Journal Logo

Online Clinical Investigation

A Comparison of Thrombosis and Hemorrhage Rates in Patients With Severe Respiratory Failure Due to Coronavirus Disease 2019 and Influenza Requiring Extracorporeal Membrane Oxygenation

Doyle, Andrew J. MBBS1; Hunt, Beverley J. MD1; Sanderson, Barnaby BSc2; Zhang, Joseph MBBS2; Mak, Sze M. MBBS3; Benedetti, Guilia MBBS3; Breen, Karen A. MD1; Camporota, Luigi PhD2; Barrett, Nicholas A. MBBS2,4; Retter, Andrew MBBS2

Author Information
doi: 10.1097/CCM.0000000000004971


Hospitalization to treat patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (coronavirus disease 2019 [COVID-19]) pneumonia has been associated with high rates of thrombosis, particularly in the critical care, despite the use of pharmacological thromboprophylaxis. The composite rates of thromboembolism in those hospitalized with COVID-19 pneumonia without extracorporeal membrane oxygenation (ECMO) are between 16% and 42% with predominantly pulmonary events (1–4). Rates of thrombosis were notably higher if screening for asymptomatic deep vein thrombosis (DVT) was performed (4).

Patients with COVID-19 pneumonia who develop acute respiratory distress syndrome (ARDS) appear to have two thrombotic pathologies in their lungs: pulmonary emboli (PE) and in situ thrombosis, known as immunothrombosis. Immunothrombosis manifests as arteriolar and microvascular in situ thrombosis as a consequence of intense inflammation in the absence of embolization (1,2). Such pulmonary microthrombi are a recognized histological feature of ARDS of various etiologies (1,2,5). At present, it is unclear whether immunothrombosis and other thromboses are present to a greater extent in COVID-19 when compared with other viral pneumonias. Indeed, high rates of thrombotic events have also been described in critically ill patients with influenza A (H1N1) infections previously (6–8). Although a postmortem study of seven patients with COVID-19 pneumonitis demonstrated a nine-fold increase in the presence of pulmonary microthrombi in comparison to seven patients with influenza, there are limited data indicating whether rates of thrombosis are different compared with influenza and other causes of severe ARDS (9).

Entry of the SARS-CoV-2 virus into cells is principally via the angiotensin-converting enzyme 2 receptor, which is present on a number of cells, including pulmonary pneumocytes, endothelium, macrophages, and lymphocytes (10). Patients with severe COVID-19 requiring mechanical ventilation have been shown to have a prothrombotic state with high levels of fibrinogen, factor VIII, von Willebrand factor, and other acute phase proteins (1,11). d-dimer levels are also extremely high in those with moderate and severe COVID-19, and early data showed this to be associated with an increased level of thrombotic events and associated with mortality (4). d-dimers are produced when fibrin is degraded by plasmin. In plasma, the activation of plasmin is by tissue plasminogen activator. In acute lung injury, pneumocytes release urokinase-type plasminogen activator providing an alternative mechanism of d-dimer production (1,2).

ECMO is used in patients when oxygenation can no longer be safely maintained with conventional mechanical ventilation, and its use has been described in COVID-19 (12,13). ECMO causes a prothrombotic state with activation of procoagulant factors on the large synthetic surface area of the oxygenator; this includes platelet activation and red cell fragmentation leading to the release of prothrombotic molecules such as free heme (14). IV unfractionated heparin (UFH) at therapeutic levels are used to reduce membrane occlusion by microclot. The use of anticoagulation can increase the risk of major hemorrhage, particularly with an acquired coagulopathy due to their underlying critical illness. However, anticoagulation use is still necessary to prevent and treat thrombotic complications during the use of ECMO.

We compared the rate of thromboembolic and hemorrhagic events in patients receiving venovenous ECMO for acute hypoxic respiratory failure (AHRF) during the COVID-19 pandemic to a matched cohort of patients with influenza also requiring venovenous ECMO. We specifically examined the frequency of pulmonary thromboembolism and catheter-associated DVT following decannulation due to their high rates in the critical care setting. Furthermore, we studied the rate of major hemorrhage prior to and during ECMO to assess the potential safety and intensity of anticoagulation during ECMO in patients with COVID-19 pandemic when compared with its use in those with influenza.


Subjects and Study Design

This retrospective observational study was registered as a service evaluation at Guy’s and St Thomas’ NHS Foundation Trust, London (Reference Number—10802) and ethical approval was not required in accordance with U.K. ethical guidelines (, Accessed May 31, 2020). Patients with severe COVID-19 infection requiring ECMO between March 1, 2020, and May 31, 2020, were compared with a historic cohort of patients with influenza pneumonia and ARDS treated between June 1, 2012, and May 31, 2020. All patients greater than 18 years old with COVID-19 and influenza were considered eligible for review. The data were extracted from a prospectively maintained electronic registry. The viral diagnosis was confirmed from polymerase chain reactions obtained by nasopharyngeal swab or bronchoalveolar lavage.

Clinical Details, Imaging, and Assessment of Bleeding

Clinical data were collected from electronic records on Phillips ICIP IntelliSpace CCA (Amsterdam, The Netherlands) and iSOFT Clinical Manager System (DXC Technology, Tysons, VA). The electronic patients’ records for identified patients were reviewed for clinical details including age, gender, Sequential Organ Failure Assessment (SOFA) and Acute Physiology and Chronic Health Evaluation (APACHE) II scores, radiologically identified thrombosis, and major bleeding events (intracranial hemorrhage or bleeding with ≥ 20 mL/kg of red cell concentrate within 24 hr) in accordance with standardized Extracorporeal Life Support Organization (ELSO) definitions (15).

All patients requiring ECMO at our institution underwent a whole-body CT as soon as possible after cannulation. This normally occurs within 4–6 hours of cannulation. The imaging was performed with either a Somatom Force 384 (2 × 192 sections; Siemens Healthineer, Erlangen, Germany) or Brilliance iCT (256 sections; Philips Healthcare, Amsterdam, The Netherlands). CT pulmonary angiography (CTPA) is performed by bolus trigger at 110 Hounsfield units. Ultrasound Doppler scanning was performed within 24–72 hours of decannulation of the cannulas used for ECMO at both catheter insertion sites to assess for DVT (16). Additional imaging during ECMO was performed if there was clinical concern for bleeding or thrombosis.

ECMO Details and Anticoagulation Practice

All patients reviewed in this study received venovenous ECMO. The ECMO pump used was the Cardiohelp system (Getinge, Gothenburg, Sweden) and BE HLS 7.0 oxygenator (Getinge). Cannulation was performed using bifemoral cannulae as previously described (Medtronic, Dublin, Ireland) (17).

For anticoagulation during ECMO, IV UFH infusion was used with 50 international units/kg as a bolus at the time of cannulation with 2,500 units in the priming fluid for the circuit in keeping with ELSO guidelines unless contraindicated (18). Prior to November 1, 2019, subsequent UFH infusions were monitored by activated partial thromboplastin time ratio (APTTr) with a target level of 1.5–2.0 if evidence of no thrombosis and hemorrhage and 2.0–2.5 if a thrombotic event. From November 1, 2019, to present, anti-Xa activity was used for monitoring targeting levels of 0.3–0.7 U/mL if no evidence of thrombosis or hemorrhage were present and 0.6–1.0 unit/mL if there was a thrombotic event (supplementary material, UFH infusions were not given if evidence of major bleeding. Patients were also prescribed pharmacological thromboprophylaxis with subcutaneous low molecular weight heparin unless contraindicated in critical care prior to ECMO use (19).

Laboratory Testing of Hemostatic Parameter

Testing for APTT, prothrombin time, d-dimer, and Clauss fibrinogen were performed using Sysmex CS2100i (Norderstedt, Germany) as per manufacturer protocols. Full blood count testing was performed using DxH 900 Hematology (Beckman Coulter, Brea, CA). Blood samples were taken after initiation of ECMO and following heparin boluses. The tests were performed within the first 4 hours of ECMO initiation.

Outcome Variables

The clinical outcomes were the diagnosis of pulmonary thrombosis/embolism and the presence of catheter-associated DVT after decannulation from ECMO in all patients confirmed radiologically. Additional thrombotic events were also recorded in patients with COVID-19 infection. Major hemorrhage was recorded preceding and during the use of ECMO (15). The index date was the date on which venovenous ECMO was initiated. Data were censored on June 26, 2020.

Statistical Considerations

Statistical analysis was performed using SPSS Version 26.0 (IBM, Armonk, NY). Null hypothesis significance testing was employed to examine variable distributions between the two groups. Data were analyzed for distribution spread. Mann-Whitney U test was performed on nonparametric data and chi-square test for frequencies of categorical variables were performed. Fisher exact test if group frequencies were less than 5. An α-significance level of 0.05 was used for all statistical tests.


Patients Demographics and Outcomes

A total of 131 patients were identified—51 had COVID-19 and 80 had influenza pneumonia requiring venovenous ECMO due to AHRF. Within the influenza group, there were 43 cases of H1N1 influenza, 31 of other influenza A infections, and six cases of influenza B. The clinical characteristics of the two cohorts are described in Table 1. All patients with influenza had either been discharged from critical care or died (83% survival from intensive treatment unit [ITU]). One patient with COVID-19 still required ECMO at date of review (74% survival from ITU).

TABLE 1. - Clinical and Hemostatic Characteristics of Patients With Coronavirus Disease 2019 and Influenza at Initiation of Venovenous Extracorporeal Membrane Oxygenation Showing Median and Interquartile Ranges
Clinical Feature Influenza (n = 80) Coronavirus Disease 2019 (n = 51) p
Median age (yr) 48.3 (38.7–57.3) 46.1 (35.6–53.2) 0.161
Gender (male), n (%) 47 (59) 38 (74.5) 0.065
Median Sequential Organ Failure Assessment score 9 (1–17) 6 (3–16) 0.081
Median Acute Physiology and Chronic Health Evaluation II score 17 (10–27) 14 (5–22) < 0.001
Median days from symptom onset to ECMO 8 (5–13) 14.0 (11–17) < 0.001
Survival from intensive treatment unit, n (%) 66 (82.5) 37 (74.0) 0.272
Median duration of ECMO 13 (10–19) 13 (8–21) 0.450
Median admission activated partial thromboplastin time ratio (normal range 0.8–1.2) 1.8 (1.2–3.1) 1.3 (1.1–1.7) 0.006
Median admission prothrombin time ration (normal range 0.8–1.2) 1.1 (1.0–1.2) 1.1 (1.0–1.2) 0.902
Mean admission platelets (normal range 150–400 × 109/L) 148 (104–205) 238 (164–317) < 0.001
Mean admission fibrinogen (normal range 1.7–3.9 g/L) 5.0 (3.4–6.5) 7.1 (5.4–8.8) < 0.001
Mean admission d-dimer (normal range 0–0.55 mg/L fibrinogen equivalent unit) 9.0 (3.5–18.9) 9.1 (5.2–39.9) 0.083
Mean admission C-reactive protein (normal range 0–4 mg/L) 226 (146–326) 320 (225–362) 0.006
Presence of overt disseminated intravascular coagulation (≥ 5), n (%) 5 (6.8) 0 0.007
ECMO = extracorporeal membrane oxygenation.

The clinical and hemostatic characteristics of the patient groups are shown in Table 1. Seventy-five percent of patients with COVID-19 ARDS requiring ECMO were men. The median time of symptom onset to ECMO initiation was 14 days in patients with COVID-19 in comparison to 8 days in those with influenza. The SOFA and APACHE II scores were notably higher in those with influenza requiring ECMO, but both groups had similar duration of ECMO.

Those with COVID-19 had a lower APTTr than influenza. Thrombocytopenia was less frequent in COVID-19 with a mean platelet count of 238 × 109/L in comparison to 148 × 109/L with influenza. There were no cases of disseminated intravascular coagulation (DIC) seen in the COVID-19 patients in accordance to the International Society on Thrombosis and Haemostasis overt DIC criteria (20). There appeared to be a greater acute phase response in COVID-19 patients with a mean C-reactive protein (CRP) and fibrinogen levels significantly higher than those with influenza. d-dimer levels were similarly very high in both groups at 9 mg/L fibrinogen equivalent units (FEU).

Thromboembolic Events

Whole-body CT imaging at initiation of ECMO was performed in all patients with COVID-19 and influenza pneumonias. The frequency of key thromboembolic and hemorrhagic complications associated with use of ECMO are shown in Table 2. Arterial filling defects were present on CTPA in 37% of patients with COVID-19 in comparison to 8% with influenza immediately after initiating ECMO (p = 0.0001). Fourteen of 19 (74%) of identified thrombi in COVID-19 and four of six (67%) in influenza were segmental and subsegmental filling defects.

TABLE 2. - Frequency of Hemostatic Complications at Initiation and Termination of Extracorporeal Membrane Oxygenation in Patients With Coronavirus Disease 2019 and Influenza
Type of Complication Influenza (n = 80) Coronavirus Disease 2019 (n = 51) Significance (p)
All pulmonary artery filling defects seen at initiation of ECMO 6 (8%) 19 (37%) 0.0001
2—central 2—central
0—lobar 3—lobar
4—segmental 5—segmental
0—subsegmental 9—subsegmental
Pulmonary immunothrombosis at initiation of ECMOa 4 (5%) 14 (27%) 0.0005
Intracranial hemorrhage present at initiation of ECMO 11 (14%) 8 (16%) 0.759
9—subarachnoid 7—subarachnoid
2—intracerebral 1—subdural
Deep vein thrombosis seen on Doppler ultrasound at decannulation (n = 99) 15 (25%) 20 (53%) 0.005
ECMO = extracorporeal membrane oxygenation.
aPulmonary immunothrombosis considered as image-proven segmental and subsegmental filling defects.

Six thromboses at other sites were seen in six of 80 patients (8%) with influenza at ECMO initiation (three splenic infarcts, two DVT, and one intestinal infarct) and seven thromboses in six of 51 patients (12%) with COVID-19 (three internal jugular vein thrombosis, two ischemic stroke, one DVT, and one splenic infarct).

During ECMO, 27 of 51 patients (53%) with COVID-19 had further CT imaging. These showed radiological features of PE in three cases: one showed a new subsegmental thrombus, the second showed small extension of a previous thrombus, and the third showed stable thrombus size but progressive surrounding infarction. In the latter two, they had preceding subtherapeutic anti-Xa levels from ECMO initiation for at least 2 days before their subsequent imaging.

Ultrasound Dopplers were performed in 99 patients (38 patients with COVID-19 and 61 with influenza) who survived to decannulation from ECMO. Fifteen of 61 patients (25%) with influenza and 20 of 38 (53%) with COVID-19 had catheter-associated DVT. Those who developed catheter-associated DVT were evaluated for preceding thrombotic or hemorrhagic events. Four of 20 patients (20%) with COVID-19 who had catheter-associated DVT had preceding major bleeding. Therefore, these had significant interruptions of anticoagulation with UFH infusions, whereas 16 of 20 (80%) developed catheter-associated DVT while receiving UFH anticoagulation. These findings were similar to patients with influenza—three of 15 (20%) had preceding major bleeding (p = 1.00). Eight of 20 (40%) with COVID-19 who developed post-decannulation DVT has a preceding thrombotic event in comparison to two of 15 (13%) with influenza (p = 0.067).

Hemorrhagic Events

Intracranial hemorrhage (ICH) was the most frequent major hemorrhage occurring in eight of 51 patients (16%) with COVID-19 and 11 of 80 (14%) with influenza showing no significant statistical difference in rates. Subarachnoid hemorrhage was the most common bleeding type. One non-ICH major hemorrhage at initiation of ECMO was identified in a patient with COVID-19 (a hemothorax) and none with influenza. Six major bleeding events were identified during ECMO with COVID-19—three intracerebral, two retroperitoneal, and one pleural.

Hemostatic Parameters

We compared hemostatic parameters for patients with pulmonary thromboembolism on single-energy CTPA at the time of commencing ECMO to those without (Table 3). d-dimer levels in those with pulmonary thromboembolism in COVID-19 were higher than those without (median 33.3 vs 7.2 mg/L FEU; p = 0.019). Fibrinogen levels, CRP, and platelet counts were not different between those with venous thromboembolism and without in both COVID-19 and influenza.

TABLE 3. - Comparison of Coagulation Parameters and C-Reactive Protein in Patients With and Without Filling Defects on CT Pulmonary Angiography at Initiation of Extracorporeal Membrane Oxygenation With Coronavirus Disease 2019 and Influenza
Coagulation Parameter Coronavirus Disease 2019 Influenza
Pulmonary Thromboembolism (n = 17) No Pulmonary Thromboembolism (n = 34) Pulmonary Thromboembolism (n = 6) No Pulmonary Thromboembolism (n = 74)
Platelets (×109/L), median (IQR) 252 (169–344); p = 0.50 223 (170–310) 183 (173–221); p = 0.17 145 (104–199)
PTr, median (IQR) 1.1 (1.0–1.2); p = 0.69 1.1 (1.0–1.1) 1.1 (1.0–1.1); 0.86 1.1 (1.0–1.2)
d-dimer (mg/L FEU), median (IQR) 33.3 (7.0–69.1); p = 0.019 7.2 (3.2–28.3) 16.2 (9.5–21.5); p = 0.10 6.9 (3.5–18.1)
Fibrinogen (g/L), median (IQR) 7.4 (5.3–8.6); p = 0.71 6.9 (5.5–9.0) 4.4 (3.8–6.6); p = 0.99 5.0 (3.4–6.4)
C-reactive protein (mg/L), median (IQR) 315 (225–355); p = 0.73 323 (243–369) 310 (138–381); p = 0.50 222 (149–321)
IQR = interquartile range.


Rates of image-proven thromboembolism were significantly higher in patients with ARDS requiring ECMO in severe COVID-19 infection when compared with those with influenza. This difference was largely due to higher rates of segmental and subsegmental changes on CTPA, which we interpreted as immunothrombosis but also higher rates of venous thromboembolism. Similarly, high rates of PE at 19% were described by Schmidt et al (12) in a cohort of 83 patients requiring ECMO when compared with the Extracorporeal Membrane Oxygenation for Severe Acute Respiratory Distress Syndrome; Factor VIII - FVIII (EOLIA) study, in which PE was not recognized (21). Hemorrhagic rates were similar between both groups in our study. ARDS due to both types of viral pneumonia were associated with high levels of d-dimer and fibrinogen and relatively preserved platelet counts, but those with COVID-19 had significantly higher levels of fibrinogen, CRP, and platelet count, consistent with a more profound acute phase response.

Our data support the growing body of evidence that COVID-19 is associated with both increased “macrovascular” thrombosis (DVT and PE) and “microvascular” immunothrombosis (22). We hypothesize, like others, that immunothrombosis is one of the mechanisms underlying the ARDS phenotype of patients with COVID-19, which has been described in other settings of ARDS (5,23).

Hartley et al (24) previously described levels of pulmonary thromboembolism of 9–10% in a heterogeneous cohort of conditions causing severe AHRF prior to ECMO. In this cohort, 58–64% had an infective cause of ARDS. These rates are similar to the levels described in this influenza cohort but lower than that seen in the COVID-19 cohort. Minet et al (25) described rates of PE at 19% in mechanically ventilated patients in intensive care, with predominantly asymptomatic and small volume thrombi. However, the overall anticoagulation use in this cohort was lower than at our center.

Our data support the evolving hypothesis that immunothrombosis in the lungs is a significant component in the pathogenesis of severe AHRF in both influenza and COVID-19, although more marked in COVID-19. Agarwal et al (26) described a rate of PE of 36% in those with H1N1 Influenza who were mechanically ventilated and 0% in those who were not, suggesting it may be a feature of more severe lung inflammation. Walters et al (27) have also described extensive pulmonary fibrin deposition in a murine model of the 1918 influenza infection with superimposed Streptococcal infection. High rates of both macrovascular and microvascular thrombosis may partly relate to a more prothrombotic state in COVID-19 including endothelial activation, its marked inflammatory state and hypoxia leading to hypercoagulability, although further studies are necessary to assess this (11,28). It is unclear at which stage in the natural history of COVID-19 infection that immunothrombosis develops and whether it plays a role in disease progression (21). Ongoing randomized control studies are evaluating the role of anticoagulation and antiplatelet agents in critically ill and noncritically ill patients with COVID-19 pneumonia.

We found higher numbers of catheter-associated DVT at sites of decannulation from ECMO with COVID-19 than influenza. They were seen frequently in those with a preceding thrombosis during ECMO as opposed to those who have interruption of anticoagulation due to bleeding. We recognize that catheter-associated DVT relates particularly to reduced blood flow seen when a synthetic line is present within a vein, but our results suggest that the prothrombotic phenotype of COVID-19 may have contributed too (11,28). Indeed, our rates of catheter-associated DVT are higher in COVID-19 than those previously reported in ECMO (29,30).

Intracranial hemorrhage rates were similar at the initiation of ECMO in both COVID-19 and influenza with AHRF. Rates of mortality are high due to ICH during the use of ECMO (31). In previously reported data, we demonstrated that CT screening at the time of cannulation for ECMO for ICH showed similar rates of mortality if detected early to those without ICH (32). We identified that severity of hypoxia, thrombocytopenia, and renal failure were predictors of intracranial hemorrhage prior to commencing venovenous ECMO (32). These parameters are not significantly different between the two types of viral pneumonia and potentially explain the similar rates of ICH. With increased rates of thrombotic events seen in patients with severe AHRF due to COVID-19, ICH should still be considered prior to the use of therapeutic anticoagulation.

In reviewing hemostatic parameters at the initiation of ECMO, fibrinogen, and d-dimer levels were greatly elevated in both groups but with higher fibrinogen levels during severe COVID-19 infection. Similarly, CRP levels were higher in patients with COVID-19 in keeping with a significant acute phase response. Elevated APTTr in influenza cohort may be explained by underlying DIC and/or refinement in heparin use at the time of ECMO initiation over the 8-year period that was reviewed. Heparin resistance has been described in COVID-19 (33). In addition, elevated FVIII levels in COVID-19 infection may have affected this result (11). There was no significant difference in d-dimer levels between the two groups, a similar finding to Yin et al (34) who compared d-dimer levels in COVID-19 patients with d-dimer levels in ARDS from viral pneumonias in the previous year at the same institution.

d-dimer levels were greater in those patients who had pulmonary thromboembolism compared with those without in both groups of viral pneumonia-related ARDS and appeared to be discriminatory in those with COVID-19. Furthermore, these findings are in contrast to the initial early reports on coagulation changes associated with COVID-19 suggesting the presence of DIC, particularly in this cohort of critically unwell patients in which one would expect this to be more prevalent (35).

A key strength of this study was the universal use of imaging performed as the standard of care in our center during the use of venovenous ECMO for AHRF. This allows a better standardization in the imaging rates, which is a key problem with other similar studies with COVID-19 in critical care (1–4). This has been a limitation in establishing the total frequency of thromboses with the infection up until now.

We have assessed the rates of thrombosis and hemorrhage to influenza as a comparator group to COVID-19 with a similar management protocol. In general, the underlying etiologies for those requiring venovenous ECMO are heterogeneous, which is reflected in previous studies looking at the frequency of thrombosis (24,29). Our current cohorts demonstrate homogeneity with two underlying conditions being compared. The data also suggest that the underlying pulmonary pathology in ARDS may be a consideration for assessing the potential risk of thrombotic complications when using ECMO. However, further evaluation of this is required in other conditions.

Limitations of this study are its retrospective, single-center observational design, and in particularly with regards to the historical group with influenza, leading to the possible introduction of bias. Due to the nature of the pandemic and importance of these data in the practical management of patients, we lack follow-up to 90 days in all patients with COVID-19. We would need this to truly define the rates of hospital-associated thrombosis, the definition being venous thromboembolism occurring during admission and for up to 90 days post discharge. Due to the nature of our retrieval service for ECMO, we were unable to obtain follow-up data once the patients were returned to their referring hospital.

It is also important to note that the use of anti-Xa levels allows increased time in therapeutic range for UFH when compared with the APTTr (36). There was a change in UFH monitoring from APTTr to anti-Xa over the review period of the patients with COVID-19. Anti-Xa levels were used to monitor all patients with COVID-19 and minimal numbers with influenza. Despite this potential advantage of anti-Xa monitoring, there were still higher rates of thrombosis in patients with COVID-19. We were also unable to ascertain pharmacological thromboprophylaxis in some patients prior to admission to our center.

Question also remains as to whether asymptomatic thrombosis occurs during ECMO use. Our approach is to use whole-body CT scanning at the time of ECMO initiation and further imaging if indicated based upon a change in clinical status. As we do not employ CT scanning routinely at the end of ECMO, we were unable to fully evaluate if further PE or immunothrombosis develop or if resolution occurs during the use of ECMO in all patients.


We believe that clinicians should be aware of the increased risk of thrombotic complications in patients with severe AHRF due to COVID-19 requiring ECMO when compared with ARDS due to other viruses and noninfectious diseases. Radiological features suggestive of immunothrombosis are seen in both types of viral infections but are more prevalent in patients with severe COVID-19 infection suggestive of their possible role in disease progression.

We suggest whole-body CT for screening of thrombosis and hemorrhage should be performed, if available, at the initiation of ECMO, ultrasound Dopplers at the completion of ECMO for screening of catheter-associated DVT and additionally if there is a change in clinical status to identify these complications. This will assist appropriate anticoagulation use prior to the development of major hemorrhage or progressive thrombotic events.


We would like to acknowledge the contributions of care from the Critical Care department at Guy’s and St Thomas NHS Foundation Trust.


1. Desborough MJR, Doyle AJ, Griffiths A, et al. Image-proven thromboembolism in patients with severe COVID-19 in a tertiary critical care unit in the United Kingdom. Thromb Res. 2020; 193:1–4
2. Klok FA, Kruip MJHA, van der Meer NJM, et al. Confirmation of the high cumulative incidence of thrombotic complications in critically ill ICU patients with COVID-19: An updated analysis. Thromb Res. 2020; 191:148–150
3. Thomas W, Varley J, Johnston A, et al. Thrombotic complications of patients admitted to intensive care with COVID-19 at a teaching hospital in the United Kingdom. Thromb Res. 2020; 191:76–77
4. Nopp S, Moik F, Jilma B, et al. Risk of venous thromboembolism in patients with COVID-19: A systematic review and meta-analysis. Res Pract Thromb Haemost. 2020; 4:1178–1191
5. Tomashefski JF Jr, Davies P, Boggis C, et al. The pulmonary vascular lesions of the adult respiratory distress syndrome. Am J Pathol. 1983; 112:112–126
6. Hariri L, Hardin CC. COVID-19, angiogenesis, and ARDS endotypes. N Engl J Med. 2020; 383:182–183
7. Avnon LS, Munteanu D, Smoliakov A, et al. Thromboembolic events in patients with severe pandemic influenza A/H1N1. Eur J Intern Med. 2015; 26:596–598
8. Obi AT, Tignanelli CJ, Jacobs BN, et al. Empirical systemic anticoagulation is associated with decreased venous thromboembolism in critically ill influenza A H1N1 acute respiratory distress syndrome patients. J Vasc Surg Venous Lymphat Disord. 2019; 7:317–324
9. Ackermann M, Verleden SE, Kuehnel M, et al. Pulmonary vascular endothelialitis, thrombosis, and angiogenesis in COVID-19. N Engl J Med. 2020; 383:120–128
10. Bao L, Deng W, Huang B, et al. The pathogenicity of SARS-CoV-2 in hACE2 transgenic mice. Nature. 2020; 583:830–833
11. Masi P, Hékimian G, Lejeune M, et al. Systemic inflammatory response syndrome is a major contributor to COVID-19-associated coagulopathy: Insights from a prospective, single-center cohort study. Circulation. 2020; 142:611–614
12. Schmidt M, Hajage D, Lebreton G, et al.; Groupe de Recherche Clinique en REanimation et Soins intensifs du Patient en Insuffisance Respiratoire aiguE (GRC-RESPIRE) Sorbonne Université; Paris-Sorbonne ECMO-COVID investigators. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome associated with COVID-19: A retrospective cohort study. Lancet Respir Med. 2020; 8:1121–1131
13. Mustafa AK, Alexander PJ, Joshi DJ, et al. Extracorporeal membrane oxygenation for patients with COVID-19 in severe respiratory failure. JAMA Surg. 2020; 155:990–992
14. Doyle AJ, Hunt BJ. Current understanding of how extracorporeal membrane oxygenators activate haemostasis and other blood components. Front Med (Lausanne). 2018; 5:352
15. Extracorporeal Life Support Organization. ELSO Anticoagulation Guideline (2014). 2014. Available at: Accessed July 17, 2020
16. NICE Guideline [NG 158]. Venous Thromboembolic Diseases: Diagnosis, Management and Thrombophilia Testing. London, United Kingdom, National Institute for Health and Care Excellence. 2020. Available at: Accessed July 17, 2020
17. Burns J, Cooper E, Salt G, et al. Retrospective observational review of percutaneous cannulation for extracorporeal membrane oxygenation. ASAIO J. 2016; 62:325–328
18. Hunt BJ, Levi M. Re The source of elevated plasma D-dimer levels in COVID-19 infection. Br J Haematol. 2020; 190:e133–e134
19. NICE Guideline [NG89]. Venous Thromboembolism in Over 16s; Reducing the Risk of Hospital-Acquired Deep Vein Thrombosis or Pulmonary Embolism. London, United Kingdom, National Institute for Health and Care Excellence. 2018. Available at: Accessed July 17, 2020
20. Taylor FB Jr, Toh CH, Hoots WK, et al.; Scientific Subcommittee on Disseminated Intravascular Coagulation (DIC) of the International Society on Thrombosis and Haemostasis (ISTH). Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost. 2001; 86:1327–1330
21. Combes A, Hajage D, Capellier G, et al.; EOLIA Trial Group, REVA, and ECMONet. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome. N Engl J Med. 2018; 378:1965–1975
22. Iba T, Levy JH, Levi M, et al. Coagulopathy of coronavirus disease 2019. Crit Care Med. 2020; 48:1358–1364
23. Oudit GY, Pfeffer MA. Plasma angiotensin-converting enzyme 2: Novel biomarker in heart failure with implications for COVID-19. Eur Heart J. 2020; 41:1818–1820
24. Hartley EL, Singh N, Barrett N, et al. Screening pulmonary angiogram and the effect on anticoagulation strategies in severe respiratory failure patients on venovenous extracorporeal membrane oxygenation. J Thromb Haemost. 2020; 18:217–221
25. Minet C, Lugosi M, Savoye PY, et al. Pulmonary embolism in mechanically ventilated patients requiring computed tomography: Prevalence, risk factors, and outcome. Crit Care Med. 2012; 40:3202–3208
26. Agarwal PP, Cinti S, Kazerooni EA. Chest radiographic and CT findings in novel swine-origin influenza A (H1N1) virus (S-OIV) infection. AJR Am J Roentgenol. 2009; 193:1488–1493
27. Walters KA, D’Agnillo F, Sheng ZM, et al. 1918 pandemic influenza virus and Streptococcus pneumoniae co-infection results in activation of coagulation and widespread pulmonary thrombosis in mice and humans. J Pathol. 2016; 238:85–97
28. Nougier C, Benoit R, Simon M, et al. Hypofibrinolytic state and high thrombin generation may play a major role in SARS-CoV2 associated thrombosis. J Thromb Haemost. 2020; 18:2215–2219
29. Cooper E, Burns J, Retter A, et al. Prevalence of venous thrombosis following venovenous extracorporeal membrane oxygenation in patients with severe respiratory failure. Crit Care Med. 2015; 43:e581–e584
30. Menaker J, Tabatabai A, Rector R, et al. Incidence of cannula-associated deep vein thrombosis after veno-venous extracorporeal membrane oxygenation. ASAIO J. 2017; 63:588–591
31. Fletcher-Sandersjöö A, Thelin EP, Bartek J Jr, et al. Incidence, outcome, and predictors of intracranial hemorrhage in adult patients on extracorporeal membrane oxygenation: A systematic and narrative review. Front Neurol. 2018; 9:548
32. Lockie CJA, Gillon SA, Barrett NA, et al. Severe respiratory failure, extracorporeal membrane oxygenation, and intracranial hemorrhage. Crit Care Med. 2017; 45:1642–1649
33. Dutt T, Simcox D, Downey C, et al. Thromboprophylaxis in COVID-19: Anti-FXa-the missing factor? Am J Respir Crit Care Med. 2020; 202:455–457
34. Yin S, Huang M, Li D, et al. Difference of coagulation features between severe pneumonia induced by SARS-CoV2 and non-SARS-CoV2. J Thromb Thrombolysis. 2020 Apr 3. [online ahead of print]
35. Tang N, Li D, Wang X, et al. Abnormal coagulation parameters are associated with poor prognosis in patients with novel coronavirus pneumonia. J Thromb Haemost. 2020; 18:844–847
36. Arnouk S, Altshuler D, Lewis TC, et al. Evaluation of anti-Xa and activated partial thromboplastin time monitoring of heparin in adult patients receiving extracorporeal membrane oxygenation support. ASAIO J. 2020; 66:300–306

anticoagulation; coronavirus disease 2019; extracorporeal membrane oxygenation; hemorrhage; influenza; thrombosis

Supplemental Digital Content

Copyright © 2021 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.