Coagulation dysfunction in ICU patients with coronavirus disease 2019 in Wuhan, China: a retrospective observational study of 75 fatal cases

Coagulation dysfunction in critically ill patients with coronavirus disease 2019 (COVID-19) has not been well described, and the efficacy of anticoagulant therapy is unclear. In this study, we retrospectively reviewed 75 fatal COVID-19 cases who were admitted to the intensive care unit at Jinyintan Hospital (Wuhan, China). The median age of the cases was 67 (62–74) years, and 47 (62.7%) were male. Fifty patients (66.7%) were diagnosed with disseminated intra-vascular coagulation. Approximately 90% of patients had elevated D-dimer and fibrinogen degradation products, which decreased continuously after anticoagulant treatment and was accompanied by elevated albumin (all P<0.05). The median survival time of patients treated with anticoagulant was 9.0 (6.0–14.0) days compared with 7.0 (3.0–10.0) days in patients without anticoagulant therapy (P=0.008). After anticoagulation treatment, C-reactive protein levels decreased (P=0.004), as did high-sensitivity troponin (P=0.018), lactate dehydrogenase (P<0.001), and hydroxybutyrate dehydrogenase (P<0.001). In conclusion, coagulation disorders were widespread among fatal COVID-19 cases. Anticoagulant treatment partially improved hypercoagulability, prolonged median survival time, and may have postponed inflammatory processes and cardiac injury.


INTRODUCTION
In December 2019, a pneumonia outbreak of unknown origin was found in Wuhan and quickly spread to more than 100 countries [1]. Pathogen analysis confirmed a novel enveloped RNA beta-coronavirus [2], which was named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The World Health Organization defined coronavirus disease 2019 (COVID-19) as a public health emergency of international concern. As of October 4, 2020, there have been 34,804,348 confirmed cases, including 1,030,738 deaths [3]. Recent research has shown that COVID-19 can not only cause pneumonia but also damage other organs such as the heart, liver, kidneys, and coagulation and immune system [4][5][6]. Patients suffered from critical illness often die from respiratory failure, acute respiratory distress syndrome (ARDS), shock, disseminated intra-vascular coagulation (DIC), acute renal failure, heart failure, and multiple organ dysfunction syndrome (MODS) [5,7]. Therefore, it is particularly important to protect the lungs and other organs with treatment. According to clinical observations, we have found that the coagulation dysfunction of critical COVID-19 patients is easily induced by SARS-CoV-2. A pathological report of three COVID-19 cases by minimally invasive autopsies revealed the formation of hyaline thrombus in small vessels in both lungs and extrapulmonary organs [8]. Additionally, a recent study of risk factors associated with ARDS and death in COVID-19 patients confirmed that elevated coagulation functionrelated indicators (PT and D-dimer) were significantly associated with higher risk of developing ARDS [9]. Several previous studies have described the SARS-CoV-2 genome and epidemiological characteristics of COVID-19 patients [4,10]. Patients with critical illness were characterized by rapidly progressive pneumonia, respiratory failure, and poor outcomes. COVID-19 associated coagulation dysfunction is gaining attention. A recent review proposed using low molecular weight heparin (LMWH) anticoagulant therapy for patients with severe and critical illness, although there is no clear data to confirm its efficacy [11]. This study describes the clinical and laboratory characteristics of 75 COVID-19 patients admitted to the ICU at Jinyintan Hospital of Wuhan in Hubei Province and eventually died. We also analyzed the dynamic changes of coagulation function, inflammation, and cardiac injury, and evaluated the efficacy of anticoagulation therapy in these 75 patients.

Demographic and clinical characteristics of patients
All 75 patients were confirmed to have SARS-CoV-2 infections (COVID-19, critical type) and eventually died at the Jinyintan Hospital, one of the designated hospitals for COVID-19 patients. All patients were admitted to intensive care units (ICUs) between January 20, 2020 and February 26, 2020 and died before March 10, 2020. 1%]) of the cohort had CRP concentrations exceeding 160 mg/L; additionally, the median IL-6 concentration was 11.9 pg/mL (IQR: 8.8-20.2). Renal insufficiency was detected in 11 patients (15.1%) with serum creatinine (Cr) levels exceeding 133 μmol/L. Myocardial injury markers were abnormal in 53 patients (74.6%), including increased hsTNI levels ( Table  1). Most patients had abnormal coagulation tests ( Figure  1), including higher concentrations of D-dimer (89%) and FDP (90.8%), prolonged PTs (59.7%) and decreased PTA (58.3%). Two-thirds of patients (50/75) met the DIC diagnostic criteria of JAAM (Table 1). Compared with patients who survived >7 days in the ICU, the D-dimer (P<0.05) and FDP (P<0.01) levels of patients who survived <7 days gradually increased after ICU admission ( Figure 2A, 2B), and remained at high levels 5 days before coagulation disorders were widespread among fatal COVID-19 cases. Anticoagulant treatment partially improved hypercoagulability, prolonged median survival time, and may have postponed inflammatory processes and cardiac injury.  death (all P<0.01) ( Figure 2C, 2D, all P values can be found in Supplementary Table 1).

Baseline characteristics of patients with and without anticoagulation therapy
Thirty-five patients (46.7%) who received heparin within 3 days of ICU admission were included in the anticoagulant group; the remaining 40 patients (53.3%) were included in the non-anticoagulant group ( Table 2). According to the baseline characteristics of patients in the two groups, other than the sex ratio, survival time in ICU, and IL-6 level, there were no significant differences in age, co-existing diseases, hemocytology index, functional coagulation markers, inflammatory markers, and sequential organ failure assessment (SOFA) scores AGING between the two groups before anticoagulation treatment. Importantly, the median survival time in the anticoagulant group was longer than in the nonanticoagulant group (9.0 [IQR: 6.0-14.0] days vs. 7.0 [IQR: 3.0-10.0] days) ( Table 2).

Dynamic changes in coagulation, inflammation, and cardiac injury markers within 5 days after anticoagulation treatment
Next, we analyzed the dynamic changes in coagulation markers through a 5-day period in response to anticoagulant treatment compared with the nonanticoagulant group. The results showed that the Ddimer concentration was significantly and continuously deceased after heparin use; conversely, the D-dimer concentration was continuously increased during the 5 days in patients of the non-anticoagulant group (P=0.007) ( Figure 3A). We also observed significantly decreased FDP (P<0.001) and AT-III (P=0.001), and increased PTA (P=0.022) and ALB (P<0.001) in the anticoagulant group ( Figure 3B, 3D, 3F, 3G). There were no obvious dynamic differences in PT, APTT, or PLT (all P>0.05) ( Figure 3C, 3E, 3H, all P values can be found in Supplementary Table 2).
We individually evaluated the dynamic changes of inflammatory markers with D-dimer within 5 days. For the anticoagulant group, we observed a dynamic decrease over a 5-day period in the concentration of CRP with decreased D-dimer level, while for nonanticoagulant group, the concentration remained at a high level (P=0.004) ( Figure 4A, 4B). There were no significant dynamic changes in serum IL-6, PCT, lymphocyte, or eosinophil levels in patients with or without anticoagulation therapy ( Figure 4C-4J).

DISCUSSION
In this observational study, we report the clinical and laboratory characteristics of 75 patients who died in ICUs from COVID-19. All patients were seriously ill (critical type) at admission, and the condition of some patients deteriorated rapidly, suggesting SARS-CoV-2 infection may lead to poor outcomes in critically ill patients. It was remarkable that different degrees of coagulation dysfunction could be observed in COVID-19 patients, and it was particularly significant in critical type ICU patients. SARS-CoV-2 infects human cells via angiotensin-converting enzyme 2 (ACE2) [12]. ACE2 is expressed in alveolar epithelial cells, vascular endothelial cells, and the immune system at different levels [13]. SARS-CoV-2 can be rapidly recognized after entering the body, which activates the innate immune system to clear the virus; however, excessive activation can cause a cytokine storm, damage the microvasculature by direct and indirect means [14], activate the coagulation system, and inhibit fibrinolysis and the anticoagulation system. The resulting extensive thrombosis in microvessels often leads to poor outcomes [15].

AGING
Studies have confirmed the high risk of thrombosis in COVID-19 patients [16,17], and it has been reported that approximately half of COVID-19 patients have elevated D-dimer levels during disease progression; Ddimer levels have also been found to be significantly higher in patients with severe illness [18]. Increased D-dimer levels have become an independent risk factor for death in COVID-19 patients [19]. A similar phenomenon was demonstrated in our study, in which almost all the deceased patients showed coagulation dysfunction during the course of the disease, especially the high levels of D-dimer. Two-thirds of the patients   met diagnostic criteria for DIC (JAAM), and approximately 90% of the patients had D-dimer levels >1.5 µg/mL when they were admitted to the ICU. A concurrent study confirmed that 25% (20/81) of severe patients underwent venous thromboembolism (VTE) during hospitalization, demonstrating that 1.5 µg/mL is an appropriate cut-off value to reflect the high prevalence of thrombosis in COVID-19 patients [20]. Additionally, COVID-19-related coagulation dysfunction is a dynamically changing process. We noticed that the dynamic changes in coagulation function after ICU admission, especially the continuously elevated Ddimer and FDP levels, may be associated with reduced survival time in ICU. Early identifying them and continuously monitoring the trend will predict clinical prognoses.

AGING
In total, 35/75 (46.7%) patients in this study received anticoagulant therapy (LMWH or enoxaparin) within 3 days of ICU and admission, followed by 5 days of dynamic monitoring. Compared with patients in the nonanticoagulant group, the median survival time in ICU was significantly longer for the anticoagulant group, although all patients in both groups eventually died. Additionally, we found that the dynamic changes in coagulation markers such as D-dimer, FDP, and PTA    AGING in the anticoagulant group were partially improved compared with the non-anticoagulation group. ALB has been reported to have significant anticoagulant action in vitro [21] and was negatively related to the risk of thrombosis [22,23]. In this study, we found that ALB gradually increased after anticoagulation treatment. These data showed that anticoagulant treatment effectively relieved hypercoagulability in COVID-19 patients.
Classically, the association between coagulation and inflammation has been regarded as a crosstalk process [24,25]. During inflammatory reactions, inflammation mediators are released, which activate blood coagulation and consume mass clotting factors through the 'waterfall sample cascade,' which may lead to blood coagulation disorders [26]. Meanwhile, some key components of the coagulation system can promote inflammation through direct and indirect mechanisms, such as tissue factor and fibrinogen, which are not only key in the coagulation process, but also have multiple roles in tissue damage and inflammation [25,27]. A recent study reported the vital role of inflammation in COVID-19 progression [28,29].
Here, we analyzed the dynamic changes of some inflammatory markers after anticoagulant treatment and found improved CRP expression after anticoagulant therapy, which is a frequent prognostic factor for COVID-19 that reflects the inflammatory process [30]. Moreover, the dynamic improvement of cardiac injury indictors such as hsTNI, LDH, and HBDH were further demonstrated. As the most common complication of COVID-19, cardiac injury shows viral load in the myocardium and is closely related to regional and systemic inflammatory states [31,32]. Anticoagulant therapy could relieve hypercoagulability and prevent and improve the formation of systemic microthrombi, including coronary microvascular thrombosis [33,34]. However, this recovery did not reverse the outcome of patients in this study, who had severe multiple organ failure including respiratory and other organ dysfunctions, although our results showed partial improvement in inflammation and heart damage. Dynamically monitoring levels of D-dimer and indicators of inflammation and cardiac injury could assess the efficacy of anticoagulant therapy and severity of systemic disease status. After an accurate thrombosis risk assessment, more aggressive anticoagulation strategies may be needed in early rather than in late disease stages to improve outcomes.
This study had some limitations and raises areas for further study. One of the limitations of the study was the small sample size. Interpretations of our findings might be limited due to its retrospective nature with the possible loss of data. To overcome this limitation, a prospective study design and complete data collection would be needed. Additionally, at the beginning of the epidemic, due to the serious shortage of medical resources and staff, dynamic monitoring of patient conditions was insufficient, and some patients with coagulation disorders could not receive comprehensive screening, such as vascular ultrasound and CT, according to our data at the time.

CONCLUSIONS
Coagulation disorders were widespread in critical COVID-19 patients in ICUs. According to our data, twothirds of fatal patients were diagnosed with DIC upon ICU admission. In critically ill patients, anticoagulant treatment partially improved hypercoagulability, prolonged median ICU survival time, and potentially postponed inflammation and cardiac injury.

Study design and data collection
This retrospective study included 75 patients (≥18-yearsold) who were admitted to the ICU at Jinyintan Hospital (Wuhan, China) with SARS-CoV-2 infections between January 20 and February 26, 2020 and died before March 10, 2020. All patients were diagnosed with COVID-19 (critical type) according to The WHO interim guidance and Chinese management guidelines for COVID-19 (version 6.0) [35,36]. Patients' epidemiology, demographics, clinical characteristics, laboratory and treatment data were obtained from the standard electronic medical record system. All data were collated by two researchers, and then checked and confirmed by two physicians. This study was approved by the Research Ethics Commission of Jinyintan Hospital (KY-2020-56.01).

Laboratory procedures
The

AUTHOR CONTRIBUTIONS
JRS, XZ, WZ, ZHQ and LS had the idea for and designed the study, JRS wrote the main manuscript text, WZ, ZHQ and LS analyzed the data, MZ, LJ, BS, LK and YZ collected the data, XZ, XXW and DYZ revised the manuscript and gave final approval for the version to be published.

Supplementary Tables
Supplementary Table 1