A comprehensive assessment of the coagulation profile in critically ill COVID-19 patients

A number of alterations in coagulation parameters have been described in patients with COVID-19-associated pneumonia. This is associated with an exaggerated inflammatory response that can lead to severe complications and acute respiratory distress syndrome (ARDS) [1].

Coagulopathy has been described in up to 50% of severe manifestations of COVID-19, fulfilling the criteria for disseminated intravascular coagulation (DIC) in the majority (>70%) of patients. Of note, such coagulopathy was observed substantially more often in severe cases. According to Tang et al. [2] D-dimer levels were four times higher in severely affected patients (median 2.12 μg/mL, IQR 0.77–5.27) compared to non-severe patients (median 0.61 μg/mL, IQR 0.35–1.29). Moreover Guan et al. revealed that high D-dimer levels measured at hospital admission may predict the severity of COVID-19 [3]. Laboratory findings indicate a prothrombotic condition in patients with COVID-19 and consecutively developed venous, arterial and micro-thromboses have been reported [[4], [5], [6], [7], [8]]. However, no report has yet been published demonstrating the full range of coagulation parameter alterations in critically ill COVID-19 patients.

We herein describe the full coagulation profile in four male patients (age 42–77 years) with severe and deleterious COVID-19 associated pneumonia from the University Hospital Frankfurt/Main, Germany. While one patient suffered from severe obesity, the remainder of patients were diagnosed with pre-obesity (median BMI 27.0). The youngest patient (42y) did not have any known pre-existing conditions, while the other patients suffered from arterial hypertension. One of the patients was previously diagnosed with diabetes mellitus and one had a history of urothelial cancer. The included patients were admitted to our intensive care unit (ICU) after presenting with typical symptoms of COVID-19 in the emergency department. A chest CT scan on day one after hospitalization showed typical bilateral multiple ground-glass opacities with peripheral lung and subpleural distributions in all patients. Despite immediate initiation of mechanical ventilation and critical care therapy as put forth by Poston et al. [9], patients rapidly suffered from respiratory failure and a refractory hypoxemia followed by early onset of multi-organ-failure. All patients displayed acute renal failure requiring renal replacement therapy and liver injury consistent to the pathophysiology described recently, suggesting renal tubular cells and liver cells as a target of SARS-CoV-2 [10,11]. Because of refractory hypoxemia the youngest patient received veno-venous extracorporeal membrane oxygenation support.

Therapeutic anticoagulation was administered in all patients using unfractionated heparin for achieving an aPTT between 50 and 70 s. Antithrombin was replaced to maintain a level of 80% or greater. No plasma transfusion or coagulation factor supplementation was performed in any patient. Although the patients were lacking severe pre-existing conditions, none of the patients survived.

The measurements were performed on ACL-TOP coagulation analyzers and using original reagents (Werfen, Barcelona, Spain). The activity of ADAMTS-13 and the antigen of PAI were measured with the ELISA of Haemochrom Diagnostica, Essen, Germany.

The measured coagulation profile included parameters indicative of bleeding and thrombosis, with demographic data and parameters on the severity of the disease shown in Table 1 .

Table 1.

Coagulation profile in critically ill COVID-19 patients.

	Patient 1	Patient 2	Patient 3	Patient 4	Reference range
Age (years)	77	68	42	68
BMI	38.1	26.1	26.1	27.8
Day of illness	4	5	4	4
SOFA score	12	11	12	11
Thromboplastin time (%)	65	88	47	88	70–130
aPTT (s)	37	62	69	51	25–37
Fibrinogen (mg/dL)	741	380	639	707	190–498
Antithrombin-activity (%)	49	54	101	51	80–128
D-dimer (ng/mL)	3315	12,639	13,221	4159	<500
Platelets (103/μL)	391	172	251	174	146–328
Factor II (%)	61.7	47.6	51.4	65.7	75–129
Factor V (%)	98.2	96	121.4	163.4	80–148
Factor VII (%)	57.1	73.7	24.2	68.4	48–148
Factor VIII (%)	339.6	73.7	261.8	339.5	48–139
Factor IX (%)	150.8	106.1	130	97	68–133
Factor XI (%)	64	51	60	82	69–144
Factor XII (%)	34.8	29.8	32.1	47.4	66–146
Factor XIII (%)	120.9	39	58.8	46.3	70–155
v. Willebrand antigen (%)	>600	>600	>600	536	60–150
Protein C-activity (%)	53	59	87	65	>72
Free protein S-antigen (%)	55	48	45	37	68–116
DRVVT screen (s)	53.8	42.6	54.7	49.6	28.4–45.8
DRVVT ratio	1.19	1.20	1.19	1.22	0.93–1.40
Lupus sensitive PTT (s)	35.9	35.4	58.1	33.6	23.1–38.4
Anti-cardiolipin IgM (U/mL)	2.7	10.3	<1.0	1.7	<20
Anti-cardiolipin IgG (U/mL)	6.7	3.9	4	6.4	<20
Anti-ß2-glykoprotein IgM (U/mL)	1.7	1.1	<1.1	<1.1	<20
Anti-ß2-glykoprotein IgG (U/mL)	<6.4	7.7	<6.4	<6.4	<20
ADAMTS-13 (%) activity	36	19	36	28	40–130
PAI-Ag (ng/mL)	36.2	>62.4	10.9	>62.4	7–43
Factor V-mutation	Wild type	Wild type	Wild type	Wild type
Factor II-mutation	Wild type	Wild type	Wild type	Wild type

BMI: body-mass-index; SOFA score: sepsis-related organ failure assessment score; aPTT: activated partial thromboplastin time; DRVVT: diluted-Russel-Viper-Venom-Test; PAI: plasminogen activator inhibitor; AT levels are pre-substitution values.

Consistent with published data, our results confirmed a substantial increase of D-dimers and fibrinogen levels in all patients, reflecting a response of a systemic inflammatory reaction leading to the activation of blood coagulation.

Furthermore, our results revealed an increase in von Willebrand Factor (VWF) and Factor (F) VIII. These alterations may mirror the systemic endothelial damage recently described in COVID-19 by Varga et al. [12], which has previously been reported for classic ARDS, sepsis and various inflammatory diseases [13]. A direct link between the hemostatic function of VWF and inflammation has already been described earlier [14] and the massive release of VWF may be considered as an indicator of vascular dysfunction [15]. Besides its platelet activation properties, VWF promotes leucocyte adhesion to endothelial cells [16] and large or ultra-large VWF (ULVWF) multimers activate the complement cascade [17]. Further, ADAMTS-13 activity was reduced in all patients, which is a common finding in critically ill patients [18]. Of note, the reduction of ADAMTS-13 activity was not observed to reach an activity of below 10%, indicating an absent thrombotic thrombocytopenic purpura (TTP) in the studied patients [19]. Considering the known linear relationship between ADAMTS-13 and VWF, this might explain the observed decrease.

Interestingly, the changes in the various coagulation factors revealed a different and more complex picture. While the activity of FXII and FXIII were reduced in most patients, most likely in response to increased levels of D-dimer and hyperfibrinolysis, the activity of FII and FVII were reduced in two patients, resulting in a prolonged thromboplastin time. A speculative but reasonable explanation for the observed reduction of FII and FVII may result from several underlying conditions such as liver failure, vitamin K deficiency or treatment with antibiotics. The other coagulation factors, FV, FVIII, FIX, were normal or elevated in all patients, suggesting that patients did not present all criteria of DIC in consuming coagulation factors.

In particular, no hereditary thrombophilia was diagnosed and various measurements were negative in regard to antiphospholipid antibodies resulting in no evidence of antiphospholipid syndrome.

A further observation was the reduced levels of protein C and protein S: It has been repeatedly shown that these parameters are associated with a poor outcome of sepsis [20]. A strongly increased plasminogen activator inhibitor 1 (PAI-1) level above the upper limit of detection was observed in 50% of the patients, indicating an increased risk for thromboembolic events.

In conclusion, the coagulation profile in critically ill COVID-19 patients showed a substantial activation of coagulation and fibrinolysis with highly increased levels of D-dimer and VWF as potential markers of endothelial dysfunction. No clinical signs and no laboratory alterations indicative for bleeding or findings associated with DIC (e.g. altered platelet counts or fibrinogen levels) have been detected.

Author contributions

EHA, KZ and WM analyzed the data and wrote the text.

Declaration of competing interest

There are no conflicts of interests.