Introduction

Severe and extremely severe coronavirus disease (COVID-19) is often accompanied by multiple organ failure. This process is also a result of hypercoagulation and endotheliopathy as important aspects in pathogenesis of this disease. Treatment of severe COVID-19 should be multicomponent and carried out in the intensive care unit (ICU). Combination of baseline coagulopathy and endotheliopathy with common use of muscle relaxants and prolonged immobilization increase the risk of thrombotic complications in these patients [1, 2]. Most reports are devoted to venous thrombotic complications. However, arterial thrombosis has also been reported in COVID-19 [3].

Considering certain data on pathogenesis, many experts recommend heparins in various doses for prevention of thrombotic events in high-risk patients with COVID-19. Nevertheless, there is still no high evidence level for this approach [4, 5]. High doses of anticoagulants also have a downside, i.e. hemorrhagic complications. The last ones have not been studied enough so far. Clinical characteristics of patients with hemorrhagic complications have not been described and should be the subject of further research.

Considering the ongoing pandemic, we need to understand the causes, incidence, prevention options and impact on the outcomes of thrombotic and hemorrhagic complications associated with coagulopathy in COVID-19.

The purpose of the study was to analyze the incidence and structure of thrombotic and hemorrhagic complications in ICU patients with severe and extremely severe coronavirus disease 2019 (COVID-19).

Material and methods

A retrospective single-center study included all patients with severe and extremely severe COVID-19 who admitted to the ICU of the Federal Scientific and Clinical Center for Specialized Types of Medical Care and Medical Technologies from April 06 to July 01, 2020 and from October 5, 2020 to February 21, 2021. For these periods, this center was converted into infectious disease hospital for the treatment of patients with COVID-19. There were no exclusion criteria. Diagnosis of a new coronavirus infection, assessment of severity and treatment were carried out in accordance with the clinical guidelines of the Ministry of Health of the Russian Federation "Prevention, diagnosis and treatment of a new coronavirus infection (COVID-19)" [6].

Primary endpoint was any thrombotic complication (pulmonary embolism (PE), deep and superficial vein thrombosis, peripheral and mesenteric artery thrombosis, myocardial infarction, stroke).

Secondary endpoint was any hemorrhagic complication. We analyzed the length of ICU- and hospital-stay, as well as survival in patients with thrombotic and hemorrhagic complications.

Demographic data, comorbidities, clinical characteristics, methods of therapy and laboratory data (complete blood count, complete urinalysis, biochemical blood test, and blood coagulation test) were considered upon admission to the ICU. The von Willebrand factor antigen was identified and studied in 83 patients. Thromboelastography parameters were measured according to standard technique (TEG 5000, Haemoscope Corporation, USA). In patients with hemorrhagic complications, we additionally analyzed laboratory parameters and anticoagulation therapy. Clinical severity was assessed using the SOFA scale (Sequential Organ Failure Assessment). Severity of delirium was analyzed using the CAM-ICU method (Confusion assessment method in intensive care unit).

We applied low molecular weight heparins (LMWH) for prevention of thrombotic complications (calcium nadroparin 86 anti-Xa IU/kg 2 times a day subcutaneously, enoxaparin sodium 100 anti-Xa IU/kg 2 times a day subcutaneously, parnaparin 6400 anti-Xa ME 2 times a day subcutaneously). In patients with creatinine clearance (CC) < 30 ml/min, but more than 15 ml/min, we administered enoxaparin sodium 100 anti-Xa IU/kg 1 time per day subcutaneously with control of anti-Xa activity (target values 0.6-1.0). In patients with CC < 15 ml/min, infusion of unfractionated heparin (UFH) was used with control of activated partial thromboplastin time (aPTT) within 45–60 s. To exclude thrombosis of lower limb veins, we performed ultrasound of these veins in all patients (Vivid 7 pro scanner, General Electric, USA) at the first day after admission to the ICU and then every 7 days. If PE was suspected, contrast-enhanced chest CT was carried out. Major hemorrhagic events were defined according to the criteria of the International Society on Thrombosis and Hemostasis (ISTH) [7].

Local ethics committee approved the study protocol (protocol No. 5 dated June 3, 2020).

Statistical analysis was performed using the SPSS software (version 28.0.0.0, IBM, USA). Data are presented as absolute values (percentage), mean (±standard deviation) or median with 25^th and 75^th percentiles depending on the type and distribution of data. Distribution normality was tested by the Kolmogorov–Smirnov test. Between-group differences in normally distributed data were analyzed using the Mann–Whitney U-test, χ² test and Fisher's exact test. The odds ratio (OR) and 95% confidence interval (CI) were assessed. Differences were significant at two-sided p-value<0.05.

Results

The study included 442 adults with severe and extremely severe COVID-19. Thrombotic complications occurred in 87 (19.7%) out of 442 patients. Venous thrombotic complications developed in 53 (11.9%) patients (42 (9.5%) patients with deep and superficial vein thrombosis, 11 (2.4%) patients with PE). Arterial thrombotic events occurred in 34 (7.7%) cases (ischemic stroke — 19 (4.3%), myocardial infarction — 11 (2.5%), peripheral/mesenteric thrombosis — 4 (0.9%) patients). Structure of thrombotic complications is shown in Fig. 1. Demographic data and clinical characteristics of patients with and without thrombotic complications are presented in Table 1.

Fig. 1. Structure of thrombotic complications (n=87).

Table 1. Clinical characteristics and outcomes in patients with and without thrombotic complications

Note. Data are presented as median and percentiles [0.25; 0.75], absolute (n) and relative (%) frequencies. CT — computed tomography; ICU — intensive care unit; CAD — coronary artery disease; CKD — chronic kidney disease; SOFA — The Sequential Organ Failure Assessment.

There were significant differences in age between patients with and without thrombosis (69.5 [58; 78] and 72 [60; 82] years, respectively, p = 0.043). Nevertheless, both categories of patients are in the same age group.

Mechanical ventilation (78.2 and 64.0%, p=0.012), injections of neuromuscular relaxants (64.3 and 47.3%, p=0.005) and norepinephrine (74.7 and 58%, p=0.005) for more than a day were significantly more common in patients with thrombotic complications compared to those without similar events (Table 1). Delirium was also more common in patients with thrombosis (42 (48.3%) and 114 (32.1%) cases, respectively; p=0.005). Incidence of hemorrhagic complications in patients with and without thrombosis was similar.

Patients with thrombotic complications had significantly higher concentrations of ferritin (p=0.026), C-reactive protein (p=0.004) and troponin T (p=0.012) at admission to the ICU (Table 2). There was a tendency to higher level of von Willebrand factor antigen in patients with thrombotic complications (415 [362; 692] and 367 [272; 457], respectively; p=0.07).

Thromboelastography data are available in 37 patients (Table 2). In patients with thrombotic complications, thromboelastography data demonstrated a hypercoagulability profile with decrease of reaction time (R) and excess of the upper limit of maximum amplitude (MA). However, we found no significant between-group differences for any of these parameters (Table 2).

Table 2. Laboratory data among patients with and without thrombotic complications

Note. Data are presented as median and percentiles [0.25; 0.75]. R — reaction time; k — kinetic time; MA — maximum amplitude.

Patients with thrombotic complications were characterized by significantly longer ICU-stay (11 [7; 23] and 6 [3; 11] days; p=0.001) and hospital-stay (21 [14; 32] and 17 [11; 23] days; p=0.001) compared to those without thrombotic complications. Survival rates were similar (30 (34.5%) and 165 (46.4%), respectively; p=0.058) (Table 1). The odds ratio (OR) of thrombotic complications in patients with delirium was 1.97 (95% CI 1.28–3.18; p=0.005). In case of norepinephrine infusion for more than a day, this value was 2.14 (95% CI 1.26 — 3.62; p=0.004), in neuromuscular relaxant infusion for more than a day — 2.01 (95% CI 1.24 — 3.27; p=0.004), in mechanical ventilation — 2.02 (95% CI 1.16-3.51; p=0.013).

Hemorrhagic complications developed in 23 (5.2%) patients including major events in 15 (65.2%) cases (Table 3). Gastrointestinal bleeding was the most common hemorrhagic complication (39.1%) (Fig. 2).

Table 3. Clinical characteristics and laboratory parameters in patients with hemorrhagic complications

Note. Data are presented as median and percentiles [0.25; 0.75], absolute (n) and relative (%) frequencies. ICU — intensive care unit; SOFA — The Sequential Organ Failure Assessment.

Fig. 2. Structure of hemorrhagic complications (n=23).

Bleedings occurred in 8 [5; 24] days after admission to the ICU. Age of patients was 78 [58; 83] years, 43.5% were men, body mass index was 27.4 [25; 31] kg/m2 (Table 3). The most common comorbidities in patients with hemorrhagic complications were arterial hypertension (56.5%), coronary artery disease (39.1%), and cerebrovascular disease (21.7%). Therapy of COVID-19 included glucocorticosteroids in 78.3% of patients, monoclonal antibodies to interleukin-6 receptor or interleukin-6 — in 43.5% of patients.

Patients with bleeding had multiple organ failure (SOFA score 4). It should be noted that 14 (60.8%) patients underwent mechanical ventilation, 6 (26%) — renal replacement therapy, 18 (78.2%) — norepinephrine infusion for more than a day. Laboratory findings included increase of serum D-dimer, aPTT prolongation and increased serum creatinine (upper normal limit). Survival of patients with hemorrhagic complications was 36.8%.

Discussion

In this retrospective study, the incidence of thrombotic complications in ICU patients with severe and extremely severe COVID-19 was 19.7%. Mechanical ventilation and norepinephrine infusion were more common in patients with thrombotic complications. Length of ICU- and hospital-stay was longer while survival was similar in both groups. Incidence of hemorrhagic complications was 5.2%. Thrombotic complications were less common in this study compared to other previous studies in China [8, 9] or Europe [2, 10, 11]. Prevention of thrombotic complications in these studies included anticoagulants in prophylactic doses.

In our study, lower limb vein thrombosis was the most common thrombotic complication. These findings are consistent with other studies not only in patients with COVID-19 [10], but also in other ICU patients [12]. Clinical diagnosis of lower extremity vein thrombosis is difficult in ICU patients. Nevertheless, Doppler ultrasound ensures timely and regular screening and diagnosis of these complications. A systematic review of 2,928 ICU patients found significantly higher incidence of venous thrombosis of the lower extremities in case of routine ultrasound compared to clinical diagnosis alone (56.3 vs. 11.0%, p<0.001) [13].

PE is another common potentially fatal venous thrombotic complication in patients with COVID-19 [14]. Incidence of PE is 12.6 — 29%. This complication is more common in ICU patients with COVID-19 compared to the hospital [13, 15]. Higher incidence of PE reported in the literature may be due to different local examination protocols. Some hospitals routinely perform CT angiography in all patients at admission [16]. In our study, CT angiography was performed only in patients with suspected PE and only in those ones who could be transported to the CT room. Moreover, diagnosis of pulmonary embolism is difficult in patients with acute respiratory distress syndrome due to baseline lung damage. It is not always possible to differentiate the causes of progressive respiratory failure in these patients [14].

In our study, incidence of arterial thrombosis was higher than in other studies. Helms J. et al. [11] reported only 4 (2.6%) complications among 150 ICU patients with COVID-19. Nevertheless, the authors emphasize that arterial thrombosis can lead to severe disability and poor outcomes [3, 17].

Pathogenesis of arterial thrombosis is unclear. There is a question whether the same mechanisms underlie venous and arterial thrombosis. Azouz E. et al. [3] revealed that arterial thrombi intraoperatively removed in patients with COVID-19 mainly consisted of platelets. These data can underlie the use of antiplatelet agents for prevention of arterial thrombotic complications, but further studies are needed to confirm this hypothesis.

We found similar survival in patients with and without thrombotic complications. However, ICU- and hospital-stay was longer in patients with thrombotic complications. Potential benefit of anticoagulation for COVID-19 is based on potential role of hypercoagulability syndrome and endothelial dysfunction in pathogenesis of disease and subsequent minor and major thrombosis [1]. One of the first studies devoted to effectiveness of different doses of anticoagulants in prevention of thrombotic events in patients with severe COVID-19 revealed that therapeutic anticoagulation reduced the incidence of thrombotic events and in-hospital mortality compared to standard prophylactic or intermediate doses [18]. Other studies have also demonstrated that higher doses of anticoagulants were associated with lower mortality in patients with COVID-19 [19, 20].

However, these data were not confirmed in three randomized trials [21], which were terminated early due to no effect of therapeutic doses of anticoagulants on the outcomes. The authors found no difference in survival of patients receiving therapeutic and prophylactic doses of anticoagulants (62.7 and 64.5%, respectively; OR 0.84; 95% CI 0.64–1.11). Moreover, there were no differences in the number of days without ventilation. There were fewer severe thrombotic events in case of therapeutic anticoagulation compared to standard doses of heparins (6.4% vs. 10.4%). Nevertheless, odds ratio of thrombotic events or mortality was similar in both groups (40.1 and 41.1%, respectively; OR 1.04; 95% CI 0.79-1.3). The need to increase the prophylactic dose should be carefully weighed against the risk of thrombotic and hemorrhagic complications in patients with severe and extremely severe COVID-19. In our study, hemorrhagic complications occurred in 23 (5.2%) patients. According to early studies [2, 9, 11], hemorrhagic events developed in 0-7.5% of patients with COVID-19, and this complication is poorly described regarding clinical characteristics, time of development and anatomical location. Al-Samkari H. et al. [15] analyzed 400 patients with severe COVID-19 and reported overall bleeding rate of 7.6% and major bleeding rate of 5.6%.

In our study, gastrointestinal bleeding was the most common event (9 (39.1%) out of 23 patients with hemorrhagic complications). Shah A. et al. [10] obtained similar results. The authors reported gastrointestinal bleeding in 6 (40%) out of 15 patients with hemorrhagic complications (3.2% of the total cohort). Hemorrhagic complications are well studied in critically ill patients without COVID-19 [22]. Risk factors of these complications were increased aPTT, thrombocytopenia, therapeutic doses of anticoagulants, antiplatelet therapy, renal replacement therapy and recent surgery. Certain drugs used for anticoagulation are not associated with bleeding. We also found hemostatic disturbances in patients with hemorrhagic complications (2-fold increase of serum D-dimer and aPTT prolongation). However, high risk of hemorrhagic complications persists in patients with COVID-19 and organ dysfunction even if laboratory signs of coagulopathy are absent [10]. In addition, such treatments as extracorporeal membrane oxygenation are accompanied by high risk of gastrointestinal bleeding and hemorrhagic stroke [23].

Pathogenesis of hemorrhagic complications in COVID-19 is unclear. Intrapulmonary microscopic hemorrhages prevail in patients with COVID-19. Perhaps, pulmonary intravascular coagulopathy is a trigger mechanism for systemic hemostasis disorders, disseminated intravascular coagulation and hemorrhages [24]. Apparently, virus characteristics also determine clinical manifestations of disease. Indeed, signs of coagulopathy and hemorrhagic complications were rare in infected patients during SARS-CoV-1 outbreak in 2002 [25]. On the contrary, incidence of bleeding was high in patients with Ebola fever followed by damage to liver and peripheral vessels [26]. Some authors suggest that hemorrhagic events in patients with severe COVID-19 receiving anticoagulation occur after regression of inflammation and decrease of serum acute phase proteins, primarily fibrinogen. Decrease of serum fibrinogen leads to restoration of resistance to heparin, increase of its concentration and higher risk of hemorrhagic complications [27]. In this regard, it is still unclear whether our results represent a true increase in bleeding risk due to COVID-19-associated immune mechanisms, anticoagulation or severity of disease and treatment approaches. Therefore, any potential benefit of anticoagulation beyond the general standard of care in patients with COVID-19 should be explored and evaluated in randomized trials.

Elevated C-reactive protein, ferritin and troponin determine severity of COVID-19 and are closely associated with inflammation and thrombosis [24]. It is unclear whether these high values in our study are a direct consequence of SARS-CoV-2 infection or secondary consequences of systemic inflammation and severity of disease. Considering a tendency to increase of von Willebrand factor antigen in patients with thrombosis, we emphasize the necessity of further study of the parameter in patients with COVID-19.

Thromboelastography can be essential to choose therapeutic strategy for major bleeding [28]. We found hypercoagulability profile in patients with thrombotic complications without significant between-group differences. Shah A. et al. [10] obtained similar results. However, they analyzed a small sample. In this regard, it is necessary to determine the value of thromboelastography for assessing the risk of thrombotic and hemorrhagic complications, determining the indications and choosing the anticoagulation mode for COVID-19 in large randomized trials.

Conclusion

We found a high incidence of thrombotic complications in patients with severe COVID-19 despite therapeutic anticoagulation. Venous thrombosis of the lower extremities prevails among venous thrombotic complications. Diagnosis of pulmonary embolism in ICU patients can be difficult. Therefore, the true incidence of pulmonary embolism is often underestimated. Arterial thrombosis complicates COVID-19, result disability and poor outcomes. Effectiveness of antiplatelet agents is still unclear. The risk of thrombotic complications is 2 times higher in norepinephrine infusion and injections of neuromuscular relaxants for more than a day, mechanical ventilation and development of delirium. Most hemorrhagic complications are clinically significant and occur predominantly in patients with severe organ failure. The causes of bleeding in patients with COVID-19 are unclear. Therefore, when prescribing anticoagulants, physicians should be wary due to the risk of hemorrhagic complications. Thromboelastography can be valuable for assessing the risk of thrombotic and hemorrhagic complications.

Author contribution:

Concept and design of the study — Bychinin M.V.

Collection and analysis of data — Avdonin P.V., Korshunov D.I.

Statistical analysis — Mandel I.A.

Writing the text — Bychinin M.V.

Editing — Klypa T.V., Mandel I.A.

Laboratory studies were carried out within the framework of the Russian Science Foundation grant No. 21-15-00441.

The authors declare no conflicts of interest.