key: cord-1020350-iqwmz3bu
authors: Len, Polina; Iskakova, Gaukhar; Sautbayeva, Zarina; Kussanova, Aigul; Tauekelova, Ainur T.; Sugralimova, Madina M.; Dautbaeva, Anar S.; Abdieva, Meruert M.; Ponomarev, Eugene D.; Tikhonov, Alexander; Bekbossynova, Makhabbat S.; Barteneva, Natasha S.
title: Meta-Analysis and Systematic Review of Coagulation Disbalances in COVID-19: 41 Studies and 17,601 Patients
date: 2022-03-11
journal: Front Cardiovasc Med
DOI: 10.3389/fcvm.2022.794092
sha: c9c6f953242393b0fbcbb12849e3f4caf7dd9ec4
doc_id: 1020350
cord_uid: iqwmz3bu

INTRODUCTION: Coagulation parameters are important determinants for COVID-19 infection. We conducted meta-analysis to assess the association between early hemostatic parameters and infection severity. METHODS: Electronic search was made for papers that addressed clinical characteristics of COVID-19 patients and disease severity. Results were filtered using exclusion and inclusion criteria and then pooled into a meta-analysis to estimate the standardized mean difference (SMD) with 95% confidence interval (CI) for D-dimers, fibrinogen, prothrombin time, platelet count (PLT), activated partial thromboplastin time. To explore the heterogeneity and robustness of our fundings, sensitivity and subgroup analyses were conducted. Publication bias was assessed with contour-enhanced funnel plots and Egger's test by linear regression. Coagulation parameters data from retrospective cohort study of 451 patients with COVID-19 at National Research Center for Cardiac Surgery were included in meta-analysis of published studies. RESULTS: Overall, 41 original studies (17,601 patients) on SARS-CoV-2 were included. For the two groups of patients, stratified by severity, we identified that D-dimers, fibrinogen, activated partial thromboplastin time, and prothrombin time were significantly higher in the severe group [SMD 0.6985 with 95%CI (0.5155; 0.8815); SMD 0.661 with 95%CI (0.3387; 0.9833); SMD 0.2683 with 95%CI (0.1357; 0.4009); SMD 0.284 with 95%CI (0.1472; 0.4208)]. In contrast, PLT was significantly lower in patients with more severe cases of COVID-19 [SMD −0.1684 with 95%CI (−0.2826; −0.0542)]. Neither the analysis by the leave-one-out method nor the influence diagnostic have identified studies that solely cause significant change in the effect size estimates. Subgroup analysis showed no significant difference between articles originated from different countries but revealed that severity assessment criteria might have influence over estimated effect sizes for platelets and D-dimers. Contour-enhanced funnel plots and the Egger's test for D-dimers and fibrinogen revealed significant asymmetry that might be a sign of publication bias. CONCLUSIONS: The hemostatic laboratory parameters, with exception of platelets, are significantly elevated in patients with severe COVID-19. The two variables with strongest association to disease severity were D-dimers and fibrinogen levels. Future research should aim outside conventional coagulation tests and include analysis of clotting formation and platelet/platelet progenitors characteristics.

In <2 years of the outbreak, the COVID-19 pandemic took almost five million lives (1). SARS-CoV-2 is the seventh member of the large coronavirus family capable of inducing human disease (2) . Viral respiratory infections, including severe respiratory acute syndrome coronavirus (SARS-CoV), Middle East respiratory syndrome coronavirus (MERS-CoV), SARS-CoV-2, may induce coagulopathy and lead to intravascular thrombi and deposition of fibrinogen (3) (4) (5) (6) (7) .

The coagulation system is activated and dysregulated during COVID-19 infection; however, characteristics of COVID-19 associated coagulopathy are different from other coagulation disorders (8) . Evidence of abnormal COVID-19associated coagulation parameters has already appeared in early publications from Wuhan (9) (10) (11) (12) and was supported by further publications worldwide [rev. (8, 13) ]. The reported hemostatic abnormalities with COVID-19 infection include increased D-dimers (DD) levels (14, 15) , fibrinogen levels (16, 17) , changes in the quantities and levels of activation of platelets (14, (18) (19) (20) , elevated von Willebrand factor (VWF) [ (21, 22) , rev. by Becker et al. (23) ] and other coagulation parameters. Persistently elevated risk of thromboembolic events, even with the initiation of prophylactic anti-coagulation, suggest the presence of hypofibrinolysis, in addition to detected hypercoagulability in COVID- 19 (24, 25) . It is also consistent with the results of thromboelastography (TEG) reported by Wright et al. (26) .

There is a need for early detection of elevated coagulation biomarkers to optimize risk stratification of patients with COVID-19. Importantly, multiple findings suggest that thromboembolic complications should be considered as a cause of clinical deterioration in severe COVID-19 infection and potentially beyond hospital discharge (27) (28) (29) . However, due to the limitations of early pandemic research with small study sizes and highly heterogeneous datasets, it is still debated.

The aim of the current study was to evaluate the validity of stratification based on early coagulation parameters provided by different clinical studies worldwide. We also aimed to use subgroup analysis to explore whether differences in patient subpopulations partly explained heterogeneity in results. The coagulation parameters of our own cohort were included in a systematic meta-analysis of published studies. Furthermore, we also discuss a complex mechanism of hemostatic dysregulation that fuels the coagulation cycle.

In this study, we aimed to identify the relationship between coagulation biomarkers taken at admission and severity of COVID-19 in adult (above 18) patients by estimating the effect size of five laboratory coagulation tests: D-dimers (DD), platelet count (PLT), fibrinogen (FIB), activated partial thromboplastin time (APTT), and prothrombin time (PT).

To broaden our search, we included Ovid, PubMed, Web of Sciences, and Google Scholar in the study for the period ending September 30th, 2021. They were thoroughly scanned using the following keywords: SARS-CoV-2, COVID-19, coagulation, severity, characteristics, features, D-dimers, platelets, fibrinogen, cohort, observational, retrospective. First, satisfying search results were exported to an Excel table, and duplicates were eliminated. Second, we conducted a preliminary review of retrieved articles' abstracts. Papers that focused on pregnant women, children, or specific age groups were avoided. Similarly, works that focused on the effect of medical treatment or patients with particular commodities were not included in the metaanalysis. We also excluded studies that compared mortality among groups of COVID-19 patients. While diagnostic and classification of COVID-19 severity follow official guidelines, the differentiation between whether COVID-19 caused deaths might be ambiguous. In order to avoid such bias, we only considered studies that grouped patients according to the severity of their symptoms. Then, each publication was checked according to the inclusion criteria: sample size, severity assessment criteria, presence of at least one coagulation marker of interest expressed as a continuous variable. To be included, a study should have had more than 15 patients and divided them according to the disease severity-WHO guidelines, intensive care unit (ICU) admission, disease aggravation, or need for oxygen therapy. Next, a list of biomarkers should have included at least some of the coagulation parameters: PLT, DD, FIB, APTT, PT, thrombin time (TT), Activated thrombin (AT), Factor VIII (FVIII), and VWF. Finally, the quality of selected articles was assessed with the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (30) (Supplementary Table 1 ). Questions 5, 6, 12, 13 were not applicable to selected studies due to their observational/retrospective nature; Question 8 was not applicable to studies with dichotomous exposure values, like ICU admission and need for oxygen therapy. Hence, studies with a quality score of <5 were excluded from the analysis. Search and quality check were conducted independently by two authors. Disagreements were resolved during a joint discussion with a third author.

The primary goal was to retrieve data from each article about two groups of patients differing in disease severity by: group size, stratification criteria, and results of the laboratory tests. Apart from these, we kept a record of the country, hospital, and admission time to identify and exclude duplicate patients.

From a large cohort of 560 patients consecutively admitted from June 2020 to August 2021 due to coronavirus infection to the cardiology department of the National Research Center for Cardiac Surgery, Nur-Sultan, Kazakhstan, we selected 451 patients with a confirmed diagnosis of COVID-19, aged ≥18 years, and those with the onset of the disease ≤21 days, all of whom were hospitalized for at least 24 h with COVID-19. All demographic, clinical, laboratory data were extracted from the electronic records of the NRCSC. Laboratory tests included PLT, PT, international normalized ratio, APTT, FIB, DD. Blood samples were collected at admission by the clinical team. Clotting tests were performed according to standard methods. For coagulation tests (PT, international normalized ratio, APTT, FIB), we used AVATUBE R with sodium citrate, and it was measured on Sysmex CS-2500 (Sysmex, Japan). Blood for DD was collected in AVATUBE R with sodium citrate, and D-dimers were quantified on Cobas R 6000. PLT counts were measured on Sysmex XS-500 (Sysmex, Japan) with blood collection performed in AVATUBE R with K2 EDTA. The confirmation of SARS-CoV-2 infection was done by real-time quantitative reversetranscription polymerase chain reaction (RT-PCR) assay on nose/throat swab or sputum samples using CFX96 R Real-Time System (Bio-Rad Laboratories, Inc., USA).

The COVID-19 population was initially divided into four groups according to disease severity following local guidelines "Diagnosis and treatment of Covid-19 in Adults" (http:// www.rcrz.kz/index.php/ru/2017-03-12-10-51-13/klinicheskieprotokoly) (Supplementary Data 1) . The study protocol was approved by the medical ethics committee of the National Research Center for Cardiac Surgery, Nur-Sultan, Kazakhstan waiving the need for written informed consent due to the retrospective design.

All data analysis was performed in the RStudio 1.4.1717 with R version 4.1.0 (Integrated Development Environment for R. RStudio, PBC, USA). In most articles, results were reported as the first, second (median), and third quartiles, reflecting that original data might be abnormally distributed. Since the authors of the primary studies do not report the underlying distributions, and there are no established methods to analyze the difference of medians, we performed data transformation according to the method described by McGrath et al. (31) . The workflow takes advantage of the formulas proposed by Luo et al. (32) and Wan et al. (33) for estimation of mean and standard deviation, respectively. Both methods assume the input data to be normally distributed, so the package also includes the preceding transformation step by the Box-Cox method. In the scope of this meta-analysis, we compare two groups: more and less severe. Patients classified as severe, critical, admitted to ICU, with observed aggravation of a disease, or in need of oxygen therapy were assigned to the first group. The second group included mild, moderate, outpatients, general ward patients, recovered and stable patients (Supplementary Table 2 ). If, for example, a primary study had more than two groups-mild and moderate, severe and critical-then mild patients were combined with moderate, and severe were combined with critical.

The "meta" and "metafor" packages were used for the metaanalysis. Heterogeneity was assessed using the I 2 , tau-squared, and prediction intervals. Following the recommendations of Veroniki et al. (34) , Paule-Mandel (PM) tau-squared estimator and Q-Profile (QP) tau-squared confidence interval estimator were used in this meta-analysis. To account for different measurement scales, the standardized mean difference (SMD), or Hedges g was estimated. Before pooling point estimates, each study was assigned a weight based on inverse variance approach-smaller variance gives higher precision, thus yielding greater weight for a study. For datasets with high heterogeneity, the Knapp-Hartung adjusted random-effects model is implemented. In case of low heterogeneity, the fixedeffects model is applied.

We also performed several sensitivity analyses using the "dmetar" package. First, we recalculated the effect size using the leaveone-out method. Second, basic outliers, studies whose confidence intervals do not overlap with the pooled effect confidence interval, were removed. Third, we ran the diagnostic for influential cases with Graphic Display of Study Heterogeneity (GOSH) plots and excluded them from the meta-analysis. To further explore the heterogeneity of the data, we conducted subgroup analyses: based on grouping criteria and location.

We used the counter-enhanced Funnel plots and Egger's test by linear regression to assess the possibility for publication bias in selected studies.

The stepwise process of selection is depicted in Figure 1 . Initially, there were 4,056 papers that matched search parameters. After the removal of duplicates and studies that did not meet the exclusion criteria, only 195 unique studies remained. A more thorough review based on the inclusion criteria has proven 125 studies to be unfitting. In addition to 70 that were left, 15 studies were extracted from the references lists of previously published meta-analyses and systematic reviews. Finally, we performed a quality check and removed duplicate patients, which resulted in 40 studies being included in the meta-analysis. Besides original articles, we have included data from the National Research Center for Cardiac Surgery located in Nur-Sultan, Kazakhstan. This data set included data from 451 patients with acute COVID-19.

Overall, this meta-analysis comprises 17,601 patients from 40 research papers and the NRCSC. Of the total, 24 studies originated from China, while the rest included data from France, Germany, Italy, Mexico, Singapore, South Korea, Turkey, the UK, and the USA. Most of the articles reported the division of patients according to the WHO guidelines. Out of the rest, 12 articles made groups based on the ICU admission, three focused on disease aggravation, and three on the need for oxygen therapy. This information is depicted in Figure 2 in more detail.

We collected data for five coagulation parameters: PLT, DD, FIB, APTT, and PT. In all cases, the heterogeneity was extremely high (I 2 > 80%), so the Knapp-Hartung adjusted random-effects model was applied in all parts of the meta-analysis. DD, FIB, The sizes of red squares are proportional to weights of respective studies, which are adjusted random-effects weights: reciprocal of the sum of squares of study's standard error and the variance of the distribution of true effect sizes. Pooled effect size and its confidence intervals are shown by blue diamond at the bottom. Detailed output of the meta-analysis is given in Table 1 .

Neither the analysis by the leave-one-out method nor the influence diagnostic has identified studies that solely cause a significant change in the effect size estimates. Removal of several outliers based on the pooled and individual confidence intervals allowed significant decrease in heterogeneity, yet it did not affect the interpretation of the results. Adjusted effect sizes and heterogeneity measurements are summarized in Table 2 . All influence diagnostic plots, such as Baujat and leave-one-out plots, Cook's distance, Covariance Ratio, etc., are provided in the Supplementary Figures 1-16 along with resulting Forest plots.

GOSH plots were built for all five models. K-means, DBSCAN, and Gaussian Mixture Model were implemented as clustering methods. This diagnostic was not applied on PLT and DD since they only have one cluster each ( Figures 6C-E) . GOSH plots of the prothrombin time effect size identified four studies, three of which (36, 39, 40) pull effect size to the left (Figures 6F-H) , and one (41) pulls effect size to the right ( Figure 6I ). We have excluded these studies and recalculated the pooled effect size for each parameter. Results are summarized in Table 2 .

To explore the possible sources of heterogeneity and test the robustness of our data, we have conducted subgroup analyses based on articles' country of origin and criteria for severity stratification. Results are summarized in Table 3 . Subgroup analysis showed no significant difference between articles originated from different countries but revealed that severity assessment criteria might have an influence on overestimated effect sizes for PLT and DD. High heterogeneity levels seem to be persistent oven different subgroups. Being the only exception, values of APTT obtained from non-Chinese articles were less heterogenous-47.1% compared to 84.7% in studies originated in China.

Contour-enhanced funnel plots for five coagulation parameters are depicted in Figure 7 , and the results of Egger's test are summarized in Table 4 . Based on the linear regression results, significant asymmetry is observed in published studies that included results for PLT, DD [Intercept (p-value): −1.297 (0.0336), 3.182 (0.0013)], and somewhat significant . While it is impossible to predict whether asymmetry is coming solely from publication bias, contour-enhanced funnel plots for DD and FIB (Figures 6B,C) show that the number of papers with insignificant findings is very small (p > 0.1). This may advocate for the presence of some publication bias. In contrast, Figure 6A illustrates that even though the funnel plot for PLT has some asymmetry, there are plenty of publications with insignificant SMDs, thus decreasing the possibility of publication bias.

To further explore the robustness of our findings, we have conducted several additional tests. As reported by Zwetsloot et al. (42) , plotting SMD against its standard error to assess the publication bias can result in distorted funnel plots, leading to false-positive conclusions. Since SMD is necessary for our analysis and cannot be substituted, we used a sample sizebased precision estimate instead of the standard error-1/ √ n. The resulting funnel plots (Supplementary Figure 16) showed no significant distortion. Similar corrections were made to Egger's test: since standardized mean difference and standard error are not independent, we can expect the test to yield some false-positive results. In our calculations, we used the corrected SE suggested by Pustejovsky and Rodgers (43) . Results are given in the Supplementary Table 3 . The only model that was affected by this correction was FIB (intercept 2.142 with a p-value of 0.0526 compared to the initial 1.872 and 0.0493), indicating that previously concluded asymmetry might have been a falsepositive. Interpretation of results for other parameters remained the same. 

There is a need to identify early indicators of the risk of critical patients progressing with COVID-19 to ICU. Abnormal coagulation parameters, including DD and FIB levels, and changes in PLT and PT, are observed in many patients with COVID-19 infection at admission [rev. (8, 13) ]. However, as discussed by different authors, these indicators provide contradictory information regarding risk stratification and prognosis (44) (45) (46) . Several previous meta-analysis reviews are researching abnormal coagulation and the severity of COVID-19. The majority of analyzed studies were conducted at the beginning of the pandemic and examined the Chinese population of patients [before April 2020- (14, (47) (48) (49) ; June 2020- (50) ].

Our meta-analysis study found that DD and FIB levels, APTT, and prothrombin time were significantly higher in the severe group of COVID-19 patients. In contrast, PLT was significantly lower in these patients. There was no evidence of effects across different geographic subgroups, except for APTT values.

In this meta-analysis, we systematically analyze evidence for the utility of lowered PLT levels on admission in relation to the severity of infection. Platelets play a crucial role in the maintenance of hemostasis, and contribute to thromboinflammatory processes (51) . At different stages of Frontiers in Cardiovascular Medicine | www.frontiersin.org viral infection, changes in platelet production or destruction may result in coagulation imbalances leading to pro-thrombotic events or in platelet dysfunction and bleeding risks (52) and increased activation of platelets (19) . Thrombocytopenia (platelet count <150 × 10 9 /L) seems to represent an important marker of COVID-19 severity (8, 14, 18, 53) and mortality outcome (14, 18) . The reasons for thrombocytopenia are multi-factorial may include early suppression of platelets progenitors' production and damage in bone marrow and lungs, hemophagocytosis and other reasons. Significant thrombocytopenia in COVID-19 is less frequent than it was reported for the largest SARS-1 cohort (45%) (54) . SARS-1 induced platelet depletion was related to direct viral infection of megakaryocytes and hematopoietic progenitors [rev. by (4) ]. However, almost half of patients developed thrombocytosis following initial thrombocytopenia during the SARS epidemics in 2003 (55) . It is possible that opposite PLT changes may reflect different phases of COVID-19 infection (56)-acute and convalescent phases, though the explanation has not been established yet. The reactive thrombocytosis that may occur during the convalescent phase requires time observation, including a period preceding patient discharge. The increase of immature reticulated platelets reported recently in COVID-19 patients by different groups (57, 58) may play a role in the limiting effects of anti-platelet therapies (59) (60) (61) . Change in platelets parameters (increased mean volume) (62) and reactivity (63, 64) are suggested to be associated with a severe COVID-19 infection (65) . In addition to platelet changes, there are an increasing number of studies related to the dysfunction of megakaryocytes (platelets progenitors) in COVID-19 patients. Thus, elevated and abnormal megakaryocytes (MKs) were found in post-mortem and autopsy studies in the lungs (66) (67) (68) (69) , heart (68), brain (70, 71) , bone marrow (72, 73) . Recently, the concept that megakaryopoiesis occurs in the lung parenchyma with significant contribution of the lungs in platelet biogenesis has emerged (74) (75) (76) . There is a close association between platelet biogenesis and alveolarization of the lung (77, 78) , and hypoxia-induced thrombocytopenia associated with reduction of lung megakaryocytes and impaired generation of platelets in the lungs (79) . The SARS-CoV-2 virus elicited up-regulation of megakaryocyte progenitors and elevated circulating megakaryocytes in blood in severe COVID-19 (80, 81) , hypothetically through infection of early megakaryocyte progenitors. Recently, Fortmann et al. (80) detected infected calprotectin+ spike+ megakaryocytes in the blood of patients with severe COVID-19. We hypothesize that viral RNA next may be transferred from megakaryocytes to platelets and circulating plasma EVs. The strong supporting argument is a recent finding of Brunet-Ratnasingam et al. (82) that plasma SARS-CoV-2 RNA is strongly associated with increased mortality. There is a current gap of knowledge regarding platelet non-canonical function and regeneration, and further research is crucial for our understanding of COVID-19 pathogenesis and stratification of increased thromboembolic risk.

Concordant with other studies, we found DD and FIB levels significantly elevated in severe COVID-19 infection. However, DD levels results may suffer from some publication bias as revealed by meta-regression analysis. It may be related to the variability of cut-offs for varying DD measurement assays used in the different studies [(83); rev. (84)]. The usefulness of hemostasis tests, particularly, D-dimers, is limited by interlaboratory variability reflecting differences in methodologies, antibody sources and specificity, different detection methods, calibrators, and diagnostic thresholds (85, 86) . As confirmed by autopsy results, the elevated DD levels can be associated with fibrin deposits within pulmonary extravascular space and alveoli (67), but they may be non-specific to intravascular fibrin formation (87) . High DD levels, together with elevated neutrophils, were reported to be predictive factors of pulmonary embolism in COVID-19 patients (88) . DD and FIB levels are elevated in both COVID-19 and thromboembolism, and therefore as single tests are unspecific and unhelpful in the differentiation of these conditions. In part, it may be explained by the involvement in abnormal coagulation in COVID-19 of different systems, including endothelial cells, complement activation, and hypofibrinolysis resulting in changes undetected by routine tests (39, 88) . A significantly higher amount of neutrophil extracellular traps [and related markers such as cell-free DNA and myeloperoxidase-DNA (MPO-DNA)] was found when compared with non-COVID-19 thrombi (89) . Neutrophil extracellular traps (NETs) markers, and cell-free DNA was most closely aligned with D-dimers, with inflammatory markers (C-reactive protein, lactate dehydrogenase) and more specific markers of NETs, MPO-DNA (90) (91) (92) . The presence and high concentrations of NETs at COVID-19 patients' tracheal aspirates, blood samples (93, 94) , and different organs (73) are associated with COVID-19 severity. The formation of NETs in the lungs and bloodstream can be critically associated with thrombosis (95) . Platelet activation and endothelial cell damage (96, 97) in COVID-19 patients are also resulting in elevated levels of extracellular vesicles (EVs) (98-100) expressing tissue factor (101, 102) and associated with severity of disease (99, 101, 103) .

PT has been included in the International Society on Thrombosis and Hemostasis (ISTH) criteria for diagnosing disseminated intravascular coagulation (DIC) (104) , prolonged at COVID-19 with acute distress respiratory syndrome (ARDS) (10) and is significantly elevated in severe COVID-19 patients (105) . Progressive prolongation of PT is considered a predictor of fatal outcome.

The first reports from China described a shortened APTT which is considered to be a predictive marker of hypercoagulation. Later studies reported prolonged APTT (45, 106) . A prolongation of APTT may indicate a deficiency of clotting factors (II, V, VIII, IX, X, XI, or XII) or the presence of inhibitor (heparin therapy). Another likely explanation is the presence of antiphospholipid antibodies (aPL) or lupus anticoagulant (LA) observed in 45% of COVID-19 patients in the Mulhouse series (107) . In Bowles's study (106) , most of the patients with prolonged APTT were positive for LA (91%). However, increasing of heparin therapy usage, especially in severe COVID-19 cases, may also lead to prolonged APTT.

Discordant reports exist concerning the fibrinogen levels and severity of COVID-19 disease. There are reports that COVID-19 is associated with high fibrinogen levels (16, 108) and alternative findings (109) . Some authors reported that severe COVID-19 infection induces fibrinolysis shutdown (26, 45) , which authors consider partly responsible for high DD and FIB levels. However, there are also reports of an increase in plasminantiplasmin complexes in severe COVID-19 and findings of mild consumption coagulopathy seen in a proportion of COVID-19 patients (110) . Elevated tissue-type plasminogen activator was also found to be increased in the patients with severe COVID-19 (101) .

Fibrinogen is a critical molecule that links coagulation, complement system, and inflammation, and at COVID-19, a significant association between FIB levels and interleukin-6 was described (16) . Recent post-mortem studies of COVID-19 patients identified microvascular thrombi in multiple organs, including kidney (67) , which may lead to increased DD even in the situation of reduced fibrinolysis. Depression of fibrinolysis was reported not only in ICU patients with severe infection and clinical signs of thromboembolism (26) but also in the general unselected ICU cohort (108, 109) . Viscoelastic testing ([ROTEM, tissue-type plasminogen activator (tPA) ROTEM, TEG] and evaluation of clot formation by Clot Waveworm analysis have demonstrated a hypercoagulable state, characterized by increased clot stiffness and severely impaired fibrinolysis (17, 108, 111, 112) . However, no consistent association between abnormal VET (viscoelastometric testing) pattern at COVID-19 and clinical outcome have been demonstrated [rev. (113) ].

Elevated routine coagulation parameters taken at admission may help with the initial differentiation of severe and nonsevere COVID-19 patient groups. The hypothetical mechanisms of fibrinolysis shutdown implicate lowering of fibrinolysis factors (plasminogen) and elevated amounts of inhibitors of fibrinolysis (such as α2-antiplasmin and plasminogen activator inhibitor PAI-1).

Longitudinal observational studies of COVID-19 infection are particularly important for understanding infection dynamics and disease outcomes (114) (115) (116) (117) (118) (119) . Hardy et al. (117) reported daily changes of relevant parameters of hemostasis of severe COVID-19 patients, including standard hemostatic tests (D-dimers, PT) as well as functional integrative tests for thrombin generation and fibrinolysis. Importantly, the day of hospital and/or ICU admission does not represent the same timepoint in the COVID-19 disease course that increases variability and complicates interpretation of studies results.

Although our findings were consistent across different geographic groups, they were tempered by significant heterogeneity. The DD, FIB, and PLT elevation trends speak about the necessity of investigating the potential of additional specific biomarkers for the stratification of COVID-19 patients. We suggest that thorough characterization of platelets and their progenitors (megakaryocytes), as well as clot formation testing could aid in the assessment of risk forthromboembolic events in COVID-19 patient groups.

Our findings should be interpreted with some limitations. Firstly, most of the included studies were either retrospective and observational by design, thus prone to recall or misclassification bias. Second, such studies often imply recruitment by convenience sampling, in which case the representativeness of a study is questioned. In the context of this meta-analysis, several different endpoints were allowed: severity guidelines, ICU admissions, oxygen therapy requirement, and disease aggravation. In an attempt to overcome this issue, we have conducted a subgroup analysis. Third, uncontrolled variables might also pose a limitation-both the disease severity and coagulation parameters can be affected by several potential confounding variables, such as comorbidities, age, etc. Fourth, data transformation was applied to those studies that reported their results in quartiles or maximum and minimum values. This could potentially introduce some degree of error in the pooled estimate. Fifth, the random-effects model was used due to the high heterogeneity of research. Although sensitivity and subgroup analyses were conducted, some residual heterogeneity may affect the interpretation of results. Finally, the underlying mechanisms of coagulation disbalances at COVID-19 still need to be investigated.

Our findings support using a combination of coagulation parameters for risk stratification of patients with COVID-19 infection at the time of admission, in particular DD and FB levels. However, additional evaluation using clotting formation methods or combination with specific biomarkers of production and activation of platelet and platelet progenitors is required. This information may help physicians triage patients with thromboembolism. An important consideration is an intermediate and long follow-up of COVID-19 patients.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics Committee of the National Research Center for Cardiac Surgery (NCJSC), Nur-Sultan, Kazakhstan. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

PL: conceptualization, methodology, formal analysis, investigation, data curation, and writing original draft. GI, ZS, and AK: methodology, data curation, formal analysis, and investigation. ATT, MS, AD, MA, and MB: methodology, formal analysis, investigation, data curation of NRCSC cohort, and writing a draft. EP and AT: methodology, analysis, and funding. NB: conceptualization, methodology, formal analysis, investigation, data curation, finding, and writing a draft. MB and NB: supervision. All authors read and edit a final draft. All authors contributed to the article and approved the submitted version.

NB and AT were supported by FDCRGP SSH2020028 and NB was funded by OPCRP2020018 grants from Nazarbayev University.

We are very grateful to Sadyk Khamitov and Aliya Sailybayeva from the Research Department of NCJSC National Research Center for Cardiac Surgery for their excellent organizational help and to Alexander Kovaldgi from 2nd Mathematical Liceum for helpful discussion. 

Pulmonary post-mortem findings in a series of COVID-19 cases from northern Italy: a two-centre descriptive study

Pulmonary pathology of severe acute respiratory syndrome in Toronto

Thrombocytopenia in patients with severe acute respiratory syndrome

Review: viral infections and mechanisms of thrombosis and bleeding

Guidelines for Investigation of Cases of Human Infection With Middle East Respiratory Syndrome Coronavirus (MERS-CoV)

Coagulation disorders in coronavirus infected patients: COVID-19, SARS-CoV-1, MERS-CoV and lessons from the past

Coagulopathy of coronavirus disease-2019

Analysis of dynamic disturbance in blood coagulation function of patients with coronavirus disease 2019: a retrospective observational study

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

Predictive factors of severe coronavirus disease-2019 in previously healthy young adults: a single-center, retrospective study

Clinical characteristics and laboratory indicator analysis of 67 COVID-19 pneumonia patients in Suzhou, China

Coagulation abnormalities and thrombosis in patients with COVID-19

Thrombocytopenia is associated with severe coronavirus disease 2019 (COVID-19) infections: a meta-analysis

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

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

Hypercoagulability of COVID-19 patients in intensive care unit: a report of thromboelastograpgy and other parameters of hemostasis

Thrombocytopenia and its association with mortality in patients with COVID-19

Antibodyinduced procoagulant platelets in severe COVID-19 infection

Platelet count rose while D-dimer levels dropped as deaths and thrombosis declined -an observational study on anticoagulation shift in COVID-19

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

Von Willebrand factor multimer formation contributes to immunothrombosis in Coronavirus disease 2019

COVID-19 and biomarkers of thrombosis: focus on von Willebrand factor and extracellular vesicles

Impaired fibrinolysis in critically ill COVID-19 patients

Fibrinolysis shutdown correlation with thromboembolic events in severe COVID-19 infection

Incidence of thrombotic complications in critically ill ICU patients with COVID-19

Thrombotic complications of patients admitted to intensive care with COVID-19 at a teaching hospital in the United Kingdom

Venous and arterial thromboembolic complications in COVID-19 patients admitted to an academic hospital in

Estimating the sample mean and standard deviation from commonly reported quantiles in meta-analysis

Optimally estimating the sample mean from the sample size, median, mid-range, and/or mid-quartile range

Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range

Methods to estimate the between-study variance and its uncertainty in meta-analysis

The clinical implication of dynamic neutrophil to lymphocyte ratio and D-dimer in COVID-19: a retrospective study in Suzhou China

Clinical features and risk factors for severe-critically ill COVID-19 adult patients in Jiangsu

Potential factors for prediction of disease severity of COVID-19 patients. medRxiv

Epidemiological and clinical characteristics of 1663 hospitalized patients infected with COVID-19 in Wuhan, China: a single-center experience

Complement activation and endothelial perturbation parallel COVID-19 severity and activity

Peripheral blood inflammatory markers in predicting prognosis in patients with COVID-19. Some differences with influenza A

Clinical and laboratory features of COVID-19: predictors of severe prognosis

Standardized means differences cause funnel plots distortion in publication bias assessments. eLife

Testing for funnel plot asymmetry of standardized mean differences

Early coagulation tests predict risk stratification and prognosis of COVID-19

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

Identification of parameters in routine blood and coagulation tests related to the severity of COVID-19

Changes in blood coagulation in patients with severe coronavirus disease 2019 (COVID-2019): a meta-analysis

Metaanalysis of coagulation parameters associated with disease severity and poor prognosis of COVID-19

Coagulation dysfunction is associated with a severity of COVID-19

Hemostatic changes in patients with COVID-19: a meta-analysis with meta-regressions

Platelet biology and functions: new concepts and clinical perspectives

Platelets and infection-an emerging role of platelets in viral infection

Platelet-to lymphocyte ratio is associated with prognosis in patients with coronavirus disease-19

A major outbreak of severe acute respiratory syndrome in Hong Kong

Haematological manifestations in patients with severe acute respiratory syndrome: retrospective analysis

ISTH interim guidance on recognition and management of coagulopathy in COVID-19

Immature platelets in patients hospitalized with COVID-19

Immature platelets as a biomarker for disease severity and mortality in COVID-19 patients

Reticulated platelets and uninhibited COX-1 and COX-2 decrease the antiplatelet effects of aspirin

Immature platelet fraction (IPF) determined with an automated method predicts clopidogrel hyporesponsiveness

Newly formed reticulated platelets undermine pharmacokinetically short-lived antiplatelet therapies

Mean platelet volume predicts severe COVID-19 illness

Platelet expression and function in patients with COVID-19

Platelets amplify endotheliopathy in COVID-19

Platelet activation in critically ill COVID-19 patients

Pulmonary and systemic involvement in COVID-19 patients assessed with ultrasound-guided minimally invasive autopsy

Vander Heide RS. Pulmonary and cardiac pathology in African American patients with COVID-19: an autopsy series from New Orleans

Megakaryocytes and platelet-fibrin thrombi characterize multi-organ thrombosis at autopsy in COVID-19: a case series

Increased number of pulmonary megakaryocytes in COVID-19 patients with diffuse alveolar damage: an autopsy study with clinical correlation and review of the literature

Assessing brain capillaries in coronavirus disease

Entrapped megakaryocytes in the microvasculature of brain tissues are not specific to COVID-19 but can be seen across a spectrum of acute lung injuries

A proof of evidence supporting abnormal immunothrombosis in severe COVID-19: naked megakaryocyte nuclei in the bone marrow and lungs of critically ill patients

Viral presence and immunopathology in patients with lethal COVID-19: a prospective autopsy cohort study

Circulating megakaryocytes: delivery of large numbers of intact, mature megakaryocytes to the lungs

The lung is a site of platelet biogenesis and a reservoir for hematopoietic progenitors

Lung megakaryocytes are immune modulatory cells

Close association between platelet biogenesis and alveolarization of the developing lung

Laboratory markers associated with COVID-19 progression in patients with or without comorbidity: A retrospective study

Low arterial oxygen partial pressure induces pulmonary thrombocytopenia in patients and a mouse model

Megakaryocytes are a novel SARS-CoV2 infection target and risk factor for mortality and multi-organ failure. medRxiv

Longitudinal multi-omics analyses identify responses of megakaryocytes, erythroid cells, and plasmablasts as hallmarks of severe COVID-19

Integrated immunovirological profiling validates plasma SARS-CoV2-RNA as an early predictor of COVID-19 mortality. medRxiv

D-dimer: simple test, tough problems

Issues surrounding age-adjusted D-dimer cutoffs that practicing physicians need to know when evaluating patients with suspected pulmonary embolism

Reporting of D-dimer data in COVID-19: some confusion and potential for misinformation

Management of the thrombotic risk associated with COVID-19: guidance for the hemostasis laboratory

The source of elevated plasma D-dimer levels in COVID-19 infection

D-Dimer level and neutrophils count as predictive and prognostic factors of pulmonary embolism in severe non-ICU COVID-19 patients

Vascular endothelial damage in the pathogenesis of organ injury in severe COVID-19

Neutrophil extracellular traps in COVID-19

Neutrophil extracellular traps contribute to immunothrombosis in COVID-19 acute respiratory distress syndrome

Complement and tissue factor-enriched neutrophil extracellular traps are key drivers in COVID-19 immunothrombosis

SARS-CoV2-triggered neutrophil extracellular traps mediate COVID-19 pathology

Circulating markers of neutrophil extracellular traps are of prognostic value in patients with COVID-19

Thrombosis: tangled up in NETs

Endothelial injury and glycocalyx degradation in critically ill coronavirus disease 2019 patients: implications for microvascular platelet aggregation

Endotheliopathy is induced by plasma from critically ill patients and associated with organ failure in severe COVID-19

Circulating platelet-derived extracellular vesicles are a hallmark of SARS-Cov-2 infection

Dissemination of extreme levels of extracellular vesicles: tissue factor activity in patients with severe COVID-19

Extracellular vesicles and citrinullated histone H3 in Coronavirus disease 2019 patients

Upregulation of pulmonary tissue factor, loss of thrombomodulin and immunothrombosis in SARS-CoV-2 infection

Circulating extracellular vesicles are endowed with enhanced procoagulant activity in SARS-CoV-2 infection

Extracellular vesicle-mediated endothelial apoptosis and EV-associated proteins correlate with COVID-19 disease severity

DIC score: statistical relationship with PT, aPTT, and simplified scoring systems with combinations pf PT and aPTT

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Lupus anticoagulant and abnormal coagulation tests in patients with COVID-19

Lupus anticoagulant is frequent in patients with Covid-19

Hypofibrinolytic state and high thrombin generation may play a major role in SARS-Cov2 associated thrombosis

Coagulation profile of COVID-19 patients admitted to the ICU: exploratory study

Prothrombotic changes in patients with COVID-19 are associated with disease severity and mortality

Evaluation of coagulation function by rotation thromboelastometry in critically ill patients with severe COVID-19 pneumonia

COVID-19 associated coagulopathy in critically ill patients: a hypercoagulable state demonstrated by parameters of haemostasis and clot waveform analysis

Viscoelastometric testing to assess hemostasis of COVID-19: a systematic review

Inference from longitudinal laboratory tests characterizes temporal evolution of COVID-19-associated coagulopathy

Hematological findings and complications of COVID-19

Biomarkers of coagulation, endothelial function, and fibrinolysis in critically ill patients with COVID-19: a single-center prospective longitudinal study

Prothrombotic disturbances of hemostasis of patients with severe COVID-19: a prospective longitudinal observational study

International comparisons of harmonized laboratory value trajectories to predict severe COVID-19: leveraging the 4CE collaborative across 342 hospitals and 6 countries: a retrospective cohort study

Association of coagulation disturbances with severity of COVID-19: a longitudinal study

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.Publisher's Note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.Copyright © 2022 Len, Iskakova, Sautbayeva, Kussanova, Tauekelova, Sugralimova, Dautbaeva, Abdieva, Ponomarev, Tikhonov, Bekbossynova and Barteneva. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.