Introduction
Novel coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, occurs in various clinical manifestations ranging from mild upper respiratory tract damage to severe pneumonia with significant respiratory failure that requires invasive respiratory support [1–3]. As known, severe course of disease is caused by overproduction of proinflammatory cytokines (cytokine storm) and systemic hyperinflammatory reaction that significantly predetermines disease course, prognosis and outcome [4, 5]. It is noted that autopsy studies demonstrate extensive inflammatory capillary thrombosis of pulmonary vessels, different from embolic thrombi in pulmonary arteries, that seems to contribute to hypoxemia and respiratory failure progression 6.
There is accumulated evidence that acute respiratory distress syndrome (ARDS) in COVID-19 differs from typical ARDS in that initial hypoxemia is associated with early loss of surfactant, dysregulation of neutrophil extracellular trap formation, alveolar cell damage, and local fibrin deposition [7–9]. In addition to the specific changes in pulmonary vessels, severe COVID-19 is characterized by systemic hypercoagulation represented by arterial and venous thrombotic events, while their contribution to mortality from COVID-19 is unknown. Also, data about the association of pathological changes in the hemostatic system with the development of ARDS are insufficient, and there are few works in this direction [8, 11].
Understanding the pathophysiology of coagulopathy and phasing of the changes in the hemostatic system with the development of ARDS in COVID-19 will allow to identify potential targets for therapeutic intervention, exposure to which can reduce the risk of both intrapulmonary and systemic thrombosis.
The study objective was to determine association between the severity of respiratory failure and pathological changes in the hemostatic system in COVID-19 patients.
Materials and methods
A prospective observational two-center study was conducted on the basis of FSBI Federal Center for Traumatology, Orthopedics and Endoprosthetics of the Ministry of Health of the Russian Federation, Barnaul and City Hospital No. 4 named after N.P. Gull, Barnaul. The study period was from November 2020 to January 2022.
Inclusion criteria: Patients with a confirmed diagnosis of COVID-19 (positive laboratory test result for the presence of SARS-CoV-2 RNA using nucleic acid amplification tests (NAAT) or SARS-CoV-2 antigen using immunochromatographic assay); severe and extremely severe course of disease; need for respiratory support in intensive care unit (ICU); lung damage confirmed by computed tomography (CT); age over 18 years.
Exclusion criteria: Refusal of the patient and/or the relatives to participate in the study; age less than 18 years; septicemic condition at ICU admission [12], for women – pregnancy and lactation.
The morbidity as well as the algorithm of diagnosis and treatment were determined according to the temporary guidelines for the prevention, diagnosis and treatment of novel coronavirus infection, version 15 (dated February 22, 2022) [13].
The study included 225 patients at admission to the intensive care unit of a specialized hospital. During the follow-up, 21 patients were excluded from the analysis due to insufficient laboratory and clinical data reflected in the medical history, that did not allow to study the clinical course and outcome of the disease. Among the other 204 patients in the group, 168 patients had severe disease and 36 patients had extremely severe disease at ICU admission. The patients were divided into two groups according to disease outcome: fatal group (n = 106) and survived group (n = 98). The relief of acute respiratory failure (ARF) was achieved in the survived patients in ICU with the use of high-flow oxygenation and non-invasive ventilation that made it possible to transfer the patients to the therapeutic department. In the deceased patient group, 94 patients required noninvasive ventilation at admission, then they were transferred to invasive ventilation during disease progression, and 12 patients initially received respiratory support in the form of invasive ventilation.
The average age of patients was 65 years (minimum age 22 years, maximum age 97 years). The men to women ratio is 91:113. The time from hospital admission to ICU transfer ranged from 0 to 41 days. Median (Me) length of stay in the intensive care unit before the therapeutic department transfer or death was 7.0 (95% CI for Me — 6.0–7.0, Q1–Q3: 4–10 days).
All patients included in the study had acute respiratory distress syndrome (ARDS) that manifested as tachypnea, a decrease in SpO2 of less than 90%, cyanosis of the nasolabial triangle, PaO2/FiO2 index < 300 mmHg, the results of computed tomography and chest X-ray. 128 patients showed lung damage from 50 to 75 % that corresponds to CT-3. Lung damage ranged from 75 to 100 % in 76 patients that corresponds to CT-4. To analyze the severity of respiratory failure, the formulas and criteria presented in the 2012 Berlin definition of ARDS were used (Table 1).
Severity of ARDS | Oxygenation index calculated with the PaO2/FiO2 ratio* |
---|---|
Mild | 200 mmHg < PaO2/FiO2 ≤ 300 mmHg* with PEEP or CPAP ≥ 5 cmH2O |
Moderate | 100 mmHg < PaO2/FiO2 ≤ 200 mmHg with PEEP ≥ 5 cmH2O |
Severe | PaO2/FiO2 ≤ 100 mmHg with PEEP ≥ 5 cmH2O |
To assess dynamics of the clinical picture of the disease and to study the hemostatic profile, the main time points were determined: I point – the first day, admission to intensive care unit, II point — 3–5 days, III point — 7–10 days of treatment.
In addition to the peripheral blood indicators regulated by the guidelines, 58 different peripheral blood parameters were performed and analyzed within the framework of the study, including the indicators that reflect the functioning of various parts of the hemostatic system and the values related to the immune system and the complement system. The present paper demonstrates an analysis of only the most significant parameters of hemostatic reactions.
Laboratory examination technology
Venous blood samples were collected from the ulnar vein into VACUETTE tubes with sodium citrate buffer in the ratio of 9:1 (9NC Coagulation sodium citrate 3.2 %, Greiner-Bio-One, Austria). The blood was centrifuged at 1400 g for 15 min at room temperature. Before conducting enzyme immunoassays, plasma was stored at –40 °C in a deep-freeze refrigerator MDF-192 Ultra-low temperature freezer (Sanyo, Japan) during one day up to two months.
The platelet count in the blood was evaluated using the DREW-3 automated hematology analyzer (Drew Scientific, USA). Concentration of von Willebrand factor (vWF) (vWF Ag reagent kit, Siemens Healthcare Diagnostics, Germany), factor II (prothrombin) activity (with the use of factor II deficient plasma and thromboplastin reagent Thromborel, Siemens Healthcare Diagnostics, Germany), factor XIII activity (Berichrom Factor XIII reagent kit), D-dimer concentration (D-dimer Red-700 reagent kit, Helena Bioscience, Great Britain) and α2-antiplasmin (α2-AP) activity (Berichrom α2-Antiplasmin reagent kit, Siemens Healthcare Diagnostics, Germany) were evaluated with Sysmex CA-1500 and Siеmens BCS XP automated coagulometers (Sysmex Corporation, Japan).
ADAMTS-13 metalloproteinase activity (TECHNOZYM ADAMTS-13 Activity ELISA test kit), tissue plasminogen activator (t-PA) and plasminogen activator inhibitor type 1 (PAI-1) antigens (t-PA Combi Actibind ELISA test kit and TECHNOZYM® PAI-1 Actibind® ELISA test kit, respectively, Technoclone GmbH, USA) were evaluated with Real-R automated photometer (Vector-Best-Baltika, Russia).
Fluoroscan Ascent plate fluorimeter (ThermoFisher SCIENTIFIC, Finland) equipped with a dispenser with Thrombinoscope 3.0.0.26 software was used to perform TGT. Coagulation of the studied blood plasma was carried out in the presence of 5.0 pmol of tissue factor and 4 µmol of phospholipids (PPP-Reagent 5 pM, Thrombin Calibrator, FluCa-Kit, CAT Thrombinoscope BV, the Netherlands). Thrombin generation was recorded by measuring a fluorogenic substrate (Z-Gly-Gly-Arg-AMC) signal. The following indicators were taken into account: PTC (Peak thrombin - peak thrombin concentration, nmol/L) — the maximum thrombin concentration per unit time, and ETP (endogenous thrombin potential, nmol/min) — an indicator estimating the area under the thrombin generation curve that counts on the features of inactivation of this enzyme. All sample tests were carried out in two parallel versions, the average was automatically taken into account.
The tactics of intensive care, resuscitation, manipulation and transportation of patients with severe and extremely severe COVID-19 were carried out according to the Guidelines of the All-Russian public organization “Federation of Anesthesiologists and Reanimatologists” 14.
Also, it should be noted that before peripheral blood sampling for analysis at the first point, the patients did not receive anticoagulation. Further, anticoagulation is represented by: unfractionated heparin in 57.4 % (n = 117) of cases and low molecular weight heparin (LMWH: sodium enoxiparin and calcium nadroparin in equal proportions) in 42.6 % (n = 87) of cases. Anticoagulants were prescribed in a weight-adjusted therapeutic dose. At the second and third study points, blood sampling was carried out in the morning, at least 6 hours after the last intake of anticoagulants. The selected time range is due to the pharmacokinetics of the drugs: the duration of action of UFH is 4–6 hours [15], the half-life of LMWH is 5–6 hours [16].
Statistical analysis
Statistical data processing was carried out using the statistical software package MedCalc Version 20.110 (a product of MedCalc Software Ltd, Belgium). The Kolmogorov-Smirnov test was used to check the variation series for normality of distribution. Nonparametric methods were used for statistical while the results of the analyzed parameters of hemostatic tests and respiratory index values did not correspond to the normal distribution. The data of laboratory and instrumental indicators are presented in the form of median (Me), 95 % confidence interval (95 % CI) and interquartile range [Q1–Q3]. The degree of correlation between individual parameters of the hemostatic system and the data of the oxygenation index was determined with the Spearman correlation coefficient. Multiple regression analysis with the Kolmogorov-Smirnov test was used to determine association between respiratory failure and general disorders in the hemostatic system.
The study was approved by the local Ethics Committee of Altai State Medical University, Protocol No. 9 dated October 23, 2020.
Results
The results of a dynamic study of the indicators of various parts of the hemostatic system are presented in Figures 1, 2, 3 and 4.
When coagulative hemostasis was studied (Fig. 1 and 2), no statistical differences were found in comparing the fibrinogen concentration ratio at all study points in the comparison groups: p = 0.6004, p = 0.4707 and p = 0.6127, respectively. At the same time, the activity of the precursor of thrombin, factor II (prothrombin), in the deceased group was statistically significantly lower compared to the survived patients. Decreased prothrombin activity in blood plasma was specific to both groups by 7–10 days of ICU stay, but significantly lower values were characteristic of the group of deceased patients (89 % vs. 126 %, p = 0.0095).
When the activity of fibrin stabilizing factor (F XIII) was analyzed, the medians of this indicator were found to be comparable at the first and second study points (p = 0.5240 and p = 0.4630, respectively). However, its activity was significantly higher (p = 0.0043) in the group of deceased patients than in that of survived patients on days 7–10.
Analysis of the thrombin formation dynamics based on the values of the thrombin generation test indicates that thrombin generation (according to the Peak thrombin indicator) is generally higher in the group of deceased patients than in that of deceased patients, but there was no statistically significant difference.
A significantly higher concentration of von Willebrand factor (p < 0.0001) and definitely lower activity of ADAMTS-13 metalloproteinase (p = 0.0037) were determined in the fatal group on the very first day of ICU stay (Fig. 3). Both comparison groups are characterized by an increased von Willebrand factor level, with a significant difference in the indicator values at the second (p = 0.0001) and third (p = 0.0242) study points. At the same time, the concentration of ADAMTS-13 metalloproteinase remained stable in the survived group by 7–10 days of ICU stay (p = 0.0530), and it decreased slightly by 19 % in the fatal group (p = 0.2788).
Platelet count analysis showed that their number had no statistically significant differences in the comparison groups (survived and deceased patients) on the first day of the study — at ICU admission (p = 0.3254). The platelet count decreased in both groups by 3–5 days of following up, without reaching statistical significance (p = 0.3304). On 7–10 days in the group of deceased patients, the platelet count decreased by 53.2 % in respect of the first study point (p = 0.0001), with a significant difference in the same indicator in the survived group (p = 0.0251).
The study of dynamics of the indicators reflecting the fibrinolysis system function (Fig. 4) allowed to establish that multidirectional reactions were noted in the comparison groups, given its regulation complexity. According to the findings, significantly higher t-PA concentrations (1.6 times median) and significantly lower α2-antiplasmin activity were determined in the fatal group at the first study point that may indicate fibrinolysis activation in the systemic circulation. Later, however, by 7–10 days of ICU stay, the t-PA concentration significantly decreased (by 16 %; p = 0.0482) in patients with a fatal outcome with an increased concentration of its inhibitor (PAI-1) by 29 % (p = 0.0610). In contrast, the survived group was characterized by significantly increased t-PA concentration (p = 0.0028) with identical PAI-1 values on the first and 7–10 days of ICU stay (p = 0.1064). It should be noted that the activity of α2-antiplasmin, the main plasmin inhibitor, was comparable by study points in the comparison groups.
The concentration of D-dimer, one of the probable fibrin degradation products, was determined to be significantly higher in the fatal group at the first study point (p = 0.0011), while the values of this indicator in 2 and 3 study points were comparable to each other (p = 0.3030 and p = 0.4179, respectively). Along with this, by 7–10 days of ICU stay, a dynamic increase in D-dimer concentration was revealed by 26.6 % in the group of deceased patients (p = 0.4822) and by 83.9 % in that of survived patients (p = 0.2412).
Summing up the above, we can conclude that severe and extremely severe COVID-19 is associated with dysfunction in all parts of the hemostatic system that is characterized by endothelial cell activation, hypercoagulation shift and fibrinolysis inhibition.
When the respiratory index value was assessed, it was determined that patients with a fatal outcome had its significant decrease compared to the survived group. Thus, by 7–10 days in the survived group, the oxygenation index was statistically significantly higher by 2.15 times (p = 0.0001) than in the fatal group (Fig. 5).
Considering that the main objective of our study was to determine the association between the severity of respiratory failure and pathological changes in the hemostatic system, a correlation analysis of the respiratory index value and the analyzed parameters of the hemostatic system was carried out at the selected study points to identify the direction and connection strength (Table 2).
Hemostatic system parameters | I point | II point | III point | |||
---|---|---|---|---|---|---|
Spearman’s coefficient | p | Spearman’s coefficient | p | Spearman’s coefficient | p | |
Activity of factor II, % | 0.2110 | 0.0040 | 0.261 | 0.0120 | 0.3690 | 0.0268 |
Activity of factor XIII, % | 0.0490 | 0.5562 | –0.0783 | 0.4984 | –0.2550 | 0.1597 |
Peak thrombin, nmol/L | 0.0308 | 0.7054 | 0.0233 | 0.8357 | –0.1960 | 0.3172 |
Endogenous thrombin potential (ETP) | 0.0948 | 0.2517 | 0.0330 | 0.7744 | –0.1070 | 0.6121 |
Platelet count ×109, % | 0.0353 | 0.6252 | 0.0730 | 0.4728 | 0.3140 | 0.0561 |
Concentration of factor vWF, % | –0.2220 | 0.0019 | –0.278 | 0.0043 | 0.3880 | 0.0232 |
Activity of ADAMTS 13, ME/ml | 0.2430 | 0.0013 | -0.0033 | 0.9741 | 0.0800 | 0.6670 |
Concentration of t-PA (Ag), ng/ml | –0.2120 | 0.0024 | –0.0688 | 0.4858 | 0.1160 | 0.5013 |
Concentration of PAI-1 (Ag), ng/ml | –0.1910 | 0.0327 | –0.3880 | 0.0013 | –0.180 | 0.3883 |
Activity of α2-antiplasmin, % | 0.1470 | 0.0502 | 0.1340 | 0.2287 | 0.1120 | 0.5291 |
Concentration of D-dimer, ng/ml | –0.1420 | 0.0470 | –0.1840 | 0.1370 | –0.2270 | 0.3351 |
According to the findings, a reliable positive correlation of the activity of factor II, ADAMTS 13 metalloproteinase, α2-antiplasmin and a reliable negative correlation of the concentration of factor vWF, PAI-1 and D-dimer with the respiratory index at various study points were identified. At the same time, connection strength for all the parameters was weak, that does not make it possible to identify the dominant dysfunction in the hemostatic system that determines the severity of respiratory failure.
To better predict the possible association of the analyzed parameters of the hemostatic system with the respiratory index value at the study points, a multiple regression analysis was performed considering the multicollinearity of factors and their backward selection in the model (Table 3).
Study point | Independent variable | Ratio | t | р | R2 | R |
---|---|---|---|---|---|---|
1 day | Constant | 260.9 | — | — | 0.2097 | 0.4316 |
Concentration of factor vWF, % | –0.1892 | –3.236 | 0.0017 | |||
3–5 days | Constant | 219.3 | — | — | 0.3524 | 0.6024 |
Concentration of PAI-1 (Ag), ng/ml | –0.6498 | –3.582 | 0.0010 | |||
7–10 days | Constant | 492.3 | — | 0.7813 | 0.8756 | |
Activity of factor II, % | 1.3980 | 2.463 | 0.0263 | |||
Peak thrombin, nmol/l | –0.3375 | –3.304 | 0.0048 | |||
Activity of factor XIII, % | –1.6773 | –3.809 | 0.0017 |
According to the findings, the contribution of specific factors of the hemostatic system to the formation of the severity of respiratory failure at each study point was different. On the first day of ICU stay, the concentration of von Willebrand factor explains 21 % of the variability of the respiratory index value, while an inverse correlation of average strength is observed. When the patients were examined on 3–5 days, only the concentration of plasminogen activator inhibitor type 1 (PAI-1) significantly contributed to the respiratory index value variation determining it by 35 % with an average correlation between the variables. The other factors of the hemostatic system lose their significance in the formation of the severity of respiratory failure at the first two study points. Activation of the coagulative hemostasis component on 7–10 days of following up is the determining factor of the critical values of the respiratory index that determines it by 78 % with a strong correlation of the analyzed parameters.
Thus, the severity of respiratory failure in patients with a confirmed diagnosis of severe and extremely severe COVID-19 is gradually associated with endothelial dysfunction (activation) (1 day), inhibition of parietal fibrinolysis (3–5 days) and activation of the coagulative hemostasis component by 7–10 days of ICU stay.
Discussion
Pulmonary intravascular coagulation plays a major role in the pathogenesis of acute respiratory failure in COVID-19 [17]. As known, microthromboembolization of pulmonary vessels in COVID-19 patients causes a progressive decrease in lung perfusion, hypoxic vasoconstriction and increasing hypoxemia [18, 19]. At the same time, its severity can be objectively evaluated by the respiratory index, an integral indicator, that characterizes the severity of respiratory distress syndrome [20–22].
In the present work, it was found that the respiratory index was 2.15 times lower in the fatal group than in the survived group. Hence, it was of interest to identify the association of phasing and severity of disorders in the hemostatic system that determine the possible microthrombosis in the pulmonary vessels with the severity of respiratory failure.
As known, one of the main causes of thrombosis in COVID-19 is vascular endothelial damage caused by viral invasion [19] resulting in that endothelial cells increase the expression of proinflammatory cytokines and von Willebrand factor, and decreased fibrinolytic activity of mast cells is recorded [23]. Our study is consistent with the earlier studies according to a number of the indicators confirming the dysfunction of the hemostatic system. Thus, it was determined that platelet count in the blood significantly decreased in both outcome comparison groups, however, thrombocytopenia was much more significant in deceased patients on 7–10 days, the median was 97×109 versus 157×109/l in the group of survived patients (p = 0.0251). According to E.L. Bulanova et al., the presence of thrombocytopenia increases the risk of death in COVID-19 infection by 5.5 times (95% CI 2.979–10.031) 24.
According to the study results, the decreased respiratory index value on the first day was due to the increased concentration of von Willebrand factor. At the same time, significantly higher concentration ratios were found in the group of deceased patients on 3–5 and 7–10 days of following up compared with the group of survived patients. Significantly lower activity of ADAMTS-13 metalloproteinase (p = 0.0037) in the group of deceased patients was noted only at the first study point. Probably, the discrepancy of its stable activity at 2 and 3 study point is more clinically significant against the background of an increasing concentration of von Willebrand factor, that can result in the increased processes of microthrombosis specific for thrombotic microangiopathy (TMA). Consideration of coagulopathy in COVID-19 as a variant of TMA, as known, supposes a paradigm shift in therapy, for example, plasma exchange [25].
In the conducted work, it was found that severe COVID-19 is characterized by multidirectional fibrinolytic reactions depending on the duration of ICU stay, and, therefore, on the duration of respiratory support. Thus, on the first day of respiratory failure, activation of parietal fibrinolysis (increased concentration of t-PA) was established, followed by its inhibition on 3–5 days of treatment (increased concentration of PAI-1), that is especially marked in patients with a fatal outcome. It was found that it was the excessive concentration of PAI-1 that determined the severity of respiratory distress syndrome at the second study point.
A similar conclusion was made by the researchers in South Carolina. Decreased concentration of PAI-1 was associated with severe hypoxemia, fibrinolysis inhibition in the microcirculation system and subsequent micro- and macrovascular thrombosis in severe COVID-19 11.
The revealed pattern suggests that the basis for the correction of bleeding disorders in patients with severe and extremely severe COVID-19 should be not anticoagulant, but fibrinolytic therapy, that is confirmed by a few studies considering intravenous administration of fibrinolytics [6, 26, 27].
Now there is convincing evidence that hypercoagulation in COVID-19 is secondary to elevated levels of tissue factor, von Willebrand factor, fibrinogen, a decrease in ADAMTS-13 concentration with platelet activation and fibrinolysis inhibition [6, 28], that is confirmed by the findings of the present work.
On the one hand, we identified facts of common knowledge when we analyzed the parameters of the hemostatic system at 3 study point. The severity of respiratory failure, namely, critical values of the respiratory index by the 7–10 days of ICU stay are associated with activation of the coagulative component of hemostasis, that is confirmed by an increase in peak thrombin by 12.7 % in thrombin generation test and by statistically significantly higher activity of factor XIII in the group of deceased patients (p = 0.0043). On the other hand, prothrombin activity value has a negative correlation with the respiratory index. In our opinion, the nature of dynamics of FII and FXIII activity values from the moment of ICU admission to disease outcome suggests intensive consumption of the indicated factors for thrombosis, that warrants further studies.
It should be noted that there are quite few works characterizing decreased activity of factor XII. At the same time, the researchers' results are comparable to our findings. For example, in the work of Austrian researchers, it is reported that the concentration of fibrin-stabilizing factor is one of the important predictors of poor outcome in COVID-19 29.
Thus, it can be assumed that the severity of respiratory failure in patients with severe and extremely severe COVID-19 correlates with pathological changes in the hemostatic system. At the same time, there is a certain phasing of the detected hemostatic disorders that indicates primary acute endothelial damage followed by inhibition of parietal fibrinolysis and the development of secondary hypercoagulation by 7–10 days of respiratory support in the ICU. The revealed regularities of phasing of pathological changes in the hemostatic system that correlate and determine the severity of respiratory failure are a prerequisite for revising therapeutic approaches to the treatment of patients with severe and extremely severe COVID-19.
Conclusions
The severity of respiratory failure in patients with severe and extremely severe COVID-19 is associated with pathological changes in all parts of the hemostatic system.
In patients with severe and extremely severe COVID-19 receiving respiratory support, the phasing of hemostatic changes is determined that correlates with the respiratory index value and consists in endothelial dysfunction followed by inhibition of parietal fibrinolysis and secondary hypercoagulation.
The hypercoagulation shift determines the respiratory index value by 78% with a strong correlation by 7–10 days against the background of concomitant respiratory support in patients with severe and extremely severe COVID-19 (R = 0.8756). At the same time, increased thrombin generation is accompanied by a statistically significant decrease in the activity of factors II and XIII.
Disclosure. The authors declare that they have no competing interests.
Author contribution. All authors according to the ICMJE criteria participated in the development of the concept of the article, obtaining and analyzing factual data, writing and editing the text of the article, checking and approving the text of the article.
Ethics approval. This study was approved by the local Ethical Committee of Altai State Medical University (reference number: 9-23.10.2020).
Funding source. This study was not supported by any external sources of funding.
Data Availability Statement. The data that support the findings of this study are openly available in repository Mendeley Data at https://data.mendeley.com/datasets/2z3dmg4y5f