Clinical Characteristics, Complications and Outcomes of Patients with Severe Acute Respiratory Distress Syndrome Related to COVID-19 or Influenza Requiring Extracorporeal Membrane Oxygenation—A Retrospective Cohort Study

Extracorporeal membrane oxygenation (ECMO) represents a viable therapy option for patients with refractory acute respiratory distress syndrome (ARDS). Currently, veno-venous (vv) ECMO is frequently used in patients suffering from coronavirus disease 2019 (COVID-19). VV-ECMO was also frequently utilised during the influenza pandemic and experience with this complex and invasive treatment has increased worldwide since. However, data on comparison of clinical characteristics and outcome of patients with COVID-19 and influenza-related severe ARDS treated with vv-ECMO are scarce. This is a retrospective analysis of all consecutive patients treated with vv/(veno-arterial)va-ECMO between January 2009 and January 2021 at the University Medical Centre Hamburg-Eppendorf in Germany. All patients with confirmed COVID-19 or influenza were included. Patient characteristics, parameters related to ICU and vv/va-ECMO as well as clinical outcomes were compared. Mortality was assessed up to 90 days after vv/va-ECMO initiation. Overall, 113 patients were included, 52 (46%) with COVID-19 and 61 (54%) with influenza-related ARDS. Median age of patients with COVID-19 and influenza was 58 (IQR 53–64) and 52 (39–58) years (p < 0.001), 35% and 31% (p = 0.695) were female, respectively. Charlson Comorbidity Index was 3 (1–5) and 2 (0–5) points in the two groups (p = 0.309). Median SAPS II score pre-ECMO was 27 (24–36) vs. 32 (28–41) points (p = 0.009), and SOFA score was 13 (11–14) vs. 12 (8–15) points (p = 0.853), respectively. Median P/F ratio pre-ECMO was 64 (46–78) and 73 (56–104) (p = 0.089); pH was 7.20 (7.16–7.29) and 7.26 (7.18–7.33) (p = 0.166). Median days on vv/va-ECMO were 17 (7–27) and 11 (7–20) (p = 0.295), respectively. Seventy-one percent and sixty-nine percent had renal replacement therapy (p = 0.790). Ninety-four percent of patients with COVID-19 and seventy-seven percent with influenza experienced vv/va-ECMO-associated bleeding events (p = 0.004). Thirty-four percent and fifty-five percent were successfully weaned from ECMO (p = 0.025). Ninety-day mortality was 65% and 57% in patients with COVID-19 and influenza, respectively (p = 0.156). Median length of ICU stay was 24 (13–44) and 28 (16–14) days (p = 0.470), respectively. Despite similar disease severity, the use of vv/va-ECMO in ARDS related to COVID-19 and influenza resulted in similar outcomes at 90 days. A significant higher rate of bleeding complications and thrombosis was observed in patients with COVID-19.


Introduction
The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged in 2019 and caused an ongoing worldwide pandemic [1,2]. Although most infected patients have an asymptomatic or mild course of coronavirus disease 2019 (COVID-19) a considerable number of patients required hospitalisation [1,3,4]. About 5% of COVID-19 patients need treatment in the intensive care unit (ICU), mainly for respiratory support for varying degrees of ARDS as well as other forms of organ failure [1,5,6]. These patients have a high risk of mortality, especially when invasive mechanical ventilation (MV) is needed [6][7][8][9][10]. For patients who are experiencing progressive and refractory respiratory failure, veno-venous extracorporeal membrane oxygenation (vv-ECMO) may be considered as rescue therapy [11].
The use of vv-ECMO is an established rescue therapy in patients suffering from severe acute respiratory distress syndrome (ARDS) refractory to conservative management including optimised ventilator settings, prone positioning and pulmonary vasodilatory treatment [12]. Early initiation of vv-ECMO as well as early referral to ECMO centres has been proven to be beneficial in these patients [13][14][15].
Use of vv-ECMO in patients with ARDS related to viral infections was previously reported during the influenza A (H1N1) pandemic as well as the Middle East respiratory syndrome coronavirus (MERS-CoV) outbreaks [16,17]. Clinical features regarding clinical symptoms and course of disease of COVID-19 and influenza are highly variable, and recent studies described differences regarding patient characteristics and outcomes [18,19]. For the comparison of patients with severe ARDS related to COVID-19 and influenza requiring ECMO there is limited data regarding clinical characteristics, complications and outcomes [20][21][22][23].
In the present study, we aimed to investigate and compare clinical characteristics, outcomes and complications of patients with COVID-19 and influenza receiving ECMO for refractory severe ARDS in an experienced high-volume centre.

Study Population, Design and Ethics
We performed a retrospective analysis of all consecutive patients with COVID-19 or influenza admitted to the ICUs of the Department of Intensive Care Medicine at the University Medical Centre Hamburg-Eppendorf (Hamburg, Germany) between 1 January 2009 and 15 January 2021. The Department of Intensive Care Medicine cares for all critically ill adult patients of the university hospital and comprises 12 ICUs, with a total capacity of 140 beds. During the pandemic, a maximum of 3 ICUs with a total capacity of 36 beds was exclusively dedicated to the treatment of patients with COVID-19. The study was approved by the local clinical institutional review board and complies with the Declaration of Helsinki. The study was registered with the Ethics Committee of the Hamburg Chamber of Physicians (No.: WF-052/21). Owing to the retrospective character of the study and its pseudonymised data collection, the need for informed consent was waived.

Inclusion and Exclusion Criteria
We included all consecutive adult patients (≥18 years) with confirmed COVID-19 or influenza requiring vv-or veno-arterial (va)-ECMO support admitted to our department during the study period. Confirmed COVID-19 and influenza were defined as at least one positive result on reverse transcriptase polymerase chain reaction obtained from nasopharyngeal swabs and/or bronchial secretions. Patients with a non-completed ICU stay (ongoing ICU treatment at the end of the study period), without confirmed COVID-19/influenza or aged <18 years were excluded.

Data Collection
Data were collected from the department's electronic patient data management system (PDMS, Integrated Care Manager ® (ICM), Version 9.1-Draeger Medical, Luebeck, Germany) and the department's ECMO database. The extracted data included age, gender, comorbidities, admission diagnosis, length of ICU stay, treatment modalities and organ support (mechanical ventilation, prone positioning, type of ECMO (vv/va), vasopressor support and renal replacement therapy), blood products, medication such as glucocorticoid and anti-infective treatment, as well as laboratory test results.

Study Definitions and Patient Management
ARDS was defined according to the Berlin definition, using the PaO 2 /FiO 2 ratio (Horowitz index) as marker for severity [24]. Severity of illness was evaluated by Sequential Organ Failure Assessment (SOFA) [25], Acute Physiology And Chronic Health Evaluation (APACHE II) [26] and Simplified Acute Physiology II (SAPS II) [27] score. Charlson Comorbidity Index (CCI) [28] was calculated for all patients. Clinical patient management was performed according to national and international guidelines, including prone positioning in moderate to severe ARDS and restrictive fluid management following the initial resuscitation period. Vasopressor support was initiated to obtain a mean arterial pressure (MAP) of 65 mmHg using norepinephrine [12,29]. Patients with severe hypoxemic and/or hypercapnic respiratory failure in combination with severe respiratory acidosis refractory to adjunctive therapies received vv-ECMO (CARDIOHELP-System Maquet GmbH, Rastatt, Germany; Novalung, Fresenius Medical Care, Bad Homburg, Germany; Stoeckert Centrifugal Pump Console, LivaNova, Munich, Germany). Criteria for the initiation of vv-ECMO support were based on the guidelines of the Extracorporeal Life Support Organization (ELSO) and national recommendations [29,30]. Prone positioning during ECMO therapy was initiated in patients with persistent severe hypoxemia. Presence of thrombosis was defined as pulmonary embolisms or deep vein thrombosis; the diagnosis was based on clinical suspicion and the subsequent confirmative diagnostic procedures (e.g., ultrasound or computed tomography) according to local standard operation procedures. Anticoagulation was performed using continuous application of unfractionated heparin. The effect of heparin was monitored using the activated clotting time. The targeted activated clotting time was 40 to 50 s in all patients. Patient survival was obtained at ICU discharge, after 28 and after 90 days post ECMO initiation.

Statistical Analysis
Data are presented as absolute numbers and relative frequency or median and with interquartile range (IQR). We performed an exploratory analysis; categorical variables were compared via chi-square analysis or Fisher's exact test, as appropriate. Continuous variables were compared via Mann-Whitney U test. Survival function estimates were calculated using the Kaplan-Meier method and were compared by log rank test.
Statistical analysis was conducted using IBM SPSS Statistics Version 24.0 (IBM Corp., Armonk, NY, USA). A p-value of <0.05 was considered to be statistically significant. The study protocol was prepared in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology recommendations.

Study Population
Overall, 113 patients were included who suffered from severe ARDS due to influenza pneumonia or COVID-19 and were treated with vv/va-ECMO during the study period. We could identify 61 (54%) patients with influenza-related disease (2009-2021) and 52 (46%) with COVID-19 (2020-2021). The study flow chart is shown in Figure 1. Overall, 113 patients were included who suffered from severe ARDS due to influenza pneumonia or COVID-19 and were treated with vv/va-ECMO during the study period. We could identify 61 (54%) patients with influenza-related disease (2009-2021) and 52 (46%) with COVID-19 (2020-2021). The study flow chart is shown in Figure 1.

Discussion
In the present study we investigated the clinical characteristics and outcomes of critically ill patients with severe ARDS due to COVID-19 or influenza pneumonia on vv-/va-ECMO. To our knowledge this is the most comprehensive study comparing these two groups with severe viral pneumonia requiring vv/va-ECMO. Patients with COVID-19 experienced significantly higher rates of complications including thromboembolic events and bleeding on vv/va-ECMO. However, despite differences of baseline demographics and severity of illness, we found no significant differences in mortality up to 90 days after ECMO initiation.

Discussion
In the present study we investigated the clinical characteristics and outcomes of critically ill patients with severe ARDS due to COVID-19 or influenza pneumonia on vv-/va-ECMO. To our knowledge this is the most comprehensive study comparing these two groups with severe viral pneumonia requiring vv/va-ECMO. Patients with COVID-19 experienced significantly higher rates of complications including thromboembolic events and bleeding on vv/va-ECMO. However, despite differences of baseline demographics and severity of illness, we found no significant differences in mortality up to 90 days after ECMO initiation.
Previous studies reported a higher rate of bleeding events in patients with COVID-19 on vv/va-ECMO compared to influenza [21,23]. In our cohort we observed more bleeding complications in COVID-19 than in influenza patients. This may be due to a more aggressive anticoagulation protocol owing to the higher rate of venous thromboembolism (VTE) reported in our cohort. Furthermore, many previous studies on patients with COVID-19 described a high rate of VTE especially in critically ill patients [31]. The pathophysiology linked to these observations is possibly linked to the COVID-19-associated substantial dysregulation of both inflammation and coagulation [31][32][33]. Previous studies have reported that biomarkers reflecting high systemic inflammation and coagulation activation are associated with worse outcomes [32]. In our cohort, patients with COVID-19 showed higher inflammatory activation during the whole course of disease than influenza patients, potentially reflecting higher severity of illness and risk of VTE. Surprisingly, D-dimer levels were lower than those in patients suffering from influenza. The contrasting higher rates of VTE in patients with COVID-19 may have been due to more severe endothelial injury [34]. However, there were no surrogate parameters of endothelial injury available in our retrospective analysis to further evaluate this hypothesis. To date, the optimal anticoagulation strategy in critically ill patients with COVID-19 on vv/va-ECMO is much debated, but remains unknown and warrants further investigation regarding both optimal coagulation testing as well as optimal and individualised dosing and timing of anticoagulation.
Another striking finding was that 17% of patients in our cohort suffered from cardiac arrest (CA) during their ICU stay. In general, CA occurring within the ICU is less frequent and usually affects only 2% of the general critically ill population [35,36]. Although not significantly different, a higher proportion of patients with COVID-19 suffered from CA in our cohort. This is in line with different recent reports of patients with COVID-19 [35,37,38]. In a previous study, we reported an incidence of 18% for cardiac arrest in critically ill COVID-19 patients with severe ARDS; about one-third of patients were on ECMO therapy [35]. In our cohort, we observed that most patients suffered CA prior to initiation of vv/va-ECMO but also a relevant number of patients suffered CA during therapy. Cause of CA during ECMO therapy was mainly hypoxia related (3 of 7), vagal related (2 of 7), pericardial tamponade (1 of 7) and myocardial infarction related (1 of 7). This further highlights the severity of illness in this vulnerable patient cohort. The occurrence of CA in patients with ARDS requiring vv/va-ECMO is unknown and its occurrence should be further investigated in future studies.
COVID-19 and seasonal influenza pneumonia are both viral respiratory infections with highly variable clinical presentation and a course ranging from asymptomatic cases to respiratory failure with varying degrees of ARDS [18,19,39]. Several studies investigated clinical differences of hospitalised patients with influenza and COVID-19. Most studies reported that patients with influenza were significantly younger and presented with a higher burden of comorbidities [18,19]. In contrast to these findings, we observed that patients with COVID-19 in our cohort presented with a higher number of comorbidities on admission compared to patients with influenza. This may be explained by the fact that patients with multiple comorbidities may be at higher risk for a more severe course of COVID-19 disease leading to both initiation of vv-ECMO and high mortality [1,3]. Furthermore, a more liberal use of vv-ECMO in recent years due to growing experience in our centre may have contributed to applying vv-ECMO in sicker patients [40]. However, high mortality rates were also reported in patients requiring ECMO during MERS-CoV outbreaks [16,41]. Furthermore, changes of clinical practice over the time course of the COVID-19 pandemic due to expanding clinical experience and evidence, especially with respect to anticoagulation, sedation and mechanical ventilation, may also have had an impact on outcome differences between the two study groups.
Tang et al. compared ARDS patients with influenza and COVID-19 and found that patients with COVID-19 had lower severity of illness on admission and lower SOFA scoreadjusted mortality [39]. We can confirm the finding of lower severity of illness in patients with COVID-19 on admission. However, on the day of vv/va-ECMO initiation, severity of illness was comparable between both groups. Possibly a more rapid clinical deterioration of patients with COVID-19 after their initially more liberal ICU admission may explain the observed differences between the two groups regarding severity of illness on admission in our cohort.
Several studies investigated differences between COVID-19 and influenza in hospitalised patients. However, data comparing the clinical characteristics and outcome of these two groups with ARDS on vv-ECMO are scarce [20][21][22]. Reported mortality rates were mainly higher in patients with COVID-19 [20][21][22]. Clinical characteristics and differences between influenza and COVID-19 regarding severity of illness, demographic characteristics and age were comparable to our study [20][21][22]. In one study, vv-ECMO specific prognostic scores before initiation were lower than in our study, which may have been due to different ECMO entry criteria [21]. However, most previous studies included low numbers of patients with COVID-19, all treated in experienced centres and thus limiting the external validity. Furthermore, these studies evaluated patients from the early phases of the pandemic and changes in clinical practice over the course of the pandemic may further reduce generalisability. Additionally, overall experience with initiation and handling of vv-ECMO over the study period may also have had an influence. The largest study compared 53 patients with COVID-19 and 67 patients with influenza and did not show significant outcome differences [23]. Duration on vv/va-ECMO and ICU length of stay in this study was significantly longer in patients with COVID-19, which is in line with our findings. ICU treatment regarding vasopressor support, RRT and rate of tracheostomy was similar in both groups. Greater length of ICU stay for COVID-19 as opposed to influenza patients on vv/va-ECMO in that study may be explained by a higher burden of comorbidities and age.
Our study has several limitations. First, our study included a relatively small number of patients. Larger cohorts are needed to confirm our findings. Second, the data were derived from a single centre and were collected retrospectively. Third, our results in an experienced ECMO centre may not be transferable to other, less experienced, settings. Fourth, our study included patients from several waves of the COVID-19 pandemic. Changes in clinical practice over time may have influenced the outcome of critically ill patients with COVID-19 on vv-ECMO. Fifth, changes in practice and management over the study period from 2009 to 2021 may also have influenced outcomes and could also explain the difference observed in the use of prone position and neuromuscular blockade between influenza and COVID-19 group.

Conclusions
In conclusion, in our study population, 90-day outcomes of patients with severe ARDS on vv/va-ECMO were similar between COVID-19 and influenza patients despite differences in baseline demographic characteristics and comorbidities. Patients with COVID-19 on vv/va-ECMO had significantly higher complications, including thromboembolic and bleeding events, than patients with influenza on vv/va-ECMO. Further, larger studies are needed to confirm these preliminary results.   Institutional Review Board Statement: The study was approved by the local clinical institutional review board and complies with the Declaration of Helsinki. The study was registered with the Ethics Committee of the Hamburg Chamber of Physicians (No.: WF-052/21). Owing to the retrospective character of the study and its pseudonymised data collection, the need for informed consent was waived.

Informed Consent Statement:
The requirement for informed patient consent was waived due to the use of anonymised data from routine clinical care only.
Data Availability Statement: Data sharing is not applicable to this article.