Keywords
SARS-CoV-2 - COVID-19 - respiratory distress syndrome - extracorporeal - membrane
oxygenation - respiratory failure
Introduction
Since late 2019, coronavirus disease 2019 (COVID-19), caused by a novel coronavirus
strain known as severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), has
rapidly spread over the globe causing a pandemic.[1] In addition to adequate lung protective mechanical ventilation, veno-venous extracorporeal
membrane oxygenation (vv-ECMO) has been established as an invasive ultima ratio therapy
to bridge refractory severe respiratory failure since the early beginning of the pandemic.[2]
[3]
[4] Noninvasive and invasive mechanical ventilation is part of the baseline therapy
for acute respiratory distress syndrome (ARDS), however, patients with COVID-19-related
ARDS often require high pressure invasive ventilation causing an increased incidence
of ventilator-induced lung injury such as barotrauma.[5]
[6] Although global registry data suggest comparable outcome of vv-ECMO patients suffering
from COVID-19 and non-COVID-19-related ARDS, recent German data has indicated severely
impaired outcome for COVID-19 patients.[4]
[7]
[8]
[9]
[10]
[11] In fact, global data report mortality of 37% for ECMO patients suffering from COVID-19-related
ARDS.[8]
[11] In contrast, mortality in Germany is reported nearly twice as high with 71 to 73%.[7]
[9] Furthermore, increased mortality has been described between the different infective
surges for both international (approximately between 40 and 60%) as well as German
data (approximately between 60 and 75%) indicating impaired outcome with the persistence
of the pandemic.[7]
[12]
To investigate possible changes and developments in the therapy and outcome of patients
with vv-EMCO support for therapy-refractory COVID-19-related ARDS, we retrospectively
analyzed all patients treated at our center and compared pre-, peri-, and postinterventional
parameters of the four different infective waves between 2020 and 2021.
Patients and Methods
Patients, Study Design, and Follow-Up Period
All patients (n = 91) who underwent ECMO implantation for therapy-refractory COVID-19-related ARDS
between March 2020 and December 2021 in our department were analyzed. Those with any
kind of primary venoarterial ECMO support were excluded (n = 16). The remaining n = 75 patients with isolated vv-ECMO were included in the study and categorized according
to the date of the ECMO implantation. During the study period, four surges (infective
waves) of COVID-19 patients on the intensive care unit (ICU) were observed in Germany.
In line with this, patients were assigned to four study groups: ECMO implantation
between March 2020 and September 2020: first wave (n = 11); October 2020 to February 2021: second wave (n = 23); March 2021 to July 2021: third wave (n = 25); and August 2021 to December 2021: fourth wave (n = 20). Follow-up was performed for 6 months after hospital discharge for every survivor.
[Fig. 1] reflects the four infective waves of the study period in the context of the quantity
of COVID-19 patients on German ICU.
Fig. 1 Schematic illustration of the quantity of coronavirus disease 2019 (COVID-19) patients
on German intensive care units (ICUs) between March 2020 and December 2021. Throughout
the study period four infective surges (first to fourth waves) were identified. Patients
with veno-venous extracorporeal membrane oxygenation for therapy-refractory COVID-19-related
acute respiratory distress syndrome of each wave were prospectively enrolled in an
institutional database and retrospectively compared regarding their peri-interventional
morbidity and mortality.
Study Objectives
Preinterventional baseline data and concomitant diseases as well as peri-interventional
morbidity and mortality were assessed and retrospectively analyzed. Weaning from vv-ECMO
and in-hospital mortality were defined as primary endpoints. In addition, peri-interventional
adverse events including acute kidney failure, neurological complications, and bleeding
complications were defined as secondary endpoints of the study.
ECMO Implantation and Adjuvant Pharmacotherapy
Inclusion and exclusion criteria for vv-ECMO implantation followed the guidelines
of the Extracorporeal Life Support Organization (ELSO) and evolved during the ongoing
pandemic in line with the updated guidelines.[3]
[4]
[13]
[14]
[15] In line with the ELSO interim guideline our institutional standard operating procedure
covers a list of relative contraindications such as age > 65 years, body mass index > 40 kg/m2, high-dose vasopressors or immune deficiency, as well as absolute contraindications
such as severe multiorgan failure, severe neurological injury, contraindication for
anticoagulation, and preexisting life-limiting medical conditions (e.g., end-stage
malignancy).[3]
[15] In general, vv-ECMO was implanted on ICU onsite or in a remote hospital. Cannulation
was performed percutaneously with sonographic guidance. For venous drainage femoral
vein was cannulated with a 25-Fr multistage cannula in every patient. For return of
oxygenated blood, either contralateral femoral vein or internal jugular vein was cannulated
(15–19 Fr cannula). Patients with vv-ECMO implantation in a remote hospital were transferred
to our center directly after implantation and therapy was pursued onsite. Thoracic
X-ray examinations were routinely performed every few days. In addition, thoracic
as well as cerebral computed tomography scans were performed directly after ECMO implantation
and whenever clinically indicated. Pharmacotherapy of ECMO patients followed international
recommendations and varied therefore throughout the pandemic. In general, therapy
included application of antiviral medications (e.g., remdesivir), antibodies (e.g.,
convalescent plasma, casirivimab/imdevimab), immunomodulatory medications (e.g., glucocorticoids,
tocilizumab), and immunoglobulins (Igs) in case of low serum IgM levels.
Statistics
Statistical analyses were calculated with SPSS Statistics 28 (IBM Corporation, Armonk,
New York, United States). All results are displayed as mean values with the standard
deviation respectively percentage of the whole. Due to the small groups sizes Gaussian
distribution was not assumed and variables therefore compared by either nonparametric
two-tailed Kruskal–Wallis tests or Fisher–Freeman–Halton tests. In case of statistically
significant results (p < 0.05), additional post hoc analyses were performed by a Bonferroni correction or
two-tailed Fisher's exact tests. Detailed results of post hoc analyses are displayed
in [Supplementary Table S1] (available in the online version).
Results
Baseline Parameters and Concomitant Diseases of COVID-19 Patients
Baseline parameters and concomitant diseases before implantation of vv-ECMO are displayed
in [Table 1]. In addition, [Fig. 2A] shows the corresponding graphical trends of the most relevant parameters. Throughout
the pandemic, a constant decline of patient age (p = 0.05) as well as incidence of relevant concomitant diseases such as chronic renal
insufficiency (p < 0.01), chronic obstructive pulmonary disease (p = 0.01), and nicotine abuse (p = 0.01) was observed. In contrast, the percentage of obese patients increased since
the first wave. Blood gas analyses indicated hypercapnic pulmonary failure during
the whole study period and additional hypoxic pulmonary failure in waves two and three.
Table 1
Baseline parameters and preexisting concomitant diseases
|
Parameter
|
First wave
|
Second wave
|
Third wave
|
Fourth wave
|
p-Value
|
|
(n = 11)
|
(n = 23)
|
(n = 25)
|
(n = 20)
|
|
|
Age, y
|
58.6 ± 17.6
|
55.7 ± 7.7
|
53.5 ± 9.1
|
49.2 ± 10.6
|
0.05
|
|
Female gender, n (%)
|
2 (18.2)
|
3 (13.0)
|
8 (32.0)
|
6 (30.0)
|
0.41
|
|
Body mass index, kg/m2
|
29.7 ± 4.5
|
27.3 ± 5.1
|
32.2 ± 5.6
|
30.3 ± 4.5
|
< 0.01
|
|
Concomitant diseases
|
|
|
|
|
|
|
Cardiovascular disease, n (%)
|
2 (18.2)
|
5 (21.7)
|
1 (4.0)
|
4 (20.0)
|
0.23
|
|
Arterial hypertension, n (%)
|
8 (72.7)
|
20 (87.0)
|
17 (68.0)
|
7 (35.0)
|
< 0.01
|
|
Diabetes mellitus, n (%)
|
6 (54.5)
|
7 (30.4)
|
4 (16.0)
|
6 (30.0)
|
0.14
|
|
Renal insufficiency, n (%)
|
5 (45.5)
|
5 (21.7)
|
1 (4.0)
|
0 (0.0)
|
< 0.01
|
|
COPD, n (%)
|
5 (45.5)
|
14 (60.9)
|
7 (28.0)
|
3 (15.0)
|
0.01
|
|
Smoking, n (%)
|
9 (81.8)
|
18 (78.3)
|
13 (52.0)
|
7 (35.0)
|
0.01
|
|
Obesity, n (%)
|
5 (36.4)
|
7 (30.4)
|
18 (72.0)
|
9 (45.0)
|
0.03
|
|
Blood gases before ECMO
|
|
|
|
|
|
|
paO2, mm Hg
|
73.8 ± 20.4
|
56.8 ± 17.4
|
56.2 ± 13.2
|
63.9 ± 8.2
|
0.02
|
|
paCO2, mm Hg
|
75.3 ± 27.1
|
72.4 ± 33.5
|
68.5 ± 33.4
|
95.3 ± 44.1
|
0.61
|
|
FiO2, %
|
80 ± 20
|
94 ± 12
|
99 ± 3
|
99 ± 4
|
0.08
|
|
pH, /1
|
7.23 ± 0.14
|
7.23 ± 0.17
|
7.32 ± 0.12
|
7.19 ± 0.18
|
0.36
|
Abbreviations: COPD, chronic obstructive pulmonary disease; ECMO, extracorporeal membrane
oxygenation; FiO2, fraction on inspired oxygen; paO2, arterial partial pressure of oxygen; paCO2, arterial partial pressure of carbon dioxide.
Note: Table shows preinterventional baseline parameters and concomitant disease of
patients undergoing veno-venous extracorporeal membrane oxygenation (ECMO) due to
therapy-refractory coronavirus disease 2019 (COVID-19)-related acute respiratory distress
syndrome. Results of post hoc analyses for statistically significant differences are
given in [Supplementary Table S1] (available in the online version).
Fig. 2 Graphical trends and percentage change from first to fourth waves of different parameters
of patients undergoing veno-venous extracorporeal membrane oxygenation (ECMO) for
therapy-refractory coronavirus disease 2019 (COVID-19)-related acute respiratory distress
syndrome. (A) Preinterventional baseline parameters and concomitant diseases. (B) ECMO support and adjuvant therapy. (C) Outcome parameters. COPD, chronic obstructive pulmonary disease. *Parameter was
not observed during the first wave, therefore no percentage change was calculated.
ECMO Support and Adjuvant Pharmacotherapy
[Table 2] displays details about the ECMO support as well as the pharmacotherapy for COVID-19.
Graphical trends of the most important parameters are illustrated in [Fig. 2B]. We did not observe differences regarding the duration of the pre-ECMO course of
disease. In fact, patients were admitted to the ICU about 9 days after the onset of
the first COVID-19-related symptoms and ECMO was implanted about another 8 days later.
The mean duration of mechanical ventilation of about 4 days before ECMO implantation
did not change over the study period.
Table 2
ECMO support and adjuvant therapy
|
Parameter
|
Frist wave
|
Second wave
|
Third wave
|
Fourth wave
|
p-Value
|
|
(n = 11)
|
(n = 23)
|
(n = 25)
|
(n = 20)
|
|
|
Days since first symptoms and ECMO, d
|
14.8 ± 11.0
|
17.2 ± 6.5
|
15.6 ± 5.7
|
15.6 ± 7.7
|
0.63
|
|
Days since first symptoms and ICU, d
|
9.7 ± 10.6
|
8.8 ± 6.1
|
8.2 ± 8.9
|
8.6 ± 5.2
|
0.72
|
|
Days since ICU and ECMO, d
|
5.1 ± 4.9
|
8.7 ± 6.6
|
8.4 ± 5.3
|
7.0 ± 7.1
|
0.25
|
|
Days since intubation and ECMO, d
|
3.8 ± 4.0
|
4.7 ± 5.3
|
4.9 ± 4.8
|
2.7 ± 3.5
|
0.48
|
|
ECMO parameters
|
|
|
|
|
|
|
Indication
|
|
|
|
|
1.00
|
|
ARDS
|
11 (100.0)
|
23 (100.0)
|
25 (100.0)
|
20 (100.0)
|
|
|
Awake ECMO, n (%)
|
0 (0.0)
|
1 (4.5)
|
1 (4.0)
|
4 (20.0)
|
0.19
|
|
Vascular access
|
|
|
|
|
|
|
Femoro-jugular, n (%)
|
4 (36.4)
|
13 (56.5)
|
24 (96.0)
|
18 (90.0)
|
< 0.01
|
|
Configuration change, n (%)
|
2 (18.2)
|
2 (8.7)
|
0 (0.0)
|
0 (0.0)
|
0.05
|
|
Adjuvant therapy
|
|
|
|
|
|
|
Antibiotics, n (%)
|
11 (100.0)
|
22 (95.7)
|
22 (88.0)
|
20 (100.0)
|
0.40
|
|
Antiviral drugs, n (%)
|
10 (90.9)
|
14 (60.9)
|
18 (72.0)
|
20 (100.0)
|
< 0.01
|
|
Steroids, n (%)
|
5 (45.5)
|
20 (87.0)
|
25 (100.0)
|
20 (100.0)
|
< 0.01
|
|
Convalescent plasma, n (%)
|
2 (18.2)
|
5 (21.7)
|
2 (8.0)
|
0 (0.0)
|
0.10
|
|
Inotropes, n (%)
|
6 (54.5)
|
10 (43.5)
|
14 (56.0)
|
0 (0.0)
|
< 0.01
|
|
Vasopressors, n (%)
|
10 (90.9)
|
21 (91.3)
|
24 (96.0)
|
19 (95.0)
|
0.87
|
|
Prone therapy, n (%)
|
11 (100.0)
|
21 (91.3)
|
21 (84.0)
|
18/19 (94.7)
|
0.55
|
|
Prone therapy before ECMO, n (%)
|
5 (45.5)
|
14 (60.9)
|
12/22 (54.5)
|
7/17 (41.2)
|
0.63
|
|
Type of anticoagulation
|
|
|
|
|
< 0.01
|
|
Heparin, n (%)
|
11 (100.0)
|
23 (100.0)
|
22 (88.0)
|
12 (60.0)
|
|
|
Argatroban, n (%)
|
0 (0.0)
|
0 (0.0)
|
3 (12.0)
|
8 (40.0)
|
|
Abbreviations: ARDS, acute respiratory distress syndrome; ECMO, extracorporeal membrane
oxygenation; ICU, intensive care unit.
Note: Table 2 shows device parameters and adjuvant therapy of patients undergoing
veno-venous extracorporeal membrane oxygenation (ECMO) due to therapy-refractory coronavirus
disease 2019 (COVID-19)-related acute respiratory distress syndrome (ARDS). Results
of post hoc analyses for statistically significant differences are given in [Supplementary Table S1] (available in the online version).
We observed a distinct change regarding ECMO cannulation technique between the first
two and the last two waves. While femoro-femoral cannulation was used in most cases
during the first wave, femoro-jugular access was predominant since the third wave
(p < 0.01). In addition, configuration change from veno-venous to any kind of venoarterial
configuration (first wave: vav-ECMO, n = 2; second wave: vav-ECMO, n = 1 and vva-EMCO, n = 1) due to ventricular failure or refractory hypoxemia was abandoned in our center
within the second wave (p = 0.05). In contrast to that, awake ECMO was implemented and more frequently used
during the fourth wave (20% of cases).
Just as these parameters of ECMO support changed during the pandemic, pharmacotherapy
advanced a lot. While antiviral drugs were used since the very beginning, steroids
became part of the standard therapy within the second wave and were therefore used
in every patient of the last two waves. In contrast to that, application of convalescent
plasma preparations completely vanished by the development of targeted antibody therapies.
Outcome of ECMO Support
[Table 3] shows the relevant outcome parameters of ECMO support and additional graphic trends
are given in [Fig. 2C]. At time of the analyses of the data, one patient of the fourth wave was still on
vv-ECMO support (running since 118 days). Therefore, some data of the fourth wave
only cover 19 patients as indicated in [Table 3]. Overall ECMO support duration as well as ECMO support duration until successful
weaning increased during the pandemic from 11 days in the first wave to 45 days in
the fourth wave (312% increase, p < 0.01), respectively, from 9 days in the first wave to 39 days in the fourth wave
(354% increase, p = 0.08). Maximum duration of ECMO support was 163 days (fourth wave), maximum support
duration until successful device weaning was 131 days (fourth wave). While incidence
of acute kidney failure requiring hemodialysis dropped from 81.8% in the first wave
to 0% in the fourth wave (p < 0.01), incidence of other device-related adverse events such as stroke or bleeding
complications were more often observed within the third and fourth wave. Same trends
were observed for pulmonary superinfections (first wave: 0.0%, fourth wave: 75.0%,
p < 0.01). In-hospital mortality was mostly caused by therapy-refractory respiratory
failure in the first two waves (77.7% respectively 68.8% compared with 12.5 and 27.3%)
and then shifted toward multiorgan dysfunction syndrome (0.0% in first two waves compared
with 43.8 and 45.5%) and neurological injuries (0.0% respectively 12.5% compared with
31.3 and 18.2%).
Table 3
Outcome
|
Parameter
|
First wave
|
Second wave
|
Third wave
|
Fourth wave
|
p-Value
|
|
(n = 11)
|
(n = 23)
|
(n = 25)
|
(n = 20)
|
|
|
Overall ECMO support duration, d
|
10.9 ± 9.6
|
13.0 ± 10.5
|
29.4 ± 32.2
|
44.9 ± 47.0
|
< 0.01
|
|
Adverse events
|
|
|
|
|
|
|
Multiorgan dysfunction, n (%)
|
4 (36.4)
|
4 (17.4)
|
9 (36.0)
|
5/19 (26.3)
|
0.46
|
|
Sepsis, n (%)
|
8 (72.7)
|
9 (39.1)
|
15 (60.0)
|
9/19 (47.4)
|
0.26
|
|
Kidney failure with hemodialysis, n (%)
|
9 (81.8)
|
5 (21.7)
|
5 (20.0)
|
0/19 (0.0)
|
< 0.01
|
|
Major bleeding, n (%)
|
1 (9.1)
|
3 (13.0)
|
9 (36.0)
|
2/19 (10.5)
|
0.11
|
|
Ischemic stroke, n (%)
|
0 (0.0)
|
4 (17.4)
|
1 (4.0)
|
2/19 (10.5)
|
0.33
|
|
Hemorrhagic stroke, n (%)
|
0 (0.0)
|
1 (4.3)
|
7 (28.0)
|
3/19 (15.8)
|
0.06
|
|
Intracranial bleeding, n (%)
|
0 (0.0)
|
6 (26.1)
|
11 (44.0)
|
3/19 (15.8)
|
0.03
|
|
Bowl ischemia, n (%)
|
1 (9.1)
|
2 (8.7)
|
2 (8.0)
|
0/19 (0.0)
|
0.55
|
|
Pneumothorax, n (%)
|
5 (45.5)
|
7 (30.4)
|
5 (20.0)
|
10/20 (50.0)
|
0.16
|
|
Lung bleeding, n (%)
|
2 (18.2)
|
4 (17.4)
|
2 (8.0)
|
4/20 (20.0)
|
0.60
|
|
Pulmonary superinfection, n (%)
|
0 (0.0)
|
1 (4.3)
|
6 (24.0)
|
15/20 (75.0)
|
< 0.01
|
|
ECMO weaning, n (%)
|
2 (18.2)
|
9 (39.1)
|
11 (44.0)
|
8/19 (42.1)
|
0.52
|
|
Support duration till weaning, d
|
8.5 ± 2.1
|
13.4 ± 10.6
|
28.0 ± 18.6
|
38.6 ± 46.9
|
0.08
|
|
In-hospital death, n (%)
|
9 (81.8)
|
16 (69.6)
|
16 (64.0)
|
11/19 (57.9)
|
0.61
|
|
Cause of death
|
|
|
|
|
< 0.01
|
|
Respiratory failure, n (%)
|
7 (77.7)
|
11 (68.8)
|
2 (12.5)
|
3 (27.3)
|
|
|
Neurological injury, n (%)
|
0 (0.0)
|
2 (12.5)
|
5 (31.3)
|
2 (18.2)
|
|
|
Sepsis, n (%)
|
1 (11.1)
|
3 (18.8)
|
1 (6.3)
|
1 (9.1)
|
|
|
Multiorgan dysfunction, n (%)
|
0 (0.0)
|
0 (0.0)
|
7 (43.8)
|
5 (45.5)
|
|
|
Other, n (%)
|
1 (11.1)
|
0 (0.0)
|
1 (6.3)
|
0 (0.0)
|
|
|
Alive at 6 mo
|
2 (18.2)
|
7 (30.4)
|
7/23 (30.4)
|
3/13 (23.1)
|
0.87
|
|
Persistent dyspnea, n (%)
|
1 (50.0)
|
4 (57.1)
|
5 (57.1)
|
1/2 (50.0)
|
1.00
|
|
Persistent oxygen therapy, n (%)
|
0 (0.0)
|
2 (28.7)
|
2 (28.6)
|
1/2 (50.0)
|
1.00
|
|
Cognitive disorders, n (%)
|
0 (0.0)
|
3 (42.9)
|
3 (42.9)
|
1/2 (50.0)
|
0.82
|
Abbreviation: ECMO, extracorporeal membrane oxygenation.
Note: Table 3 shows outcome of patients undergoing veno-venous extracorporeal membrane
oxygenation (ECMO) due to therapy-refractory coronavirus disease 2019 (COVID-19)-related
acute respiratory distress syndrome. One patient of the fourth wave was still on ECMO
support (day 118). Therefore, parts of the outcome parameters of the fourth wave only
cover 19 patients. Results of post hoc analyses for statistically significant differences
are given in [Supplementary Table S1] (available in the online version).
Meanwhile weaning rate numerically increased since the first wave more than two times;
however, this was not statistically significant (p = 0.52). Similar trends were observed for in-hospital death (29% decrease since the
first wave, p = 0.61). Follow-up revealed no late term mortality with every patient surviving the
initial hospital stay being still alive after 6 months.
Discussion
COVID-19-related ARDS requiring vv-ECMO support is still challenging and related with
a quite dismal prognosis. Recent studies suggested even further impaired survival
from one infective wave to the next[7]
[12]: We hereby analyzed all our center's ECMO patients to investigate potential developments
and determinants in the therapy outcome. Although several advancements have been found
and weaning rate and survival numerically increased since the beginning of the pandemic
and outrun German average, in-hospital mortality stayed unsatisfyingly high compared
with global data.
Since the very beginning of the pandemic, tremendous effort has been taken to investigate
pathomechanism and therapeutic targets of SARS-CoV-2 infections.[16] Starting from repurposing well-established drugs to the development of new therapeutics
and vaccines, pharmacotherapy has evolved throughout the reported 2-year study period
in our cohort.[16]
[17] Meanwhile, emerging new virus variants lead to increased transmissions resulting
in periodic infective surges.[16] Both developments have affected our data, as we reported alteration in adjuvant
pharmacotherapy of the analyzed EMCO patients. Furthermore, ECMO therapy itself evolved
as we focused more frequently on femoro-jugular cannulation and awake EMCO from wave
to wave indicating a learning curve. Femoro-jugular instead of femoro-femoral cannulation
for vv-ECMO has been reported as advantageous in patients requiring high ECMO blood
flow which is often the case in COVID-19 patients.[18]
[19] In fact, the femoro-jugular configuration has been reported to be more efficient
in oxygenation and it is also advantageous with the prone position of the patients.[18]
[19]
[20] As COVID-19 patients often required long-term vv-ECMO support with deeply impaired
oxygenation and decarboxylation, as well as multiple prone positions, the advantages
of femoro-jugular access have gained importance and were preferred in our center during
the ongoing pandemic. On the contrary, the concept of awake ECMO is discussed as controversial
in the literature.[21]
[22] Although several case reports describe promising results for awake ECMO in patients
with COVID-19-related ARDS, recent multicenter data could not prove this presumption.[21]
[22]
[23]
[24]
[25] However, in-depth analysis of our data revealed that of all six patients with awake
ECMO, only a single patient had never been at any time on mechanical ventilation which
goes in line with the multicenter data.[21] Nevertheless, four patients were successfully weaned and one patient was still awake
on ECMO support at the end of the study period indicating a favorable weaning rate
in our cohort. Therefore, both changes, preference of femoro-jugular instead for femoro-femoral
cannulation and awake ECMO, have evolved within the pandemic with our center's increasing
experience in the treatment of COVID-19-related ARDS.
ECMO therapy is associated with a variety of severe adverse events such as bleeding
and neurological complications limiting the long-term outcome of patients.[26]
[27] Probably due to the increased ECMO support duration, we consequently observed increased
incidence of bleeding and neurological complications in the recent waves.[26]
[27] Meanwhile, reported causes of death shifted from pulmonary reasons toward more diverse
events, again underlining a learning curve as therapy for severe lung failure was
pursued and not limited to a strict time span. Hereby, we were able to report successful
weaning and discharge of patients with 4 months of continuous ECMO support. Prolonged
ECMO run time with a median duration of 14 to 20 days has been described for COVID-19
patients in the literature.[15] Interestingly, already before COVID-19 patients with ultra-long term ECMO support
of up to 260 days have been described with no disadvantages in survival compared with
patients with regular support duration less than 2 weeks supporting our concept of
continuing support in the absence of limiting adverse events regardless of the actual
run time.[28]
Fortunately, we were able to report a continuous numerical decrease of the in-hospital
mortality since the first wave which is in fact contrary and therefore superior to
both German as well as international multicentric data.[7]
[12] However, mortality of approximately 58% in the fourth wave was still distressingly
worse than global outcome data of approximately 40% reported in the literature.[8]
[11] Comparison with German multicentric data, however, revealed again favorable outcome
for our center compared with the nationwide average.[7]
[9] Therefore, question arises why outcome of ECMO for COVID-19-related ARDS is impaired
in Germany.[7]
[9]
[29] Multiple explanations have already been discussed in the literature with patient
selection and center expertise being the most valuable ones.[7]
[9]
[12]
[29] In this context, Barbaro et al reported significantly impaired outcome of COVID-19
patients with vv-ECMO for inexperienced centers.[12] In addition, several studies have underlined substantial benefits for experienced
ECMO centers with a specialized interdisciplinary ECMO team.[30]
[31]
[32] However, another study by Bailey et al reported worse outcome of high-volume ECMO
centers compared with low-volume institutions that may have been related to a selection
bias with high-volume center probably accepting a wider range of patients for ECMO
support.[33] As an experienced and well-trained center providing both veno-venous as well as
venoarterial ECMO therapy and being part of several ECMO-related research groups,
we follow the mentioned interdisciplinary specialized team approach and generally
accept a wide spectrum of patients for ECMO support.[34]
[35]
[36] In contrast to that, patient selection seems fundamentally responsible for the general
high morbidity observed at our center as well as nationwide in Germany.[7]
[9]
[37] Especially patient age and prevalence of preexisting concomitant diseases such as
chronic kidney or pulmonary diseases within the first waves are most likely related
to the increased in-hospital morbidity.[7]
[9]
[37] In addition, German health care system has never been really overburdened by COVID-19
patients and therefore was able to provide ECMO to a large number of patients with
little triage.[37] However, affected by a general learning curve, patient selection criteria for ECMO
have been revised during the continuance of the pandemic and therefore probably impacted
the outcome.[29]
[38] In fact, our center's policy of inclusion and exclusion criteria to evaluate eligibility
of COVID-19 patients for vv-ECMO from no contraindications except anticipated nonrecovery
following the general ELSO guideline for vv-ECMO support to a list of strict relative
and absolute contraindications as recommended by the evolving specific COVID-19-related
ELSO guidelines.[3]
[4]
[13]
[14]
[15]
Limitations
The single-center and retrospective design of the present study obviously limits its
scientific value. Furthermore, the correspondingly small group sizes prohibited propensity
score matching. As COVID-19 therapy, in particular pharmacotherapy, distinctly improved
throughout the pandemic, adjuvant therapy could most likely acts as a confounder of
the results. In addition, we changed our standard anticoagulation regime from heparin
to argatroban within the third wave, which may also impact the results in an uncertain
manner. While the changes in outcome after ECMO therapy are multifactorial and the
potential effects of the developments in adjuvant pharmacotherapy may exceed the potential
effects of the developments in EMCO cannulation and extracorporeal circulation, ECMO
therapy in general is a multidisciplinary team approach, and therefore we were able
to report changes in the outcome of critically ill COVID-19 patients requiring vv-ECMO.
In particular, we were able to report new insights from a German high-volume center
broadening the general knowledge about ECMO therapy for COVID-19-related ARDS.
Conclusion
vv-ECMO support is still associated with high mortality in COVID-19 patients. Although
advances in pharmacotherapy have been achieved, therapy improvements for ECMO support
had only a minor effect on the weaning rate and in-hospital mortality. However, with
greater knowledge and experience in the treatment of patients with vv-ECMO suffering
from COVID-19-related ARDS during the pandemic, several modifications were implemented.
In particular, preference for femoro-jugular cannulation and awake ECMO combined with
preexisting expertise in ECMO therapy and careful patient selection may be considered
to be associated with the observed increased event-free ECMO support duration, enabling
constant numerical improvements of the outcome contrary to the German and global trend.
