What’s new in adult extracorporeal membrane oxygenation in 2022–2023? Insights from the EuroELSO congress 2023

Research in the field of adult extracorporeal membrane oxygenation (ECMO) demonstrated remarkable productivity in recent years. Indeed, from January 2022 to March 2023, there were more than 1675 publications with the Mesh term “ECMO” cited in Pubmed. This concise review, presented at the recent Euro Extracorporeal Life Support Organization (ELSO) meeting, aims to offer a condensed overview of the main investigations pertaining to adult ECMO published during that period. This personal selection (obviously not exhaustive) was primarily driven by studies with fresh research insights in the fields of ECMO for adult respiratory failure, cardiogenic shock, and refractory cardiac arrest.

VV-ECMO for severe ARDS

At the time of the Euro ELSO meeting in Lisbon (April, 26–29, 2023), most countries had returned to daily life as it was before the pandemic. For the most severely critically ill patients, analyses conducted on single-center and multicenter international cohorts^1,2 revealed that despite longer ECMO durations and extended stays in intensive care unit (ICU), the mortality rate among COVID-19 patients who received ECMO during the first wave of the pandemic exhibited similar outcomes reported in the EOLIA trial ³ and other retrospective series involving ECMO for non-COVID ARDS. Nonetheless, less favorable outcomes were documented in patients treated after July 2020, with potential explanations including specific SARS-CoV-2 strains, the experience of ECMO centers, patient characteristics, and ventilation management protocols. The ECMO-SURGES study, a retrospective multicenter cohort study, aimed to assess the characteristics and mortality of COVID-19 patients who underwent ECMO across 21 experienced ECMO centers in 8 European countries from January 2020 to September 2021. ⁴ Throughout this period, a total of 1345 ECMO patients were managed in these centers. Among them, the wild-type variant accounted for 51% of cases, while the Alpha, Delta, and other variants contributed 29%, 14%, and 6% of cases, respectively. Patients with different variants demonstrated comparable demographic characteristics, except for Delta variant patients who tended to be younger with fewer comorbidities. Overall, the time interval between ICU admission and intubation, the percentage of patients receiving non-invasive oxygenation strategies, and the incidence of pneumothorax showed an upward trend over time. The overall 90-day mortality rate was 42%, with specific rates of 43%, 39%, 40%, and 58% for patients with wild-type, Alpha, Delta, and other variants, respectively (p = .008). After statistical adjustment, infection with the Delta variant was independently associated with higher mortality, with the wild-type strain serving as the reference. Moreover, the time from ICU admission to intubation, reflecting the duration of non-invasive respiratory support, emerged as a stronger predictor of adverse outcomes compared to the duration of mechanical ventilation prior to ECMO initiation. Patients treated with non-invasive oxygenation strategies might have experienced more pronounced self-inflicted lung injuries, potentially due to strong respiratory efforts and significant swings in transpulmonary pressure. These factors could also contribute to the elevated pneumothorax rates observed in the later three semesters of the study. Consequently, this study highlighted the need to consider the duration of non-invasive respiratory support when establishing selection criteria for ECMO in the context of COVID-19-related ARDS. Lastly, the observed 42% 90-days mortality rate was comparatively lower than in outcomes reported in other cohorts.^2,5 This observation emphasized the need of concentrating ECMO resources in experienced ECMO centers, particularly during pandemics that place considerable pressure on the healthcare system.

The post-ICU discharge and long-term outcomes of severe COVID-19 patients on ECMO remained uncertain. Given the substantial number of survivors of critical illness post-COVID-19, it became crucial to assess the range of morbidity, especially among those severely impacted and requiring ECMO support. The EuroECMO-COVID study, an observational prospective multicenter initiative by Euro-ELSO, sought to assess the condition of patients 6 months after undergoing ECMO for COVID-19. ⁶ The study included 1215 patients (with a median age of 53 years [interquartile range (IQR) 46–60]). The median ECMO duration was 15 days (IQR 8–27), in-hospital mortality was 50%, and 74% of patients encountered at least one complication on ECMO. Among the 577 patients for whom data were available, 547 (95%) were alive at 6-month. Only 24% of in-hospital survivors had returned to full-time employment, while 13% had taken up part-time work after 6 months. Additionally, 17% of patients necessitated respiratory rehabilitation. Common residual symptoms at 6 months included dyspnea (35%) and cardiac (10%) or neurocognitive (13%) symptoms. Chommeloux et al. also made a noteworthy contribution by providing granular data on 1-year multidimensional outcomes for survivors of COVID-19-related ARDS treated with ECMO. ⁷ This prospective multicenter study highlighted that the 62 survivors of COVID-19 ECMO were relatively young (with a mean age of 47 years) and had limited comorbidities. A majority had been employed full-time before contracting COVID-19. Analogous to the literature regarding non-COVID-19 ARDS outcomes, recovery of normal to nearly normal pulmonary function and thorax computed tomography imaging was observed within 6 months, albeit a residual lower diffusion capacity persisted at 12 months. Pulmonary outcomes contrasted with reported physical and psychological impairments. Indeed, 38% of patients had returned to work 1-year post-ECMO. Anxiety and depressive symptoms were notable (42–44%), and posttraumatic stress disorder (42%) prevalence at 1 year surpassed that of individuals with non-COVID-19 ARDS treated with ECMO. Furthermore, only 31% of patients reported a return to baseline sexual function at 1 year. The attribution of these symptoms to ECMO, COVID-19 itself, or a combination of both still requires clarification through more extensive dedicated studies. Nonetheless, these two pivotal studies underscored the need for comprehensive post-ECMO follow-up programs, which address physical and psychological rehabilitation needs in patients (and potentially their families) who have been supported by ECMO for a prolonged period.

The complex decision-making process surrounding the eligibility of adult patients for ECMO is still challenging for physicians. For instance, severely obese patients may challenge the risk–benefit balance because of anticipated difficulties with cannulation and reaching sufficient blood flow. Therefore hesitancy to consider ECMO in the daily care of these patients persists and severe obesity (BMI > 40 kg/m²) was even stated as a relative contraindication in the recent ELSO guidelines for ECMO in COVID-19. ⁸ Based on an analysis of the ELSO registry involving 18,529 patients, Peetermans and colleagues provided compelling insights into that frequent question. ⁹ They reported that patients classified with obesity class II or higher, who received respiratory ECMO, had lower mortality rates and shorter hospital stays compared to those with a BMI < 35 kg/m². Interestingly, these trends persisted despite an increased occurrence of cardiovascular, device-related, and renal complications. These findings remained consistent across various statistically adjusted analyses. Notably, the association between BMI and outcomes exhibited a nonlinear pattern, and no upper BMI limit indicated futility. This noteworthy study highlighted the persistence of the “obesity paradox” – i.e., the superior outcomes seen in obese critically ill patients compared to their non-obese counterparts - within the context of ECMO for respiratory failure. Consequently, a high BMI should not refrain from using ECMO in this population.

Similarly, patients with hematological malignancies, particularly those undergoing hematopoietic cell transplantation (HCT), often encounter contraindications to ECMO. Addressing this challenge, an international panel of experts published a consensus statement in the Lancet Respiratory Medicine to offer guidance to intensivists and hematologists struggling with the ECMO candidacy decision-making process in this specific population. ¹⁰ The key points outlined were: (1) HCT should no longer be an absolute contraindication for ECMO; (2) The evaluation of ECMO eligibility should involve a collaborative effort encompassing an interprofessional team, which includes the patient or their surrogate decision-makers; and (3) ECMO should be proposed solely for patients receiving HCT due to non-malignant conditions or malignancies in documented remission, and only when the risk of recurrence at the time of ECMO assessment is low.

Beyond the meticulous selection process for VV-ECMO, substantial uncertainties persist concerning various parts of the daily management of VV-ECMO patients. A significant contribution toward refining blood product management in this context came from Martucci et al. ¹¹ Through a multicenter, prospective, international cohort study involving 604 patients on ECMO for respiratory failure, the study found that blood transfusions were highly prevalent, with 83% of patients receiving at least one packed red blood cell (PRBC) unit. Furthermore, PRBC transfusions were administered on 31% of ECMO days, with a mean pretransfusion hemoglobin concentration of 8.1 g/dL. Based on a time-dependent Cox model, an elevated risk of death in the ICU was associated with hemoglobin concentrations below 7 g/dL compared to higher hemoglobin levels. Importantly, PRBC transfusions were consistently associated with reduced mortality only when administered with a hemoglobin concentration lower than 7 g/dL. Conversely, no substantial impact on mortality reduction was observed for transfusions triggered by higher hemoglobin levels. These crucial findings indicate that a hemoglobin threshold of 7 g/dL for PRBC transfusions seems to be safe in that ECMO population. However, further investigations are warranted to confirm these results.

VA ECMO for refractory cardiogenic shock

The first extensive randomized controlled trial focusing on veno-arterial extracorporeal membrane oxygenation (VA ECMO) in cases of cardiogenic shock was published in 2023. ¹² This randomized controlled trial (RCT) aimed to compare immediate VA-ECMO application with early conservative therapy in patients experiencing rapid deterioration (Stage D-E of the Society for Cardiovascular Angiography and Interventions (SCAI) classification) or severe cardiogenic shock (Stage D of the SCAI classification), defined by echocardiographic, hemodynamic, and metabolic criteria. The primary endpoint was a composite measure including death from any cause, resuscitated cardiac arrest, or the implementation of another form of mechanical circulatory support (including VA-ECMO in the conservative arm) within 30 days. The trial spanned 8 years (2014-2022). Among the 117 subjects analyzed, 58 were assigned to immediate VA-ECMO, and 59 received no immediate VA-ECMO. Among these, 45 patients met the criteria for rapidly deteriorating cardiogenic shock, while 72 experienced severe cardiogenic shock. Acute myocardial infarction was the most common cause of cardiogenic shock in both arms. The composite primary endpoint did not significantly differ between the immediate VA-ECMO group (63.8%) and the early conservative group (71.2%). Additionally, all-cause mortality at 30 days showed no substantial difference between the two groups (50.0% vs 47.5%; Hazard Ratio, 1.110 [95% Confidence Interval, 0.660–1.866]). In the early conservative group, 39% of patients required VA-ECMO support, and among them, 52% did not survive. It is important to note that this trial lacked sufficient statistical power to detect a 50% reduction in the primary endpoint within the VA-ECMO group, as hypothesized by the investigators. This anticipated reduction was arguably considered excessive, a point acknowledged by the authors themselves. Furthermore, this trial encompassed patients with cardiogenic shock from various causes. The medical community eagerly awaits larger ongoing studies, which are specifically focused on acute myocardial infarction, including trials like ECLS-SHOCK (comparing VA-ECMO with conservative therapy), ¹³ the ANCHOR trial (evaluating VA-ECMO plus intra-aortic balloon pump vs conservative therapy) (NCT04184635), and the DanGer Shock trial ¹⁴ (comparing Impella with conservative therapy), anticipated in the forthcoming years.

The precise timing of left ventricle unloading remains a central concern for physicians treating patients with VA-ECMO. An evaluation involving 421 patients experiencing cardiogenic shock and treated with VA-ECMO and active left ventricle (LV) unloading across 18 tertiary care centers in four countries aimed to assess the relationship between the timing of active LV unloading through Impella and VA-ECMO implantation and patient outcomes within this group. ¹⁵ Early LV unloading was characterized by Impella implantation within 2 h after VA-ECMO initiation, while delayed LV unloading occurred between 2 and 24 h post-VA-ECMO onset. Nearly three-quarters of the studied population underwent early active LV unloading. Early active LV unloading was significantly associated with a reduced 30-days mortality risk and an increased likelihood of successful mechanical ventilation withdrawal, though not associated with more complications. Importantly, the relative mortality risk increased, and the probability of successful ventilation weaning decreased proportionally with the time interval between VA-ECMO implantation and the (delayed) initiation of active LV unloading. These interesting findings warrant validation through future randomized controlled trials.

ECMO on cardiopulmonary resuscitation in refractory out-of-hospital cardiac arrest

Both the ARREST ¹⁶ and Prague trials ¹⁷ have made significant contributions to evaluating the efficacy of ECMO in refractory OHCA. The ARREST trial stressed the effectiveness of ECMO within a tightly controlled and dedicated environment, adhering to stringent inclusion criteria such as an initial rhythm of ventricular fibrillation, no ROSC after three defibrillation shocks, and an estimated transfer time to the emergency department shorter than 30 min. Similarly, the Prague study demonstrated that under similar circumstances, outcomes of standard advanced cardiac life support could surpass initial if done with advanced logistics. However, the effectiveness of ECPR might face challenges if inclusion criteria are expanded, as seen in cases involving non-shockable rhythms. The latest trial addressing this subject was published in January 2023, known as the INCEPTION trial. ¹⁸ This multicenter, randomized, controlled study carried out in the Netherlands allocated OHCA patients to either receive ECPR or conventional CPR. Eligible participants had received bystander CPR, exhibited an initial ventricular arrhythmia, and had not achieved a return of spontaneous circulation within 15 min of initiating CPR. The primary outcome measured was survival with a favorable neurological outcome, defined as a Cerebral Performance Category score of one or two on Day-30. Among the 160 patients randomized, 70 were assigned to undergo ECPR while 64 received conventional CPR. At 30 days, 14 patients (20%) in the ECPR group exhibited favorable neurological outcomes, compared to 10 patients (16%) in the conventional CPR group (odds ratio: 1.4; 95% confidence interval: 0.5 to 3.5; p = .52). Noticeably, the median time from arrest initiation to ECMO flow onset was 74 min (IQR: 63–87) in the INCEPTION study, whereas it was 58 min (IQR: 43–70) in the Prague trial. Similarly, the median (IQR) cannulation procedure time was notably shorter in the Prague Study compared to the INCEPTION trial (12 (9–15) versus 20 (11–25) minutes). The INCEPTION study underscored the intricacies of implementing ECPR in real-world scenarios, even in highly developed countries, emphasizing that outcomes achieved in specialized centers might not be universally reproducible. Moreover, this study highlighted the need for efficient logistics and experience in performing ECPR for OHCA to ensure optimal resource utilization and the effectiveness of ECPR in saving lives.