Introduction
The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) already caused over 10 million infections and 500,000 deaths as reported in July 2020.[1] SARS-CoV-2 infection causes hyperactivity of immune and nervous systems and an activation of the coagulation cascade associated with severe hypercoagulability, leading to multiple organ failure.[2] The hypercoagulability leads to an increased risk of venous and arterial thromboembolism with a reported incidence of 31% for intensive care unit (ICU) patients with COVID-19.[3]
[4] Furthermore, COVID-19 acute respiratory distress syndrome (ARDS) patients developed significantly more thrombotic complications than non-COVID-19 ARDS patients (11.7 vs. 2.1%, p < 0.008).[5] Certain laboratory parameters, such as D-dimer and fibrin degradation products (FDP), have been shown to identify patients at risk of thromboembolic and bleeding events (TEBE).[2]
[6]
[7] Thereupon studies investigated different regimes for prophylactic anticoagulant therapy in COVID-19 patients at risk of TEBE. The use of heparin is associated with a significantly lower mortality and better prognosis.[8]
Extracorporeal membrane oxygenation (ECMO) plays a role as a rescue therapy.[9]
[10] In severe non–COVID-19-induced ARDS patients, there is existing conclusive evidence linking (venovenous [V-V]) ECMO support to better patient outcomes and decreased mortality.[11]
[12]
[13] However, ECMO treatment itself increases the risk of TEBE[14] and experience of ECMO treatment in COVID-19 patients remains limited.[9]
[15]
[16]
[17]
The aim of this study was to describe TEBE in critically ill COVID-19 patients receiving ECMO support. Furthermore, we aimed to identify pre-ECMO characteristics that influence the incidence of TEBE in COVID-19 patients treated with ECMO.
Materials and Methods
Patients and COVID-19 Diagnosis
This single-center retrospective observational study included all adult inpatients (≥18 years) between March 1 and April 20, 2020, who were diagnosed with COVID-19 according to the WHO interim guidance[10] and developed severe COVID-19 disease with ARDS requiring ECMO support. Severe cases of COVID-19 with ARDS were identified according to the Berlin Definition of the European Society of Intensive Care Medicine from 2011.
The study was approved by the local ethic commission and the requirement for informed consent was waived by the ethical board due to emerging worldwide public health crisis and retrospective nature of the study.
Data Collection
Beside demographics, medical history, and treatment regimes, time course of laboratory, mechanical ventilation (MV) parameters, and ECMO settings were collected through the entire length of hospital stay (LOS) and outcomes were extracted from our patient data management systems (Philips IntelliSpace Critical Care and Anesthesia and Siemens Medico). Outcomes comprised mortality and complications such as TEBE, organ failure, super infections, and other pathologies. We also gathered data about the amount of packed red blood cells (PRBC) and albumin units. PRBC units contained 200 to 300 mL and albumin units were 100 mL with a concentration of 20%.
Laboratory Analysis
Detailed laboratory analyses were performed daily and included complete blood count electrolytes; measures of hemostasis; hemolysis markers; biochemical tests of cardiac, renal, and liver function; NT-pro-brain natriuretic peptide (NT-proBNP); interleukin-6 (IL-6); procalcitonin (PCT); C-reactive protein (CRP); fibrinogen; and D-dimers. Blood gas analyses were performed in intervals of 1 to 2 hours. The data were collected afterward, and some laboratory parameters were not always documented in the system over the entire ICU stay. However, the pre-ECMO documentation was complete and missing values exist only in measurements after ECMO initiation.
Extracorporeal Membrane Oxygenation Settings
Critically ill COVID-19 patients, who presented with commonly accepted ECMO indications as suggested by the Extracorporeal Life Support Organization (ELSO),[18] in whom all other treatments options have been exhausted (lung protective invasive MV; prone positioning; neuromuscular blockade; and inhaled nitric oxide [iNO] rescue therapy) were considered for ECMO treatment. Decision on initiation of ECMO treatment was judged by consensus following discussion in the ECMO team consisting of internal medicine intensivist, cardiothoracic surgeon, and pneumologist.
The type of ECMO engines were iLA ACTIVVE XLUNG kits (XENIOS, Heilbronn, Germany) and Cardiohelp HLS systems Version 7.0 (Maquet Cardiopulmonary GmbH, Rastatt, Germany). Data about ECMO settings, utilization (i.e., V-V or veno-arterial [V-A]), cannulation sites, utilization switch (i.e., from V-V to V-A or V-VA), and details on ECMO's weaning were extracted. Our standard approach for the treatment of isolated respiratory failure is V-V ECMO.
Anticoagulation Regime
Hemostasis parameters were measured daily and included activated partial thromboplastin time (aPTT), international normalized ratio (INR), platelet count, fibrinogen, antithrombin, d-dimer, and activated clotting time (ACT). aPTT was measured three times daily and ACT was measured four times daily as a control for adequate coagulation and not as a target value. In addition, factor XIII was measured three times a week. Although these measurements were performed and controlled regularly, the documentation in the system was not complete for all variables.
In all patients with VV-ECMO, we primarily administered 400 IE/h of unfractionated heparin (UFH) if no other relevant indications for a higher anticoagulation target, like atrial fibrillation or mechanical heart valve prosthesis, were present. If no bleeding complications occurred, we tolerated an ACT up to 180 seconds. If necessary, UFH was reduced or paused. Other target values were an aPTT between 40 and under 60 seconds, platelet count above 50 G/L, and fibrinogen above 150 mg/dL, and we kept the INR under 1.5. In case of bleedings, we adjusted the target values by using fresh frozen plasma or PRBC. We reduced the ACT under 160 seconds, normalized the INR, and raised the platelet count to at least 80 G/L and fibrinogen to more than 200 mg/dL.
Thromboembolic and Bleeding Events
TEBE include all thromboembolic events (TEE) and bleeding events (BE) that occurred during ECMO therapy. TEE comprised pulmonary artery embolism (PAE), venous or arterial thrombosis, and ischemic or hemorrhagic stroke. For BE, we documented upper respiratory tract (URT) bleedings and divided them into endobronchial and mucosal bleedings. We also reported the incidence of pericardial tamponade, hemothorax, and cannulation side or gastrointestinal bleedings (GIB).
Screening of the above-mentioned complications was conducted daily due to our standard clinical documentation. BE were differentiated into major and minor in case of URT or cannulation side bleedings. Therefore, all bleedings that led to the use of two or more units of whole blood or red blood cells and/or a fall in hemoglobin more than 1.24 mmol/L were identified as major according to the definition of a major bleeding from the International Society on Thrombosis and Haemostasis (ISTH).[19] We chose the ISTH classification due to its applicability for patients treated with anticoagulants and ECMO being a nonsurgical treatment.
Transfusion of PRBC was very frequent and in most patients, we aimed to maintain a hemoglobin value of 9 g/dL. However, as our main ECMO target aside from ultraprotective ventilation would be a sufficient oxygen delivery, we calculated the ratio of oxygen delivery (Do2) to oxygen consumption (Vo2) several times daily and aimed for a ratio of ≥4:1 or at least 3:1. Thus, in some patients, a transfusion of PRBC was required due to low Do2:Vo2 ratios even if hemoglobin was around 10 g/dL.
Confirmation of SARS-CoV-2
Throat-swab specimens, tracheal secretions, or bronchoscopic alveolar lavage were obtained for SARS-CoV-2 from each suspected patient immediately at hospital admission. The real-time reverse transcription polymerase chain reaction (RT-PCR) assay was used to confirm COVID-19 infections.
Statistical Analysis
Categorical variables are presented as absolute numbers (n) and percentages (%). Continuous variables were tested for normal distribution with the Kolmogorov–Smirnov test and presented as median and interquartile range (IQR).
For comparison between patients with and without TEE or BE, univariate analyses were performed using the Mann–Whitney U test for continuous variables and Fisher's exact test for categorical variables.
All statistical comparisons were two-sided. A p value < 0.05 was considered significant. Statistical analysis was performed with SPSS Version 26 (IBM Corp., Armonk, New York, United States). The figure for time courses of laboratory parameters using mean values was created with Microsoft Excel for Mac, Version 16.41 (Microsoft, Redmond, Washington, United States).
Results
Baseline Characteristics
We included 17 patients with COVID-19 who received ECMO from March 1 to April 20, 2020. The median age was 57 years (IQR: 52–62), median BMI was 28 kg/m2 (IQR: 25–32), 6 patients (35%) were female, and the most prevalent comorbidity was arterial hypertension in 13 patients (76%). Nine patients (53%) had fever at admission and the most prevalent typical COVID-19 symptom was dyspnea with 12 patients (71%). Four patients (24%) were nicotine users, and 3 patients (18%) were suffering from COPD. Data on MV and laboratory tests prior to receiving ECMO showed a median PaO2/FiO2 ratio of 0.82 (IQR: 0.68–1.04) and the median duration of MV before ECMO was 3 days (IQR: 2.5–17). Eight patients (47%) had iNO inhalation prior to ECMO support and all patients received antibiotic and prone position treatment. We found a median PaO2 of 68 mm Hg (IQR: 54–74), PaCO2 of 66 mm Hg (IQR: 45–78), d-dimer of 4547 µg/dL (IQR: 1,720–15,395), and fibrinogen of 656 mg/dL (IQR: 421–717). Furthermore, we found a median platelet value of 226 (IQR: 189–421) and the median PTT was 33 seconds (IQR: 28–37). The Respiratory ECMO Survival Prediction (RESP) score was estimated before ECMO therapy and predicted a median score of –1 (IQR: –3 to 1). Further parameters and details are presented in [Table 1].
Table 1
Baseline characteristics (pre-ECMO)
|
Variable
|
Total (n = 17)
|
TEE (n = 7)
|
No TEE (n = 10)
|
p-value[a]
|
BE (n = 10)
|
No BE (n = 7)
|
p-value[a]
|
|
Age (y), median (IQR)
|
57 (52–62)
|
58 (55–67)
|
57 (47–62)
|
0.364
|
57 (50–61)
|
60 (53–62)
|
0.536
|
|
Female sex, n (%)
|
6 (35)
|
1 (14)
|
5 (50)
|
0.160
|
4 (40)
|
2 (29)
|
1.000
|
|
BMI (kg/m2), median (IQR)
|
28 (25–32)
|
26 (25–31)
|
30 (26–36)
|
0.193
|
28 (25–33)
|
30 (25–31)
|
0.887
|
|
LVEF (%), median (IQR)
|
55 (50–56)
|
54 (49–57)
|
55 (50–55)
|
0.740
|
55 (50–56)
|
55 (50–56)
|
0.740
|
|
LOS pre-ECMO (d), median (IQR)
|
5 (4–19)
|
5 (3–21)
|
5 (4–14)
|
1.000
|
4.5 (4–14)
|
13 (3–21)
|
0.740
|
|
Arterial hypertension, n (%)
|
13 (76)
|
5 (71)
|
8 (80)
|
1.000
|
7 (70)
|
6 (86)
|
0.603
|
|
Coronary artery disease, n (%)
|
1 (6)
|
0 (0)
|
1 (10)
|
1.000
|
0 (0)
|
1 (14)
|
0.412
|
|
Diabetes mellitus, n (%)
|
5 (29)
|
2 (29)
|
4 (40)
|
1.000
|
3 (30)
|
3 (43)
|
0.644
|
|
Peripheral arterial disease, n (%)
|
1 (6)
|
1 (14)
|
0 (0)
|
0.412
|
1 (10)
|
0 (0)
|
1.000
|
|
Cerebrovascular disease, n (%)
|
0 (0)
|
0 (0)
|
0 (0)
|
1.000
|
0 (0)
|
0 (0)
|
1.000
|
|
Nicotine users, n (%)
|
4 (0.24)
|
1 (14)
|
3 (30)
|
0.603
|
1 (10)
|
3 (43)
|
0.250
|
|
History of pneumonia, n (%)
|
5 (0.29)
|
1 (14)
|
4 (40)
|
0.278
|
2 (20)
|
3 (43)
|
0.593
|
|
COPD, n (%)
|
3 (0.18)
|
1 (14)
|
2 (20)
|
1.000
|
1 (10)
|
2 (29)
|
0.537
|
|
Hyperlipoproteinemia, n (%)
|
2 (0.12)
|
1 (14)
|
1 (10)
|
1.000
|
2 (20)
|
0 (0)
|
0.485
|
|
Atrial fibrillation, n (%)
|
6 (0.35)
|
4 (57)
|
2 (20)
|
0.162
|
3 (30)
|
3 (43)
|
0.644
|
|
Chronic kidney disease, n (%)[b]
|
2 (12)
|
1 (14)
|
1 (10)
|
1.000
|
1 (10)
|
1 (14)
|
1.000
|
|
Prior myocardial infarction, n (%)
|
1 (06)
|
0 (0)
|
1 (10)
|
1.000
|
0 (0)
|
1 (14)
|
0.412
|
|
Prior PCI, n (%)
|
1 (6)
|
0 (0)
|
1 (10)
|
1.000
|
0 (0)
|
1 (14)
|
0.412
|
|
Prior pulmonary hypertension, n (%)
|
2 (12)
|
1 (14)
|
1 (10)
|
1.000
|
1 (10)
|
1 (14)
|
1.000
|
|
Prior cardiac surgery, n (%)
|
1 (6)
|
1 (14)
|
0 (0)
|
0.412
|
1 (10)
|
0 (0)
|
1.000
|
|
Prior lung surgery, n (%)
|
1 (6)
|
0 (0)
|
1 (10)
|
1.000
|
0 (0)
|
1 (14)
|
0.412
|
|
History of malignancy, n (%)
|
1 (6)
|
0 (0)
|
1 (10)
|
1.000
|
0 (0)
|
1 (14)
|
0.412
|
|
Immunosuppressive medication, n (%)
|
1 (6)
|
0 (0)
|
1 (10)
|
1.000
|
0 (0)
|
1 (14)
|
0.412
|
|
Typical symptoms, n (%)
|
17 (100)
|
5 (71)
|
6 (60)
|
1.000
|
6 (60)
|
5 (71)
|
1.000
|
|
Fever at admission, n (%)
|
9 (53)
|
4 (57)
|
5 (50)
|
1.000
|
5 (50)
|
4 (57)
|
1.000
|
|
Temperature (°C), median (IQR)
|
38.6 (37.1–38.9)
|
38.6 (36.9–38.8)
|
38.3 (37.3–39.1)
|
0.669
|
38.3 (36.8–38.9)
|
38.7 (37.3–38.9)
|
0.740
|
|
Cough, n (%)
|
10 (59)
|
5 (71)
|
5 (50)
|
0.622
|
6 (60)
|
4 (57)
|
1.000
|
|
Dyspnea, n (%)
|
12 (71)
|
6 (86)
|
6 (60)
|
0.338
|
8 (80)
|
4 (57)
|
0.593
|
|
Diarrhea, n (%)
|
3 (18)
|
1 (14)
|
2 (20)
|
1.000
|
2 (20)
|
1 (14)
|
1.000
|
|
Antiviral treatment, n (%)
|
3 (18)
|
2 (29)
|
1 (10)
|
0.537
|
2 (20)
|
1 (14)
|
1.000
|
|
Inotropes, n (%)
|
12 (71)
|
4 (57)
|
8 (80)
|
0.593
|
7 (70)
|
5 (71)
|
1.000
|
|
Vasopressor, n (%)
|
15 (88)
|
6 (86)
|
9 (90)
|
1.000
|
8 (80)
|
7 (100)
|
0.485
|
|
iNO inhalation, n (%)
|
8 (47)
|
3 (43)
|
5 (50)
|
0.581
|
6 (60)
|
2 (29)
|
0.335
|
|
Mechanical ventilation settings
|
|
MV pre-ECMO (d), median (IQR)
|
3 (2.5–17)
|
3 (3–20)
|
3.5 (1.8–12)
|
0.536
|
3.5 (2.8–13)
|
3 (2–20)
|
0.962
|
|
pO2:FiO2 (ratio), median (IQR)
|
0.82 (0.68–1.04)
|
0.76 (0.6–0.83)
|
0.97 (0.7–1.2)
|
0.161
|
0.91 (0.7–1)
|
0.83 (0.5–1)
|
0.813
|
|
Pinsp (mbar), median (IQR)
|
28 (25–30)
|
28 (26–30)
|
28 (25–30)
|
0.536
|
28 (27–30)
|
26 (22–30)
|
0.536
|
|
PEEP (mbar), median (IQR)
|
14 (11–15)
|
14 (13–15)
|
13.5 (9.5–15)
|
0.669
|
14 (13–15)
|
12 (10–15)
|
0.315
|
|
VTe (mL), median (IQR)
|
320 (219–477)
|
347 (295–500)
|
293 (153.8–455)
|
0.193
|
338 (258–520)
|
295 (157–440)
|
0.193
|
|
Vf (n/min), median (IQR)
|
24 (22–30)
|
24 (22–24)
|
25 (21.3–31.3)
|
0.813
|
24 (22–27)
|
24 (19–35)
|
1.000
|
|
CVP (mm Hg), median (IQR)
|
14 (11–15.5)
|
11 (10–15)
|
15 (13–17)
|
0.109
|
14 (11–15)
|
15 (10–18)
|
1.000
|
|
ECMO settings
|
|
V-V, n (%)
|
16 (0.94)
|
6 (86)
|
10 (100)
|
0.412
|
9 (90)
|
7 (100)
|
1.000
|
|
V-A, n (%)
|
1 (0.06)
|
1 (14)
|
0 (0)
|
0.412
|
1 (10)
|
0 (0)
|
1.000
|
|
Fem.-Jug., n (%)
|
7 (0.41)
|
3 (43)
|
4 (40)
|
1.000
|
3 (30)
|
4 (57)
|
0.350
|
|
Fem.-Fem., n (%)
|
1 (0.06)
|
1 (14)
|
0 (0)
|
0.412
|
1 (10)
|
0 (0)
|
1.000
|
|
Right IJV dual cannula, n (%)
|
9 (0.53)
|
3 (43)
|
6 (60)
|
0.637
|
6 (60)
|
3 (43)
|
0.637
|
|
Cytokine absorption, n (%)
|
8 (47)
|
2 (29)
|
6 (60)
|
0.335
|
5 (50)
|
3 (43)
|
1.000
|
|
Blood gas and laboratory tests
|
|
pO2 (mm Hg), median (IQR), (ref. 80–105)
|
68 (54–74)
|
68 (58–82)
|
65 (52–71)
|
0.364
|
68.5 (62–77)
|
55 (49–72)
|
0.161
|
|
pCO2 (mm Hg), median (IQR), (ref. 35–45)
|
66 (45–78)
|
77 (55–80)
|
60 (34–70)
|
0.088
|
68 (51–77)
|
66 (32–88)
|
0.813
|
|
pH, median (IQR), (ref. 7.35–7.45)
|
7.3 (7.2–7.4)
|
7.3 (7.2–7.3)
|
7.3 (7.2–7.4)
|
0.364
|
7.3 (7.2–7.3)
|
7.3 (7.2–7.4)
|
0.475
|
|
BE (mmol/L), median (IQR), (ref. -2–2)
|
1.8 (–4.2 to 6.5)
|
2.3 (–5.5 to 6.5)
|
–1.6 (–4 to 6.8)
|
0.669
|
–1.6 (–5.5 to 6.8)
|
2.1 (–3.5 to 6.5)
|
0.536
|
|
HCO3 (mmol/L), median (IQR), (ref. 22–29)
|
26 (24–35)
|
31 (25–35)
|
25 (22–33)
|
0.270
|
26 (24–35)
|
31 (25–34)
|
0.813
|
|
Na+ (mmol/L), median (IQR), (ref. 135–147)
|
145 (142–151)
|
145 (144–151)
|
143 (137–152)
|
0.536
|
144 (140–151)
|
146 (142–151)
|
0.601
|
|
K+ (mmol/L), median (IQR), (ref. 3.5–5)
|
4.4 (3.8–4.65)
|
4.5 (4.2–4.7)
|
4.1 (3.8–4.7)
|
0.230
|
4.4 (3.8–4.6)
|
4.4 (3.8–5.1)
|
1.000
|
|
Hb (g/dL), median (IQR), (ref. 12–17.5)
|
9.3 (7.9–11)
|
9.4 (8.4–12.5)
|
8.5 (7.5–10.5)
|
0.230
|
9.2 (7.9–10.4)
|
9.9 (7–12.1)
|
0.887
|
|
WBC (/nL), median (IQR), (ref. 4.5–11)
|
11.1 (9.8–15.35)
|
10.4 (9.6–13.4)
|
12.4 (9.9–18.6)
|
0.315
|
10.2 (9.5–15.3)
|
13.2 (10.4–16)
|
0.315
|
|
Platelet (G/L), median (IQR), (ref. 150–450)
|
226 (180–421)
|
213 (145–290)
|
232.8 (192–471)
|
0.740
|
209 (145–238)
|
452 (240–560)
|
0.007[a]
|
|
Creatinine (mg/dL), median (IQR), (ref. 0.84–1.21)
|
1.1 (0.8–1.7)
|
0.9 (0.8–1.8)
|
1.2 (0.9–2)
|
0.740
|
1.2 (0.9–2.3)
|
1.1 (0.7–1.6)
|
0.669
|
|
BUN (mg/dL), median (IQR), (ref. 7–20)
|
64 (41–114)
|
62 (29–116)
|
65 (44–116)
|
0.669
|
48.5 (33.5–135.5)
|
66 (62–77)
|
0.601
|
|
GLDH (U/L), median (IQR), (ref. 1–10)
|
8.6 (3.6–12)
|
5.6 (3.2–12)
|
8.7 (3.8–12.5)
|
0.887
|
7.1 (3–12.5)
|
8.8 (3.9–12)
|
0.740
|
|
CK (U/L), median (IQR), (ref. 22–198)
|
234 (60–657)
|
167 (45–234)
|
592 (66.3–1672.5)
|
0.193
|
201 (66–590)
|
598 (46–4155)
|
0.417
|
|
CK-MB (U/L), median (IQR), (ref. 5–25)
|
21 (18–40)
|
27 (18–53)
|
19 (12–34)
|
0.270
|
22 (16.5–53.5)
|
20 (18–27)
|
1.000
|
|
TpT (µg/mL), median (IQR), (ref. <22)
|
22 (12–59)
|
22 (9–52)
|
31.5 (11.8–73.3)
|
0.601
|
21 (13.5–56.8)
|
50 (10–65)
|
0.887
|
|
Bilirubin (mg/dL), median (IQR), (ref. 0.2–0.8)
|
0.5 (0.41–1.21)
|
0.4 (0.4–0.7)
|
0.7 (0.5–1.7)
|
0.109
|
0.5 (0.4–2.2)
|
0.6 (0.4–0.8)
|
0.887
|
|
ALT (U/L), median (IQR), (ref. 19–33)
|
37 (29–46)
|
39 (33–45)
|
32.5 (25.3–104.3)
|
0.740
|
34.8 (19.8–44.5)
|
39 (32–99)
|
0.364
|
|
AST (U/L), median (IQR), (ref. 5–40)
|
66 (38–140)
|
66 (37–91)
|
93.5 (36.5–174.3)
|
0.475
|
63.5 (37.8–108.5)
|
91 (35–153)
|
0.887
|
|
fHb (mg/L), median (IQR), (ref. <50)
|
36 (21–47)
|
25 (20–60)
|
36 (22.3–44)
|
0.887
|
30.5 (19.8–52.5)
|
36 (23–44)
|
0.740
|
|
LDH (U/L), median (IQR), (ref. 140–280)
|
403 (309.5–559.5)
|
403 (314–459)
|
398 (303–692)
|
0.813
|
382 (303–525.3)
|
403 (322–639)
|
0.887
|
|
PTT (s), median (IQR), (ref. 60–70)
|
33 (28–37)
|
35 (29–37)
|
33 (28–36)
|
0.740
|
35 (28–39)
|
29 (28–34)
|
0.193
|
|
INR, median (IQR), (ref. 2–3)
|
1.3 (1.2–1.3)
|
1.3 (1.2–1.3)
|
1.3 (1.2–1.3)
|
0.887
|
1.2 (1.2–1.3)
|
1.3 (1.3–1.4)
|
0.133
|
|
ATIII (%), median (IQR), (ref. 60–123)
|
66 (52–76)
|
66 (54–87)
|
68 (46.8–76)
|
0.887
|
66.5 (55.5–78.8)
|
55 (40–76)
|
0.475
|
|
IL-6 (µg/mL), median (IQR), (ref. 0–16.4)
|
256 (93–642)
|
122 (104–931)
|
286.8 (70.6–544.4)
|
1.000
|
253.7 (98.7–497.7)
|
317.6 (79.4–1329)
|
0.813
|
|
CRP (ng/mL), median (IQR), (ref. <10)
|
187 (116–284)
|
211 (162–280)
|
165 (108–295)
|
0.813
|
243 (86–295)
|
183 (120–276)
|
0.536
|
|
PCT (%), median (IQR), (ref. 0.22–0.24)
|
2.2 (0.56–6.14)
|
0.6 (0.5–5.6)
|
5.1 (0.8–7.1)
|
0.230
|
1.6 (0.6–5.8)
|
5.2 (0.4–9.1)
|
0.740
|
|
D-dimer (µg/dL), median (IQR), (ref. <250)
|
4,547 (1,720–15,395)
|
11,567 (1,876–25,097)
|
4,174 (1,626–7,950)
|
0.475
|
5,193 (1,732–34,031)
|
4,547 (1,420–12,046)
|
0.669
|
|
Fibrinogen (mg/dL), median (IQR), (ref. 200–400)
|
656 (421–717)
|
656 (432–721)
|
569 (382.5–712.5)
|
0.669
|
569 (426.3–723.8)
|
656 (303–717)
|
0.740
|
|
NT-proBNP (pg/mL), median (IQR), (ref. <300)
|
1,765 (522–4,274)
|
1,765 (438–7,365)
|
2,089 (495–4,199)
|
1.000
|
1,914 (389–4,668)
|
1,765 (605–4,425)
|
0.887
|
|
ICU risk scores
|
|
Resp score, median (IQR)
|
–1 (–3 to 1)
|
–2 (–4 to 0)
|
0 (–3 to 2)
|
0.315
|
–0.5 (–2.5 to 0)
|
–2 (–3 to 2)
|
0.813
|
|
GCS, median (IQR)
|
3 (3–3)
|
3 (3–3)
|
3 (3–3)
|
0.669
|
3 (3–3)
|
3 (3–3)
|
0.740
|
|
NEMS, median (IQR)
|
39 (39–45)
|
39 (39–44)
|
39.5 (37.8–46.3)
|
0.475
|
39 (37–44)
|
39 (39–45)
|
0.601
|
|
NEMS + SAS, median (IQR)
|
43 (41–48)
|
42 (41–47)
|
46 (42–48)
|
0.230
|
43 (41–48)
|
47 (42–49)
|
0.315
|
|
RASS, median (IQR)
|
–4 (–4 to 4)
|
–4 (–4 to 4)
|
–4 (–4.6 to 4)
|
0.193
|
–4 (–4.1 to 4)
|
–4 (–4 to 4)
|
0.887
|
|
SAPS II, median (IQR)
|
46 (36–57)
|
52 (35–57)
|
42 (36.5–51.8)
|
0.475
|
49 (35–57.8)
|
44 (37–52)
|
0.740
|
|
SAS, median (IQR)
|
2 (2–3)
|
2 (2–2.5)
|
2.4 (1.9–3)
|
0.962
|
2 (2–3.3)
|
2.5 (2–3)
|
0.962
|
|
SOFA, median (IQR)
|
9 (7–12.5)
|
11 (7–13)
|
9 (7–11.5)
|
0.669
|
11 (7–13.8)
|
9 (7–11)
|
0.536
|
|
Core-10-TISS, median (IQR)
|
17 (14–21)
|
16 (10–21)
|
17.5 (14–21.3)
|
0.536
|
17 (13.8–21.3)
|
17 (14–21)
|
0.962
|
Abbreviations: ALT, alanine transaminase; AST, aspartate transaminase; ATIII, antithrombin III; BMI, body mass index (kg/m2); BUN, blood urea nitrogen; CK, creatine kinase; CK-MB, creatine kinase myocardial band; COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein (mg/L); CVP, central venous pressure; ECMO, extracorporeal membrane oxygenation; Fem., femoral vein; fHb, free plasma hemoglobin; GLDH, glutamate dehydrogenase; Hb, hemoglobin; iNO, inhaled nitric oxide; ICU, intensive care unit; IJV, internal jugular vein; IL-6, interleukin-6; Jug., jugular vein; LDH, lactate dehydrogenase; LVEF, left ventricular ejection fraction; MPAP, mean pulmonary artery pressure; MV, mechanical ventilation; NT-proBNP, N-terminal prohormone of brain natriuretic peptide; PCT, procalcitonin; PHT, pulmonale hypertension; PCI, percutaneous coronary intervention; PTT, partial thromboplastin time; PASP, pulmonary artery systolic pressure; PEEP, positive end-expiratory pressure; Pinsp, maximal inspiratory pressure; ref., reference value; RHF, right heart failure; TpT, topotecan, Vf, ventilation frequency; VTe, expiratory tidal volume; WBC, white blood cells; TnT, troponin T; V-A, venoarterial; V-V, venovenous.
Abbreviations of risk score systems: Core-10-TISS, Therapeutic Intervention Scoring System; GCS, Glasgow Coma Score; INR, international normalized ratio; NEMS, Nine equivalents of Nursing Manpower Use Score; RASS, Richmond Agitation–Sedation Scale; RESP, The Respiratory Extracorporeal Membrane Oxygenation Survival Prediction; SAPS II, Simplified Acute Physiology Score II; SAS, Riker Sedation–Agitation Scale; SOFA, Sequential Organ Failure Assessment Score.
a
p-values are calculated between the two previous columns. Under 0.05 is considered significant.
b Including all patients with a MDRD-GFR < 60 mL/min.
During ECMO therapy, we observed TEE in 7 patients and BE in 10 patients. We compared the TEE group (n = 7) with a non-TEE group (n = 10), and the BE group (n = 10) with a non-BE group (n = 7). Besides patient characteristics, we included laboratory measurements, MV parameters, ECMO settings, and ICU risk scores before ECMO implantation. Baseline characteristics of all groups are provided in [Table 1].
In the first comparison between the TEE group and the non-TEE group, we found no statistically significant differences. However, there were several notable differences. In the seven patients with TEE, we found fewer female patients (14 vs. 50%, p = 0.160), more patients suffering from atrial fibrillation (57 vs. 20%, p = 0.162), and less utilization of cytokine absorption (29 vs. 60%, p = 0.335). Furthermore, the TEE group had higher median pCO2 values with 77 (IQR: 55–80) versus 60 (IQR: 34–70), with p = 0.088; and the group also had lower median RESP score values with –2 (IQR: –4 to 0) versus 0 (–3 to 2), p = 0.315.
As for the 10 patients with BE, we found one statistically significant difference and some notable differences regarding coagulation parameters. The median platelet count was significantly lower in the BE group than in the non-BE group with 209 (IQR: 145–238) versus 452 (IQR: 240–560), p = 0.007. The BE group also had a longer median PTT with 35 (IQR: 28–39) versus 29 (IQRL 28–34) seconds, p = 0.193; and the group also had a lower median INR with 1.2 (IQR: 1.2–1.3) versus 1.3 (IQR: 1.3–1.4), p = 0.133.
Outcome and Clinical Course
The median ECMO support in our 17 patients was 16 days (IQR: 10.5–22). As of June 1, 2020, nine patients (53%) have been weaned successfully from ECMO and survived to hospital discharge and eight patients (47%) died during ECMO treatment. Seven patients (41%) developed TEE and 10 patients (59%) had BE during ECMO treatment. Patients frequently developed both TEE and BE; thus, five patients (71%) of the TEE group had BE and five patients (50%) of the BE group had TEE. Furthermore, we observed acute kidney failure with renal replacement therapy in 12 patients (71%), bacterial superinfections in 8 patients (47%), and right heart failure in 7 patients (41%). More details can be found in [Table 2].
Table 2
Outcomes and clinical course
|
Variable
|
Total (n = 17)
|
TEE (n = 7)
|
No TEE (n = 10)
|
p-value[a]
|
BE (n = 10)
|
No BE (n = 7)
|
p-value[a]
|
|
Mortality, n (%)
|
8 (47)
|
3 (43)
|
5 (50)
|
0.581
|
6 (60)
|
2 (29)
|
0.335
|
|
TEE, n (%)
|
7 (41)
|
7 (100)
|
0 (0)
|
0.000[a]
|
5 (50)
|
2 (29)
|
0.622
|
|
BE, n (%)
|
10 (59)
|
5 (71)
|
5 (50)
|
0.622
|
10 (100)
|
0 (0)
|
0.000[a]
|
|
Right heart failure, n (%)[b]
|
7 (41)
|
5 (71)
|
2 (20)
|
0.058
|
4 (40)
|
3 (43)
|
1.000
|
|
Bacterial superinfection, n (%)
|
8 (47)
|
2 (29)
|
6 (60)
|
0.335
|
4 (40)
|
4 (57)
|
0.637
|
|
Fungal superinfection, n (%)
|
3 (18)
|
1 (14)
|
2 (20)
|
1.000
|
2 (20)
|
1 (14)
|
1.000
|
|
DIC, n (%)
|
4 (24)
|
2 (29)
|
2 (20)
|
1.000
|
4 (40)
|
0 (0)
|
0.103
|
|
Delirium, n (%)
|
4 (24)
|
1 (14)
|
3 (30)
|
0.603
|
2 (20)
|
2 (29)
|
1.000
|
|
Acute kidney failure, n (%)[c]
|
12 (71)
|
6 (86)
|
6 (60)
|
0.338
|
9 (90)
|
3 (43)
|
0.101
|
|
RRT initiated, n (%)
|
12 (71)
|
6 (86)
|
6 (60)
|
0.338
|
9 (90)
|
3 (43)
|
0.101
|
|
Blood products
|
|
ECMO flow (L/min), median (IQR)[d]
|
4 (3.7–4.4)
|
4 (3.2–4.2)
|
4.2 (3.7–4.4)
|
0.417
|
4.2 (3.4–4.4)
|
4 (3.8–4.4)
|
0.887
|
|
Total PRBCs (units), median (IQR)
|
32 (24–46)
|
33 (26–39)
|
31 (15–55.3)
|
0.813
|
32 (25.5–43.5)
|
33 (12–52)
|
0.669
|
|
PRBCs during ECMO (units), median (IQR)
|
28 (3–69)
|
26 (16–35)
|
24 (14–48.5)
|
1.000
|
28 (23.8–36.5)
|
14 (8–35.8)
|
0.093
|
|
PRBC per ECMO (units/d), median (IQR)
|
1.52 (1.13–2.7)
|
1.51 (1.13–1.78)
|
1.56 (1.26–1.88)
|
0.669
|
1.64 (1.52–2.7)
|
0.76 (0.5–1.78)
|
0.270
|
|
Total albumin (units), median (IQR)
|
85 (4–266)
|
9 (1–251)
|
103 (9–300)
|
0.669
|
124 (8–365)
|
12 (0–167)
|
0.230
|
|
Hospitalization details
|
|
Total LOS (d), median (IQR)
|
24 (11–56)
|
37 (14–64)
|
23 (6–41)
|
0.417
|
20 (6–56)
|
30 (14–65)
|
0.417
|
|
Total ICU stay (d), median (IQR)
|
24 (16–56)
|
37 (18–62)
|
23 (6–53)
|
0.315
|
30 (17–58)
|
24 (1–53)
|
0.669
|
|
ECMO duration (d), median (IQR)
|
16 (11–22)
|
16 (13–23)
|
14 (10–22)
|
0.601
|
16 (11–24)
|
16 (10–21)
|
0.669
|
|
iNO duration (d), median (IQR)
|
2 (0–4.5)
|
2 (0–5)
|
1.5 (0–5.5)
|
0.813
|
2.5 (0.8–7)
|
0 (0–3)
|
0.133
|
Abbreviations: BE, bleeding events; DIC, disseminated intravascular coagulation; ECMO, extracorporeal membrane oxygenation; ICU, intensive care unit; iNO, inhaled nitric oxide; iQR, interquartile range; LOS, length of stay; MI, myocardial infarction; MV, mechanical ventilation; PRBCS, packed red blood cell; RRT, renal replacement therapy; TEE, thromboembolic events.
a
p-values are calculated between the two previous columns. Under 0.05 is considered significant.
b Including all patients with right heart failure in stage 3 or 4 by the American Heart Association guidelines.
c Including all patients with acute kidney failure in stage 2 or 3 by Kidney Disease Improving Global Outcomes (KDIGO) guidelines.
d We presented the blood flow rate that was adjusted on the first day of ECMO therapy.
Outcomes, usage of blood products, and hospitalization details between the investigated groups showed no statistically significant differences ([Table 2]). Nevertheless, some relevant differences could be found and are mentioned in this section.
Starting with the TEE group (n = 7) compared with the non-TEE group (n = 10), we found comparable values for mortality (43 vs. 50%, p = 0.581) and median ECMO duration with 16 days (IQR: 13–23) versus 14 days (IQR: 10–22), p = 0.601. However, the TEE group developed right heart failure more frequently with 71 versus 20% (p = 0.058) and the median ICU stay was longer with 37 days (IQR: 18–62) versus 23 days (IQR: 6–53), p = 0.315.
As for the BE group (n = 10), we found higher mortality rates than in the non-BE group (n = 7) with 60 versus 29% (p = 0.335) and more patients who developed a disseminated intravascular coagulation (DIC) with 40 versus 0%, p = 0.103. Usage of blood products during ECMO was higher in the BE group; we also found more units of PRBC per ECMO day with 1.64 (IQR: 1.52–2.7) versus 0.76 (IQR: 0.5–1.78), p = 0.270 and more units of albumin over the whole ICU stay with 124 (IQR: 8–365) versus 12 (IQR: 0–167), p = 0.230.
Thromboembolic and Bleeding Events
During ECMO therapy, we had 12 patients (71%) who developed TEBE and 5 patients (29%) had a clinical course without TEBE. We used anticoagulation with UFH in 16 patients (94%). One patient (6%) received the thrombin-inhibitor Argatroban instead of UFH, due to heparin-induced thrombocytopenia II (HIT II). This patient developed a PAE and a major endobronchial bleeding despite altered anticoagulation. Moreover, we found two patients (12%) who already had TEBE after hospital admission but prior to ECMO implantation, and both had PAE. During ECMO, these two patients suffered again from PAE and developed bleedings.
[Table 3] shows the incidence of different TEBE during ECMO in detail. We summarized the results here. TEE occurred in 7 patients (41%) and BE in 10 patients (59%).
Table 3
Thromboembolic and bleeding events (TEBE)
|
Outcome variable
|
Total (n = 17)
|
|
Total TEBE,
n
(%)
|
12 (71)
|
|
TEE,
n
(%)
|
7 (41)
|
|
Pulmonary artery embolism
|
5 (29)
|
|
Thrombosis
|
3 (18)
|
|
arterial
|
1 (6)
|
|
venous
|
2 (12)
|
|
Ischemic or hemorrhagic stroke
|
0 (0)
|
|
BE,
n
(%)
|
10 (59)
|
|
Upper respiratory tract
|
8 (47)
|
|
endobronchial
|
5 (29)
|
|
major
|
2 (12)
|
|
minor
|
3 (18)
|
|
mucosal
|
7 (41)
|
|
major
|
2 (12)
|
|
minor
|
5 (29)
|
|
Pericardial tamponade
|
2 (12)
|
|
Cannulation side
|
5 (29)
|
|
major
|
2 (12)
|
|
minor
|
3 (18)
|
|
Gastrointestinal
|
1 (6)
|
|
Hemothorax
|
1 (6)
|
Abbreviations: BE, bleeding events; TEE, thromboembolic events.
Among the seven patients with TEE were five patients (29%) with a PAE. Three patients (18%) had thrombosis consisting of two venous cases (12%) and one arterial case (6%). Both venous cases were a thrombosis in the right vena jugularis. The patient with arterial thrombosis had a thrombus formation in the fingers of the right hand. Ischemic or hemorrhagic strokes did not occur.
Among the 10 patients with BE were 8 patients (47%) with a URT bleeding. To be precise, we observed seven patients (41%) with a mucosal URT bleeding, of which two cases (12%) were a major bleeding. Five patients (29%) had an endobronchial URT bleeding, of which two cases (12%) were major bleeding. There were five patients (29%) with a cannulation side bleeding from which two cases (12%) were a major bleeding. We also reported two patients (12%) with a pericardial tamponade, one patient (6%) with a GIB and one patient (6%) with a hemothorax.
Time Course of Coagulation Parameters
Multiple parameters were measured on a daily basis to adjust anticoagulation therapy, where necessary. Due to imperfect documentation after ECMO initiation, we could not present all parameters. However, four different variables had less than 10% missing values and could be used. Therefore, we displayed time courses of PTT, platelet count, INR, and d-dimer using mean values in [Fig. 1].
Fig. 1 Time courses of PTT, platelet count, INR, and d-dimer using mean values.
PTT values were between 32 and 35 seconds before ECMO implantation and had an increasing trend during the therapy in all groups.
Regarding platelet count, the BE group had less platelets compared with the non-BE group in the first 3 days. As mentioned before, the difference in platelet count before ECMO implantation was statistically significant. However, the platelet count in the BE group was continuously above 80 G/L. This threshold was used in ECMO patients with bleeding by our internal guidelines.
INR values were higher in patients with TEE and BE. During statistical comparison, we could not find significant differences.
D-dimer values decreased in the first day after ECMO initiation; after that, we observed an increasing trend until the end of ECMO therapy in all patient groups.
Discussion
ARDS in COVID-19 patients showed a higher risk for TEBE than in non-COVID-19 patients.[5] Considering the WHO and ELSO guidelines,[10]
[18] ECMO plays a role as a rescue therapy in case of critically ill COVID-19 patients suffering from hypoxemic respiratory failure. However, incidence and risk factors for TEBE in COVID-19 patients treated with ECMO are still uncertain.
In this study, we described TEBE in the first 17 critically ill COVID-19 patients who received ECMO in our center and reported the following key findings. TEBE occurred in 12 patients (71%) consisting of 7 patients (41%) with TEE and 10 patients (59%) with BE. PAE was the most frequent TEE with five cases (29%), and URT was the most common location for BE with eight cases (47%). Mortality between the TEE group (n = 7) and the non-TEE group (n = 10) was comparable (43 vs. 50%, p = 0.581), and higher in the BE group (n = 10) than in the non-BE group (n = 7), but without significance (60 vs. 29%, p = 0.335). From among 85 pre-ECMO variables, only platelet count was significantly lower between the BE group (n = 10) and the non-BE group (n = 7) with 209 (IQR: 145–238) versus 452 (IQR: 240–560), p = 0.007. However, we found other interesting differences that are not statistically significant but mentionable. TEE seemed to be more frequent patients with male gender, atrial fibrillation, less utilization of cytokine absorption, higher pCO2 values, and lower RESP score values. As for BE, we found longer PTT and lower INR values.
From among the TEBE, mainly BE were frequent. Regarding other studies about ECMO in patients suffering from severe ARDS without COVID-19, we also found a high incidence of BE. Two studies report BE in 46% of the patients; the first study included 124 patients and the second study had 214 patients.[11]
[13] We compared it to the 10 patients (59%) with BE of our study and no statistically significant differences could be found using Fisher's exact test. Only few studies reported the incidence of TEBE for COVID-19 patients treated with ECMO; the outcomes are not detailed enough and accurate comparisons with our study cannot be made. Especially the BE are more frequent in our study and this is probably due to the fact that we also reported minor bleedings. Nevertheless, we summarized the results of other studies in the following. Yang et al[17] reported only 3 patients (14%) with bleeding complications, whereas we had 10 patients (59%). In an American study by Kon et al,[20] there was only 1 patient (4%) with clinically significant bleeding complications, 5 patients (19%) with deep vein thrombosis (DVT), and 5 patients (19%) who developed thrombocytopenia; ECMO-associated complications in general were noted in 11 patients (41%). Bleeding complications and general ECMO complications are again much less than those in our study. However, we found four patients (24%) with thrombosis as a comparable outcome. In the study of Jacobs et al,[15] two patients (13%) died due to DIC and one patient (7%) died due to a cerebral bleeding. In another American study by Loforte et al[21] with only four patients in total, one patient (25%) suffered from GIB. In our study, we report also one patient (6%) with GIB.
We observed a successful weaning rate of 53% from ECMO and survival to hospital discharge. This is superior to the rate of other studies.[11]
[15]
[22] In the study of Yang et al,[23] only one patient (17%) survived out of the six patients treated with ECMO.[23] Marullo et al[9] analyzed data of multiple institutions and reported a lower mortality of only 17% from 333 COVID-19 patients treated with ECMO. However, the actual weaning rate is 18% and the remaining patients are still alive on ECMO.[9] Different patient selection and small sample sizes probably contribute to differences in survival rates.
The platelet count was significantly lower in patients with BE and therefore needs to be discussed. As stated previously, we only had one patient who developed a moderate thrombocytopenia between 50 and 100 G/L. Nevertheless, thrombocytopenia is a common phenomenon found in ECMO patients and its underlying mechanisms are multifactorial.[24] A decrease of platelets during the first 7 days is frequently seen and the values of our patient group were not clinically abnormal.[24] We aimed for a platelet count above 80 G/L as recommended in the ELSO guidelines.[25]
We did not find any relevant hemorrhagic or ischemic strokes; however, we found five patients with microscopic bleedings in magnetic resonance imaging (MRI) examinations of the brain. Neurological symptoms did not occur in all cases and it is a common finding in COVID-19 patients without ECMO therapy. Therefore, it cannot be regarded as an ECMO complication in this study.
As a retrospective study, there was potential bias in data acquisition. Because of a small sample size and short study period, doctors could control the gathered data and confirm its correctness as much as possible. Although our pre-ECMO and outcome variables were complete, we have missing values for laboratory parameters during ECMO therapy. Therefore, we could not display time courses of all parameters.
We studied the characteristics and outcomes of the first 17 COVID-19 patients from March and April 2020, but we did not gather enough patients to conduct a high-quality predictor analysis. Therefore, validation of our findings with larger samples is required.
