Association between loss of hypercoagulable phenotype, clinical features and complement pathway consumption in COVID-19

Background Coronavirus disease 2019 (COVID-19) features a hypercoagulable state, but therapeutic anticoagulation effectiveness varies with disease severity. We aimed to evaluate the dynamics of the coagulation profile and its association with COVID-19 severity, outcomes, and biomarker trajectories. Methods This multicenter, prospective, observational study included patients with COVID-19 requiring respiratory support. Rotational thromboelastometry findings were evaluated for coagulation and fibrinolysis status. Hypercoagulable status was defined as supranormal range of maximum clot elasticity in an external pathway. Longitudinal laboratory parameters were collected to characterize the coagulation phenotype. Results Of 166 patients, 90 (54%) were severely ill at inclusion (invasive mechanical ventilation, 84; extracorporeal membrane oxygenation, 6). Higher maximum elasticity (P=0.02) and lower maximum lysis in the external pathway (P=0.03) were observed in severely ill patients compared with the corresponding values in patients on non-invasive oxygen supplementation. Hypercoagulability components correlated with platelet and fibrinogen levels. Hypercoagulable phenotype was associated with favorable outcomes in severely ill patients, while normocoagulable phenotype was not (median time to recovery, 15 days vs. 27 days, P=0.002), but no significant association was observed in moderately ill patients. In patients with severe COVID-19, lower initial C3, minimum C3, CH50, and greater changes in CH50 were associated with the normocoagulable phenotype. Changes in complement components correlated with dynamics of coagulation markers, hematocrit, and alveolar injury markers. Conclusions While hypercoagulable states become more evident with increasing severity of respiratory disease in patients with COVID-19, normocoagulable phenotype is associated with triggered by alternative pathway activation and poor outcomes.


Introduction
Coronavirus disease 2019 (COVID-19) triggered a global pandemic by causing primarily, respiratory distress occasionally escalating into respiratory failure, thereby posing significant public health challenges worldwide (1).In this context, coagulopathy, characterized by hypercoagulation is identified as one of the key pathogenic factors in COVID-19 (2,3).Associations of COVID-19 with macro-and microthrombosis have been established (4)(5)(6), which have spurred numerous clinical trials aiming to mitigate coagulopathy induced by this disease (7).Despite these efforts, previous studies did not consistently demonstrate the efficacy of therapeutic anticoagulation against COVID-19, irrespective of disease severity (8).While the effectiveness of therapeutic anticoagulation has been explored in moderately and critically ill patients with COVID-19 (9,10), the contrasting results from these trials suggest that therapeutic anticoagulation might be more effective in non-critically ill patients than in their critically ill counterparts.This discrepancy could stem from the evolving nature of coagulopathy as the disease progresses.Hence, understanding how coagulopathy varies with disease severity is essential for optimizing therapeutic interventions aimed at controlling coagulation status.
This study aimed to investigate the changes in coagulation characteristics with increasing severity of COVID-19.

Study population and setting
This prospective observational study conducted at 14 centers in Japan, from March 2021 to March 2022 included individuals aged ≥18 years, who were hospitalized for COVID-19 pneumonia and required oxygenation.COVID-19 was diagnosed based on findings from lung imaging and a positive severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) polymerase chain reaction or antigen test.Patients were classified into four groups based on COVID-19 severity at the time of coagulation status assessment according to the WHO ordinal scale: moderately ill patients requiring supplemental oxygen, patients on non-invasive respiratory support, patients on mechanical ventilation, and patients on extracorporeal membrane oxygenation (11).Exclusion criteria included known coagulopathy, previous oral anticoagulants administration, hematologic malignancy, and thrombocytopenia (detailed criteria listed in the Supplementary Methods in Additional File 1).Patients provided pre-enrollment informed consent.The study was approved by Nagoya University Hospital Institutional Review Board (Approval No. 2020-0548).

Evaluation of coagulation profile using thromboelastometry
In addition to clinical laboratory coagulation function tests, thromboelastometry with ROTEM sigma ™ was used to evaluate the coagulation profile of participants on the day of inclusion.The maximum clot elasticity (MCE) of each component was calculated from maximum clot firmness (MCF) using the previously described: MCE = (100×MCF)/(100-MCF) (12).The platelet component of MCE is evaluated using the difference between MCE EXTEM and FIBTEM.A hypercoagulable state was defined as the MCE of the external pathway exceeding the normal reference value (13).Figure 1 illustrates responentable image of EXTEM components of the thromboelastometry test for patients with normal and hypercoagulable states.A comprehensive description of the thromboelastometry parameters can be found in the previous publicaion (14).

Outcomes and clinical data
The primary endpoint was the time to recovery, with the recovery day defined as the first day during the 28 days after enrollment, on which patients did not require any respiratory support (11).Secondary outcomes included ventilator-free days within the 28 days and in-hospital mortality.Other recorded clinical data included demographics, comorbidity, activities of daily living (15), history of COVID-19 vaccinations, respiratory mechanics, laboratory values at the time of thromboelastometry evaluation, and treatments.In patients requiring mechanical ventilation at the Nagoya University Hospital, daily laboratory results within the initial 7 days from intubation to ICU discharge were recorded to assess the association between biomarker dynamics.The collected longitudinal data were analyzed to identify the initial, the highest, and lowest value, and calculate the maximum changes(DC3 = lowest C3/initial C3 ×100%).These routine lab tests were conducted by the central laboratory, with the specific testing methods and reagents detailed in Table E1 in the Supplementary Material.

Viral load assessment and genotyping of SARS-CoV-2
To explore the relationship with coagulation profiles, plasma viral loads were analyzed using quantitative reverse-transcription polymerase chain reaction.To investigate the relationship between coagulation profiles and viral genotypes in SARS-CoV-2, we performed whole-genome sequencing using stored specimens from the lower respiratory tract.

Statistical analysis
Spearman's rank correlation test was used for statistical analysis, focusing on assessing the relationships between thromboelastometry parameters and clinical laboratory tests, such as coagulation function tests, using Spearman's rank correlation test.The results were visualized using a heatmap.This method was also employed to examine longitudinal associations between complements and laboratory biomarkers.To assess differences in thromboelastometry results by severity, the Kruskal-Wallis test was used.P values for pairwise comparisons were adjusted using the Bonferroni method.The association between hypercoagulable state and baseline respiratory and laboratory parameters was evaluated using the Mann-Whitney U test.To decrease type-I error of multiple comparisons for 29 variables, the Benjamini-Hochberg procedure with a false discovery rate of 0.05 was applied to adjust the P value.Time to recovery was evaluated using the log-rank test, with a competing event of death regarded as "not recovered" on the last observation day, following an approach similar to the Fine-Grey method (11,16).A P value less than 0.05 was considered statistically significant.All statistical analyses were performed using R software (version 4.2.2) and RStudio.

Characteristics of patients and outcomes
Of the 192 cases identified during the observation period, 166 were finally included in the study (Figure E1  1).Among the enrolled patients, 76 had moderate respiratory failure at the time of inclusion (conventional oxygen supplementation, 62; non-invasive respiratory support, 14), and 90 had severe respiratory failure (mechanical ventilation, 84; extracorporeal membrane oxygenation, 6) (Table 1).The most common comorbidities were hypertension (40%) and diabetes mellitus (28%).Prior to disease onset, 159 (96%) patients could independently perform activities of daily living.Of the enrolled patients, 26 (16%) had received two doses of SARS-CoV-2 vaccine before disease onset.Unfractionated heparin was more commonly administered in severe cases, 88 (98%), than in moderate ones, 34 (45%), and was mainly used for therapeutic dose titration.Among patients with moderate respiratory failure at the time of inclusion, 10 (13%) experienced disease progression.Of the patients who received mechanical ventilation, 59 (66%) recovered after 28 days, 11 (12%) died, and 20 (22%) became ventilator dependent.

Hypercoagulable state in patients with severe COVID-19 is associated with favorable prognosis
Figure 5 illustrates the association between hypercoagulable status and clinical outcomes.No correlation was observed between hypercoagulable state and the disease course in moderate COVID-  6).

Association between coagulation profile and complement dynamics
Figure 7 illustrates the relationship between coagulation profiles and the dynamics of complement components.Complement C3, in addition to the initial assessment, had a significantly lower minimum value in the normocoagulable group (median [IQR], 85.9 [74.8-94.1]mg/dL vs. 77.3[51.6-86.6]mg/dL, P = 0.016) (Figure 7A).Meanwhile, any significant difference in the dynamics of C4 was not found between the two coagulation profile groups.Total complement activity was significantly reduced in patients in

Discussion
In this study, we evaluated the coagulation profiles of patients with different COVID-19 severities and found that while hypercoagulation was common with worsening respiratory failure, absence of hypercoagulation was associated with poor prognosis in severely ill patients.Platelet components were mainly involved in changes in blood viscoelasticity.Furthermore, the absence of hypercoagulation was also associated with coagulopathy, with complement consumption initiated by the alternative pathway activation.
Hypercoagulable state is considered a risk for COVID-19 macromicro thrombosis, and treatment aimed at controlling hypercoagulation has been well studied so far (17).However, the efficacy of therapeutic anticoagulation and antiplatelet therapy has not been illustrated in critically ill patients (18).Our study depicted that hypercoagulable status was associated with rather favorable outcomes, which might be one of the reasons that previous randomized control trials failed to show the efficacy of anticoagulation in critically ill patients with COVID-19 (7)(8)(9)(10).The hypercoagulable state is known to be influenced primarily by changes in platelet component (12), and individualized anticoagulation therapies to control the component of hypercoagulability in critically ill patients may be promising (19,20).We demonstrated that thromboelastometry may be a simple and effective way to stratify severity and to identify different phenotypes of coagulation status.In this study, therapeutic anticoagulation and anti-inflammatory therapy were the standard of care in most critically ill patients.This may have partially led to a favorable outcome in patients with coagulable phenotype compared to the outcome in patients with normocoagulable phenotype.Whether phenotypic differences in coagulation profiles can help in individual treatment selection, such as anticoagulation therapy, needs further evaluation.
Our study showed that complement consumption was associated with normocoagulable phenotype and with poor prognosis in mechanically ventilated patients with COVID-19.Lower C3 but not C4 observed in the normocoagulable group at the initial assessment is consistent with similar results of previous in vivo studies that reported that SARS-CoV-2 triggers the alternative pathway overactivation (21).Previous studies reported indirect activation of the alternative pathway in addition to direct activation by viral components (22): viral infection of the alveolar epithelium activates the alternative pathway via the Janus kinase/ signal transducer and activator of transcription pathway (23).In our study, elevated KL-6, a biomarker of type II alveolar cells, was associated with decreased C3, and there was no significant association between the viral load and decreased C3.Based on these findings, we speculate that the host response, via type 2 alveolar epithelial cell injury rather than pathogen factors, leads to overactivation of the alternative pathway, which results in phenotypic changes in coagulation phenotype and poor prognosis.The phenotypic difference may be triggered by host heterogeneity such as genetic variants in C3 genes reported in a previous study (24).We also showed that complement consumptions were correlated with reduction in coagulation components.These findings are also consistent with those of a recent randomized trial, which reported that target C3 blockage suppressed thrombin production (25).In contrast to the findings of a previous study (26), we observed that the alternative pathway consumption was associated with consumptive coagulopathy but not with hypercoagulable profile.Higher levels of D-dimer and fibrin degradation product in relation to complement dynamics, together with the findings that maximum lysis decreases as the severity increases, may imply the activity of microthrombosis.Our findings are further supported by ex vivo experiments on virusinduced senescence, which demonstrate that the secretome enhances coagulation, promotes the formation of the lytic complement complex C5b-C9, and stimulates the formation of Neutrophil Extracellular Traps (27, 28).Additionally, a rise in C3a enhances the differentiation of CD16 positive cytotoxic T cells in COVID-19 (29).These underlying processes are implicated in causing lung damage from COVID-19 and result in delayed recovery.The higher maximum soluble thrombomodulin and lower minimum hematocrit may suggest the feature of thrombotic microangiopathy (30,31).Further studies are warranted to understand the control overactivation of the alternative pathway, especially in high-risk phenotypes.
The strength of this study was the standardized introduction of evidence-based treatments in many patients from multiple hospitals, which was useful in visualizing uncontrolled pathologies in a real-world setting.However, this study had some limitations.First, as the current cohort was a predominantly unvaccinated population, extrapolation to a post-vaccinated population may not be possible.Second, considering that the SARS-CoV-2 outbreak in the region was sporadic and the local health office conducted triage of hospitalized patients during the outbreak, selection bias may have occurred in hospitalized patients.Lastly, the present study is a hypothesis-generating analysis, and further investigation is needed to validate the stratification of severity by coagulation profile and the possibility of individualizing treatment for each phenotype.
In summary, the coagulation profile changes according to the severity of respiratory failure in COVID-19.As per our findings, normocoagulable status characterized by C3 consumption is associated with delayed recovery in patients with severe COVID-19.These findings support the heterogeneous treatment effect of anticoagulation depending on COVID-19 severity.Further studies are warranted to develop phenotype-specific treatments based on the coagulation profiles.

FIGURE 1
FIGURE 1 Representative image of thromboelastometry test.The figure shows the output of the EXTEM component of thromboelastometry, illustrating normal coagulation (A) and hypercoagulable states in patient with COVID-19 patients (B).The double arrow indicates maximum clot firmness.

5 FIGURE 6
FIGURE 5 Association between coagulation profile and outcomes.Association between (A) coagulation profile and time to recovery in moderate COVID-19 cases, (B) coagulation profile and time to recovery in severe COVID-19 cases, and (C) coagulation profile and ventilator-free days in severe COVID-19 cases.P values calculated from the Mann-Whitney U test for pairwise comparison and adjusted using the Bonferroni method.P values are displayed if P <0.05.