not all patients treated with ECMO for respiratory failure in Korea were included, selection bias is possible. In addition, long-term outcomes and quality of life could not be assessed, which warrants an extended observation period of our study populations or further epidemiologic studies. Despite such limitations, our current multicenter study, which is not based on the ELSO registry, provides information on the change in the survival rate of ECMO patients with respiratory failure and the factors associated with survival, and adds to the understanding of survival in patients who receive ECMO due to respiratory failure.

Conclusions
This multicenter study performed in Korea showed that utilization of ECMO for respiratory failure had increased over time, and that the survival rates of ECMO-supported respiratory failure patients had improved with increasing utilization of adjunctive management. Patient age and duration of ECMO were significantly associated with survival at discharge.