key: cord-0956792-qp76h6he
authors: Martín‐Rojas, Reyes María; Pérez‐Rus, Gloria; Delgado‐Pinos, Valeria Estefania; Domingo‐González, Amalia; Regalado‐Artamendi, Isabel; Alba‐Urdiales, Natalia; Demelo‐Rodríguez, Pablo; Monsalvo, Silvia; Rodríguez‐Macías, Gabriela; Ballesteros, Monica; Osorio‐Prendes, Santiago; Díez‐Martín, José Luis; Pascual, Cristina
title: COVID‐19 coagulopathy: an in‐depth analysis of the coagulation system
date: 2020-08-04
journal: Eur J Haematol
DOI: 10.1111/ejh.13501
sha: b30353723138053b9b0aa1204aaaa534e5f4c7cf
doc_id: 956792
cord_uid: qp76h6he

BACKGROUND: Abnormal coagulation parameters have been reported in COVID‐19 patients. Although the underlying mechanism of COVID‐19 coagulopathy remains unknown, it has been suggested to be a form of disseminated intravascular coagulation (DIC). OBJECTIVES: The aim of our study was to analyze the coagulation parameters of patients with COVID‐19, determine if coagulation factors consumption occurs and identify potential prognostic biomarkers of the disease. PATIENTS/METHODS: Blood samples from hospitalized patients with COVID‐19 pneumonia were collected. We performed basic coagulation tests and quantification of coagulation factors and physiological inhibitor proteins. Laboratory data were compared with clinical data and outcomes. RESULTS: The study involved 206 patients (63.6% male). D‐dimer was particularly elevated (median 450 ng/ml; IQR 222.5‐957.3). Free protein S levels were below the normal range (median 56.6%; IQR: 43.6‐68.9) and factor VIII showed an increasing trend (median 173.4%; IQR: 144.1‐214.9). However, all coagulation factors were within normal limits. We found no correlation between abnormal coagulation parameters and thrombosis, except for higher D‐dimer (HR 1.99; 95% CI 1.3‐3.1; p=0.002). CONCLUSIONS: COVID‐19 is associated with coagulopathy that correlates with poor prognosis. However, we did not demonstrate a consumption of coagulation factors, as seen in DIC.

In late December 2019, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was identified in Wuhan, China [1, 2] . Since then, this virus has rapidly spread worldwide, with more than 15 million confirmed cases of COVID-19, the clinical condition associated with this infection, Accepted Article and more than 630,000 deaths at the time of writing [3] .

The pathophysiology of COVID-19 and the underlying mechanism of its clinical manifestations remain unclear [4] . However, several studies have reported abnormal coagulation parameters, notably in patients with COVID-19 associated pneumonia and acute respiratory distress syndrome (ARDS) [5, 6] . Among the hemostatic system alterations identified, increased D-dimer levels seem to constitute an independent biomarker of poor prognosis in COVID-19, as first stated in early reports from China and confirmed in different series of patients worldwide [7] .

Additionally, these reports suggested that COVID-19 coagulopathy might be a form of disseminated intravascular coagulation.

In line with the hypothesis that coagulopathy may play a role in COVID-19 pathogenesis, recent data suggest an increased risk of developing deep vein thrombosis (DVT) and pulmonary embolism (PE) among patients with severe COVID-19 infection [8, 9] , especially those admitted in intensive care units (ICU), despite adequate thromboprophylaxis. Furthermore, post-mortem studies have highlighted the presence of pulmonary microthrombi and capillarostasis [10] .

Despite all the efforts of the scientific community to understand COVID-19 infection and its associated coagulopathy, there is still a lot to clarify on its mechanism and the way the hemostatic system plays a role on the development of this disease and its clinical complications.

The aim of our study was to analyze the coagulation parameters of patients with COVID-19 pneumonia admitted to our hospital, determine if coagulation factors consumption occurs and identify potential prognostic biomarkers of this new disease.

We conducted a retrospective cohort study performed at Gregorio Marañon Hospital in Madrid.

Consecutive blood samples from hospitalized adult patients with COVID-19 pneumonia were collected from our Hemostasis laboratory between 3 rd April and 3 rd May 2020. According to the local COVID-19 protocol, blood samples were obtained daily from every hospitalized patient with COVID-19, including one to perform a basic coagulation test. Thus, the samples collected for our study were either diagnostic or follow-up available samples from inpatients with COVID-19 pneumonia. We collected an average of 10 samples per day, that were selected on a random basis. One sample of blood was obtained from each patient.

Inclusion criteria were blood samples from patients with COVID-19 aged 18 years or older hospitalized in our center, regardless the time from admission. COVID-19 diagnosis was defined by positive PCR in nasopharyngeal swab or by radiologic and analytical findings highly Accepted Article suggestive of the disease. Patients receiving vitamin K antagonists (VKAs) less than 10 days prior to the blood sample withdrawal were excluded to avoid interference with coagulation tests results. This study was approved by our Institutional Ethics Committee and it was executed along with the international ethical recommendations for conducting research in humans following the latest revision of Declaration of Helsinki.

For each sample, we conducted a complete analysis of coagulation parameters. Demographic, clinical and routine laboratory data were obtained from the hospital electronic patient record using a standard data collection form. Routine laboratory data, according to our local COVID-19 protocol, included cell blood count, biochemistry and acute phase reactants such as C-reactive protein, procalcitonin, ferritin and interleukin-6. For each subject, clinical and routine laboratory data were collected at the time the hemostasis testing was performed. Qualitative variables are presented using frequency distribution and percentages. Quantitative variables are presented using the mean and the standard deviation (SD) if they are normally distributed or the median and the interquartile range (IQR) if they follow a non-normal distribution. 

A total of 206 hospitalized patients with COVID-19 pneumonia were analyzed. Mean age of the sample was 63.6 (SD 13.4) years and 63.6% of the patients were male (Table 1) 

The coagulation parameters assessment demonstrated normal median PT, INR and APTT.

However, D-dimer levels were particularly elevated (median 450 ng/ml; IQR 222.5-957.3) with 69.4% of patients above the normal range (0-250 ng/ml) and with the highest levels in patients on Accepted Article mechanical ventilation (median 1773 ng/ml; IQR 412-2812; p=0.03). Mean fibrinogen levels were also above the upper limit (571.2187.7 mg/dL; normal range 200-400 mg/dL) which could be due to an acute phase response, as we found a statistically significant correlation between fibrinogen levels and other acute phase reactants such as C-reactive protein (r = 0.67; p<0.01) and procalcitonin (r = 0.33; p<0.01). There was no difference in fibrinogen levels based on oxygen requirements (p=0.20)

Factor VIII levels showed an increasing trend (median 173.4%; IQR: 144.1-214.9; normal range 50-200%) and factor IX was slightly above the upper limit of the normal range (median 142.8%; IQR 122.6-170.6; normal range 60-140%). Factor VIII and IX were above the normal range in 32.9% and 56.1% of patients respectively. The highest levels of factor VIII were observed in patients receiving high-flow oxygen therapy (median 203.8%; IQR 184.7-253.7) and mechanical ventilation (median 190.6%; IQR 167.1-234.5) (p=0.01), and there were no statistically significant differences in factor IX levels regarding oxygen therapy (p=0.08).

The remaining factors of the coagulation cascade were within normal limits (60-140% Regarding physiological coagulation inhibitors, free protein S levels were below the normal range (median 56.6%; IQR: 43.6-68.9; normal range 70-140%) with 33% of patients showing levels below 50%. We observed the lowest levels of free protein S in patients receiving low-flow oxygen therapy (median 52.6%; IQR 42-63.4; p<0.001). Protein C and antithrombin levels were within normal limits, and there was no difference in these parameters based on oxygen requirements.

Data collected from routine laboratory tests exposed that lymphocytes levels were decreased in our cohort (median 1x10 3 

This article is protected by copyright. All rights reserved been observed in patients on mechanical ventilation. All the other routine laboratory data analyzed were within normal levels (Table 1) .

Abnormal coagulation and other routine laboratory parameters were associated with poor prognosis. Notably, D-dimer levels were significantly higher in non survivors (median 1472.5 versus 385 ng/ml; p=0.004). Similarly, PT was more elevated in non-survivors (median 14 versus 12.7 seconds; p<0.001). Significant differences were also observed in acute phase reactant levels between both groups (Table 1) , and lymphopenia was associated with poor prognosis (median 0.6x10 3 /L in non-survivors versus 1x10 3 /L in survivors; p=0.007).

Despite keeping within the normal range, non-survivors showed slightly lower levels of coagulation factors involved in the extrinsic (factor VII) and common (factor II and factor X) coagulation pathways, and this difference was statistically significant. We also found lower levels of protein C (median 107% versus 128%; p=0.001) and antithrombin (median 86% versus 107%;

p=0.008) in the non-survivors (Figure 1 ).

Non-survivors had lower platelet count and this was associated with poor prognosis (median 203 x10 3 /L versus 262 x10 3 /L; p=0.013). Thrombocytopenia was present in 44.4% of the nonsurvivors versus 8.5% of the survivors (p<0.001). Criteria of overt disseminated intravascular coagulation (DIC) as defined according to the ISTH DIC score [11] were met in 22.2% of the nonsurvivors versus 3.7% of the survivors (p=0.001).

The eleven patients who met criteria of overt DIC were analysed separately. The majority of these patients (54.6%) had high oxygen requirements and were either receiving oxygen via a high-flow device (9.1%) or mechanical ventilation (45.5%) (p<0.001). Platelet counts were significantly lower in this subgroup (median 98 x10 3 /L versus 259 x10 3 /L; p=0.003), D-dimer levels were higher (median 2812 versus 382 ng/ml; p<0.001) and PT was prolonged (median 16.5 versus 12.7 seconds, p<0.001). We found no statistical difference in fibrinogen levels (468193 versus 577186 mg/dL; p=0,06). Levels of protein C (median 66% versus 127%; p=0.001) and antithrombin (median 74% versus 107%; p=0.004) were significantly decreased in patients with DIC, and we found lower levels of factor II (median 77.7% versus 102.3%; p=0.002), X (median 45% versus 111.1%; p<0.001) and XII (median 83.7% versus 117.3%; p= 0.04). Acute

phase reactants showed also higher levels in patients with DIC, with C-reactive protein (median 10.8 versus 2.9 mg/dL; p=0.02), procalcitonin (median 0.3 versus 0.06g/L; p=0.001) and interleukin 6 (median 137.9 versus 33.6 pg/mL; p=0.002) significantly more elevated in this subgroup.

Patients admitted in the ICU (12.6%) had more comorbidities (100% versus 75%; p=0.004) than the rest of patients, and 15.4% of them met criteria of DIC (p=0.015). Regarding the laboratory findings, these patients had lower platelet count (median 189 x10 3 /L versus 259 x10 3 /L; p<0.001), higher levels of factor V (median 135% versus 108%; p=0.001) and lower of factor X (median 101% versus 111%; p=0.003). Similarly, acute phase reactants were also more elevated, including C-reactive protein (median 3.9 versus 2.8 mg/dL; p=0.009), ferritin (median 1038 versus 649g/L; p=0.007) and interleukin 6 (median 138.3 versus 29.1 pg/mL; p<0.001). In general, patients admitted in the ICU had a poorer prognosis, and 19.2% of them died (p=0.04) ( Table 2) .

Thrombotic events were observed in 23 patients during hospitalization, including venous thromboembolism in 18 patients (8.7%) and arterial thrombosis in 5 patients (2.4%). Venous thrombosis consisted in DVT in 10 patients (55.6%) and PE in 8 patients (44.4%), while arterial thrombosis manifested itself as myocardial infarction in 3 patients (60%) and acute lower limb ischemia in 2 patients (40%). Higher D-dimer levels were associated with higher risk of thrombotic events (1447.5 versus 380 ng/ml; p<0.001). We found no statistically significant correlation between other abnormal coagulation parameters and higher risk of thrombosis (Table   3 ).

A multivariate analysis using a Cox proportional hazards regression model was performed to adjust for potential confounding factors such as sex, age, different timing between admission and blood sample collection and comorbidities, including those that could affect coagulation test results such as an active cancer, history of chronic liver disease or abnormal liver function. This model revealed that age was an independent prognostic factor for survival (HR 1.32; 95% CI

This article is protected by copyright. All rights reserved 1.09-1.62; p=0.005) and there was a lack of association between D-dimer levels and poor prognosis (HR 0.99; 95% CI 0.99-1; p=0.02). The remaining variables were not statistically significant. Meanwhile, the time-to-event Cox model to assess the risk of thrombosis confirmed that D-dimer was an independent predictor of the development of a thrombotic event (HR 1.99; 95% CI: 1.3-3.1; p=0.002) and this was not interfered by other parameters such as sex, age or comorbidities.

Many reports have been published regarding coagulopathy in COVID-19 patients since the disease appeared [5, 6] . Tang et al [7] first detected increased PT and higher levels of fibrinogen and D-dimer in the non-survivors group in a cohort of 183 patients. Similarly, our bivariate analysis proved that elevated D-dimer is associated with poor prognosis, with a four-fold increase among non-survivors. In the same line, we observed prolonged PT and lower platelet counts in non-survivors. However, the multivariate model revealed a lack of association between D-dimer levels and death (HR 1). This could imply that early reports on COVID-19 coagulopathy could have overrated the effect of D-dimer, as they did not perform a multivariate analysis, and the results could have been biased by potential cofounders. In consonance with this hypothesis, recent publications performing a regression test have shown results similar to ours, with values of HR close to 1 for D-dimer [12, 13] .

The association of low platelet count, prolonged PT and increased D-dimer observed in the first published reports is suggestive of DIC, even if the pattern is clearly different to DIC seen in other clinical situations such as sepsis or malignant tumors. In fact, a minority of patients meet criteria for DIC according to the ISTH score -5.3% in our cohort-, in agreement with other data published [14] . This has led to the definition of a new concept coined as Pulmonary Intravascular Coagulopathy (PIC) [15] , as distinct to DIC, which suggests that the diffuse bilateral pulmonary inflammation observed in COVID-19 may justify development of a pulmonary-specific vasculopathy and abnormalities in the coagulation parameters.

One of the novelties of our study is the exhaustive analysis of all the factors involved in the coagulation cascade, as well as the determination of physiological coagulation inhibitors with the purpose to asses if a consumption coagulopathy is present in these patients. In the bivariate analysis, lower levels of factors II, VII and X were associated with increased mortality and this could explain the prolongation of PT among non-survivors, even if these results did not reproduce when adjusting these variables for potential confounding parameters in the Cox regression.

This article is protected by copyright. All rights reserved Factor VIII showed an increasing trend in the whole cohort, which could be interpreted considering the inflammatory state and systemic cytokine storm present in these patients, as factor VIII has been reported to act as an acute phase reactant [16] . However, we found no correlation between factor VIII and mortality. Other authors have reported much higher levels of Factor VIII in ICU patients [17] ; these differences could be explained by the low rate of ICU patients in our cohort (12.6%).

Likewise, a mild protein S deficiency was found in our cohort. Protein S is a vitamin K-dependent protein and a physiological coagulation inhibitor. However, free protein S also binds to TAM -Tyro3, Axl and Mer -receptors located on the surface of macrophages, that are involved in homeostasis, inflammation, interaction with viruses and pathophysiology of acute lung injury [18, 19, 20] .

Our study showed a moderate incidence of thrombotic events (11.1%) despite adequate thromboprophylaxis, in consonance with results recently published by Demelo-Rodriguez et al in a cohort of 156 patients from our hospital [21] . Further studies have reported a higher incidence of thrombosis [22] , but routine thromboprophylaxis was not standard of care. Other studies have found an increased incidence of thrombotic events, mostly in the ICU setting [9, 23] . However, no correlation was found between abnormal coagulation parameters and higher risk of thrombotic events, beyond D-dimer levels (HR 1.99). One possible hypothesis for this discordance is the fact that patients with poorer prognosis may develop pulmonary microthrombi due to COVID-19 associated coagulopathy [10] , as demonstrated by autopsy findings [24] , which could have a significant impact in mortality. However, macroscopic thrombosis such as DVT and PE may not be as influenced by coagulation parameters alteration.

Some studies suggest that inflammation-induced endothelial cell injury could lead to an activation of the fibrinolytic system [25, 26] , which may justify the elevated D-dimer levels in patients with severe COVID-19. Furthermore, it should be noted that a hyperfibrinolyic state and elevated levels of Plasminogen activator inhibitor-1 (PAI-1) were observed in the SARS-CoV epidemics in 2002 [27] and similar alterations have been recently reported regarding SARS-CoV-2 [28] .

Therefore, a complete evaluation of the fibrinolytic system could provide further information about COVID-19 associated coagulopathy.

In addition, other possible underlying mechanisms of COVID-19 coagulopathy have been recently suggested. Elevated antiphospholipid antibodies have been observed in some patients [29, 30] which could explain the endothelial damage that occurs in this disease and, thus, the coagulation abnormalities observed in these patients. The coagulation alterations observed in

This article is protected by copyright. All rights reserved antiphospholipid syndrome are similar to those of COVID-19 coagulopathy, and they probably share some pathogenic mechanisms [31] . However, a recent study by Galeano-Valle et al found no elevation of antiphospholipid antibodies in a series of 24 patients with COVID-19 pneumonia and venous thromboembolism [32] . On the other hand, resemblance to thrombotic microangiopathy (TAM) has been suggested, with reportedly increased levels of von Willebrand Factor (vWF) due to the vascular injury caused by SARS-CoV-2 endothelial infection. Furthermore, complement system activation might play a role in the endothelial damage caused by COVID-19, and the potential effect of anticomplement treatments is being studied [33] .

This study has some limitations. First, it was a single-centre cohort study with a restricted sample size, and 14.6% of patients remained hospitalized at the time of data collection. Second, coagulation parameters analysis was not performed on admission in all cases and this could bias our results, especially those concerning thrombotic risk. Third, information regarding bleeding events was not gathered. Fourth, certain conditions such as other active infections or medications could affect some of the coagulation tests results, especially those dependent on vitamin K, and that information was not gathered and could be a potential bias of our study. Finally, the study lacks a control arm and did not evaluate serum von Willebrand factor levels, which is emerging as a clinically relevant biomarker.

In conclusion, our findings confirm that severe COVID-19 infection is associated with coagulopathy that correlates with poor prognosis. However, it seems that COVID-19 coagulopathy is not a form of disseminated intravascular coagulation, as we did not demonstrate a consumption of coagulation factors in our cohort. Further studies to explore other possible mechanisms of this coagulopathy are required, including an exhaustive analysis of the fibrinolytic system and other coagulation parameters such as ADAMTS-13 or von Willebrand factor.

Additionally, it would be interesting to perform future studies to assess whether any of the features observed in COVID-19 coagulopathy are predictive of hospitalization, in order to be able to act at an early stage of the disease. 

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

China Novel Coronavirus Investigating and Research Team. A Novel Coronavirus from Patients with Pneumonia in China

Coronavirus disease (COVID-19): Situation report-187

Cytokine storm intervention in the early stages of COVID-19 pneumonia

COVID-19 Coagulopathy in Caucasian patients

The procoagulant pattern of patients with COVID-19 acute respiratory distress syndrome

Abnormal coagulation parameters are associated with poor prognosis in patients with novel coronavirus pneumonia

Humanitas COVID-19 Task Force. Venous and arterial thromboembolic complications in COVID-19 patients admitted to an academic hospital in

Incidence of venous thromboembolism in hospitalized patients with COVID-19

Post-mortem examination of COVID19 patients reveals diffuse alveolar damage with severe capillary congestion and variegated findings of lungs and other organs suggesting vascular dysfunction

Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation

Risk Factors for Mortality in Patients with COVID-19 in

Epidemiology, clinical course, and outcomes of critically ill adults with COVID-19 in New York City: a prospective cohort study

Coagulation abnormalities and thrombosis in patients with COVID-19

Immune mechanisms of pulmonary intravascular coagulopathy in COVID-19 pneumonia

The Factor VIII acute phase response requires the participation of NFkappaB and C/EBP

Hypercoagulability of COVID-19 patients in intensive care unit. A report of thromboelastrography findings and other parameters of hemostasis

Direct effects of protein S in ameliorating acute lung injury

Gas6, and protein S: roles in inflammation and hemostasis

Axl and MerTK receptor tyrosine kinases maintain human macrophage efferocytic capacity in the presence of viral triggers

Incidence of asymptomatic deep vein thrombosis in patients with COVID-19 pneumonia and elevated D-dimer levels

Prevalence of venous thromboembolism in patients with severe novel coronavirus pneumonia

Confirmation of the high Accepted Article This article is protected by copyright. All rights reserved cumulative incidence of thrombotic complications in critically ill ICU patients with COVID-19: an updated analysis

Pulmonary Vascular endothelialitis, Thrombosis and Angiogenesis in Covid-19

Fibrinolytic abnormalities in acute respiratory distress syndrome (ARDS) and versatility of thrombolytic drugs to treat COVID-19

Activation of Coagulation and Fibrinolysis in Acute Respiratory Distress Syndrome: A prospective pilot study

Mechanisms of severe acute respiratory syndrome coronavirus-induced acute lung injury

Endotheliopathy in COVID-19-associated coagulopathy: evidence from a single-centre, cross-sectional study

Accepted Article This article is protected by copyright. All rights reserved Y, Li Y, Zhang S. Coagulopathy and Antiphospholipid Antibodies in Patients with Covid-19

Clinically significant anticardiolipin antibodies associated with COVID-19

Disseminated intravascular coagulation in catastrophic antiphospholipid syndrome: clinical and haematological characteristics of 23 patients

Antiphospholipid antibodies are not elevated in patients with severe COVID-19 pneumonia and venous thromboembolism

The unique characteristics of COVID-19 coagulopathy. Version 2. Crit Care

Accepted Article Routine laboratory tests Hemoglobin (g/dL)

Factor II (%)

Factor VII (%)

Factor VIII (%)

We thank laboratory technicians for their hard work and collaboration for accomplishment of this study.

The authors have no conflict of interest to declare.

This article is protected by copyright. All rights reserved