Extracorporeal Membrane Oxygenation During the Coronavirus Disease 2019 Pandemic.

Abstract

In the midst of the 2009 influenza A H1N1 pandemic, clinicians turned to extracorporeal membrane oxygenation (ECMO) as a strategy to save lives (1–3). Based on the H1N1 experience, and the ECMO to Rescue Lung Injury in Severe Acute Respiratory Distress Syndrome (EOLIA) trial (4), ECMO capability has increased exponentially over the past decade (5), including the growth of mobile ECMO programs designed to increase access to the technology and provide care in experienced centers (6). A decade after the H1N1 pandemic, in December of 2019, a pneumonia of unknown etiology was identified in Wuhan, Hubei Province, China. A novel coronavirus was discovered to explain the index cases. On March 11, 2020, the World Health Organization declared the outbreak a pandemic. When hospitalized, nearly one in five patients infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) require ICU admission. Many of these patients require mechanical ventilation and mortality for those who do is extremely high. As coronavirus disease 2019 (COVID-19) traversed the globe, hospitals prepared for the surge. To mitigate the anticipated loss of life, the critical care community readied themselves for an unprecedented volume of acute respiratory distress syndrome (ARDS) cases. Clinicians and administrators projected the needs for ICU beds, ventilators, and ECMO machines and supplies (7). In the early days of the pandemic, informed by the experience in China and Italy, difficult decisions were the norm, as we needed to balance the demands of the projected surge with daily requests to mobilize our program to care for patients in the region and to meet internal demands. As a result, restrictive eligibility criteria for ECMO was commonplace. Despite its potential to save lives, the National Institutes of Health COVID-19 treatment guidelines concluded “there are insufficient data to recommend either for or against the routine use of ECMO for patients with COVID-19 and refractory hypoxemia” (8). Specifically, it was acknowledged that “while ECMO may serve as an effective short-term rescue therapy in patients with severe ARDS and refractory hypoxemia, there is no conclusive evidence that ECMO is responsible for better clinical outcomes in patients who received ECMO than in patients who did not receive ECMO” (8). In this issue of Critical Care Medicine, Yang et al (9) describe their experience managing COVID-19 ARDS, including the use of ECMO. Between January 8, 2020, and March 31, 2020, 129 COVID-19 patients were admitted to the ICU across two hospitals in Wuhan, China. Of 129 ICU admissions, 59 (46%) received mechanical ventilation. Of the 59 mechanically ventilated patients, 21 (36%) received ECMO. Although this is a striking proportion, it is notable that these events occurred in the early weeks of the pandemic. The average age of the 38 patients who received mechanical ventilation without ECMO was 70, with a range of 62 to 79. In contrast, the average age for those who received ECMO was 58, with a range of 42 to 67. Of the 21 patients who received ECMO, 12 (57%) received prone position ventilation prior to ECMO, and nine (43%) required vasoactive agents and eight (38%) developed acute kidney injury requiring renal replacement therapy during ECMO. The average duration of ECMO was 218 hours or 9.1 days. The criteria for ECMO initiation included the following: failure to stabilize with lung-protective ventilation and prone position ventilation and Pao2:Fio2 ratio less than 50 for 3 hours or more; Pao2:Fio2 less than 80 for 6 hours; arterial blood gas pH less than 7.25 and Paco2 greater than 60 mm Hg over 6 hours despite respiratory rate (RR) of greater than 35; pH less than 7.2 with RR greater than 35 breaths per minute and plateau pressure greater than 30; or cardiogenic shock or cardiac arrest. During ECMO, patients received IV heparin infusion with goal activated clotting times (ACTs) of 160–200 seconds and activated partial thromboplastin time no greater than two times the upper limit of normal. Although the criteria were not specified, patients deemed a high risk of bleeding had goal ACT of 130–160 seconds. Among those who received ECMO, as of April 7, nine survived (43%), six of whom had been discharged. In contrast, 14 (37%) of the 38 patients who received mechanical ventilation without ECMO survived. Survival, among those who received ECMO, was more common in patients with lower Paco2 (54 vs 63; p = 0.006) and higher pH (7.38 vs 7.23; p = 0.02) prior to ECMO initiation. As hypercapnia is reflective of dead-space ventilation in ARDS, and worsening dead-space fraction is associated with ARDS- (10) and ECMO-related mortality (11), these findings reveal that the degree of lung failure at the time of ECMO initiation predicts death. Among the 12 ECMO decedents, two patients died after cardiac arrest, six patients died as a result of progressive respiratory failure complicated by multidrug-resistant pulmonary infections, three patients died from septic shock related to drug-resistant Acinetobacter bloodstream infections, and one patient died from intracerebral hemorrhage (ICH). Future studies, designed to examine the incidence of nosocomial infections and the potential role of empiric antibiotics in COVID-19, are warranted. Regarding a separate complication, the rate of ICH observed in this case series is significantly lower than our experience during this pandemic and several centers during the H1N1 pandemic (1,2). Given the prolonged ventilator and ECMO support required to support COVID-19 patients with severe ARDS, strategies to mitigate both of these complications are urgently needed. In the EOLIA trial, as a comparison to the outcomes reported herein, 65% of patients enrolled in the ECMO group had survived by 60 days, compared with 54% in the control group. Although not statistically significant, 35 patients received ECMO as rescue therapy in the control group to avoid anticipated death, suggesting that a strategy of ECMO in cases of very severe ARDS was life-saving. Relevant to the current study by Yang et al (9), the EOLIA ECMO criteria were as follows: Pao2:Fio2 of less than 50 mm Hg for 3 hours or more, Pao2:Fio2 of less than 80 mm Hg for 6 hours or more, or an arterial pH of less than 7.25 with a Paco2 of 60 mm Hg or more for 6 hours (4). For those who received ECMO, the average duration of ECMO was 15 days, with a sd of 13 days. Despite similar ECMO eligibility criterion, the outcomes observed in the current study by Yang et al (9) juxtaposed to EOLIA and the H1N1 case series (1–3) suggest an apparent incontrovertible truth: the case-fatality rate observed among COVID-19 ARDS cases is unprecedented. Although true, as an international critical care community, we have also adapted and learned in short order, and anticipate that we will continue to do so as our experience caring for this novel disease matures. As this case series reports the results at the onset of the pandemic, we anticipate that outcomes will continue to improve as the role of ECMO becomes more defined and established. There are several limitations that warrant discussion. First, as a small, nonrandomized study, it is unclear whether the use of ECMO improved outcomes. Second, given the high rate of ECMO use and relatively low use of prone position ventilation pre-ECMO, it is plausible that timely use of prone position ventilation, recruitment maneuvers, and/or neuromuscular blockade could have alleviated the use of ECMO in some instances. Our experience, and the recent report from the Massachusetts General Hospital, wherein 5% of ventilated COVID-19 patients required ECMO (12), supports the notion that ECMO use is now relatively uncommon at this stage of the COVID-19 pandemic. Third, given the limited sample size, the authors were unable to adjust for confounders. Multicenter studies, designed to examine multiple risk factors simultaneously, are needed. Fourth, although the two centers are described as ECMO referral hospitals, it is unclear whether one or both centers were high volume ECMO centers, an important consideration given the relationship between volume and ECMO outcomes (13). Last, while the clinical scenario was reported in those who died while on ECMO, autopsy was not performed and/or reported. Nevertheless, based on published COVID-19 autopsy reports (14,15), lung failure, characterized by diffuse alveolar damage, and venous thromboembolism, likely contributed herein. In conclusion, the study by Yang et al (9) is an important early contribution to our understanding of the role of ECMO in the COVID-19 pandemic. In time, with a greater understanding of the course of the disease and effective management strategies, a goal is to minimize the need for ECMO. We appear to be making progress already on this front. At the same time, as was the case in the 2009 H1N1 pandemic and as demonstrated by Yang et al (9) herein, ECMO has an important role as a rescue therapy for patients with SARS-CoV-2 pneumonia associated with severe ARDS.