Association of D-dimer and fibrinogen magnitude with hypercoagulability by thromboelastography in severe COVID-19

Introduction: D-dimer concentration has been used to identify candidates for intensified anticoagulant treatment for both venous thromboembolism prevention and mitigation of the microthrombotic complications associated with COVID-19. Thromboelastography (TEG) maximum amplitude (MA) has been validated as an indicator of hypercoagulability and MA [≥] 68 mm has been utilized as a marker of hypercoagulability in other conditions. We evaluated the relationship between coagulation, inflammatory, and TEG parameters in patients with COVID-19 on extracorporeal membrane oxygenation (ECMO). Methods: We performed a single center retrospective analysis of consecutive patients that received ECMO for the treatment of COVID-19. TEG, inflammatory, and coagulation markers were compared in patients with and without thrombotic complications. Correlation tests were performed to identify the coagulation and inflammatory markers that best predict hypercoagulability as defined by an elevated TEG MA. Results: 168 TEGs were available in 24 patients. C-reactive protein and fibrinogen were significantly higher in patients that developed a thrombotic event versus those that did not (p=0.038 and p=0.043 respectively). D-dimer was negatively correlated with TEG MA (p<0.001) while fibrinogen was positively correlated (p<0.001). A fibrinogen > 441 mg/dL had a sensitivity of 91.2% and specificity of 85.7% for the detection of MA [≥] 68 mm. Conclusions: In critically ill patients with COVID-19, D-dimer concentration had an inverse relationship with hypercoagulability as measured by TEG MA. D-dimer elevation may reflect severity of COVID-19 related sepsis rather than designate patients likely to benefit from anticoagulation. Fibrinogen concentration may represent a more useful marker of hypercoagulability in this population.

CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020 markedly elevated in other conditions including sepsis in the absence of clinically evident 68 mm. [14] In this study we evaluated the relationship between commonly utilized laboratory 1 1 0 parameters and hypercoagulability as represented by TEG MA. We performed a retrospective medical record review of consecutive hospitalized patients   Only patients on ECMO were evaluated as it is standard practice at our institution to obtain daily 1 1 8 TEGs as part of the routine care of these patients. TEG was performed using the TEG 5000  . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020 Data was abstracted in a structured format by one of the authors (SP) and entered into a 1 2 3 customized data form. Data was collected when available for standard TEG parameters (R-time   1  2  4 and MA), D-dimer, fibrinogen, CRP, and ferritin. D-dimer values > 20 ug/mL FEU were 1 2 5 recorded as 20 ug/mL FEU and R-time reported as > 60 min was denoted as 60 min. TEG and 1 2 6 inflammatory markers or coagulation studies were compared if drawn within 24 hours. For  Distribution of all continuous data was examined for normality using visual inspection and the interquartile range and compared using the Wilcoxon rank sum test. Categorical data are 1 4 0 presented as counts with proportions and compared using Fisher's exact test (two-tailed). A p 1 4 1 value < 0.05 was considered statistically significant. parametric modeling, the natural log of these variables is used in graphical presentations and 1 4 4 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 29, 2020. presented using a receiver operator characteristic curve, summarized using the area under the  ECMO or after this therapy was discontinued (9 deep vein thromboses, 2 pulmonary emboli, and outcomes between groups are displayed in Table 1. Patients with imaging confirmed 1 5 8 macrothrombosis were significantly more likely to have higher median CRP and fibrinogen . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10.1101/2020.07.27.20162842 doi: medRxiv preprint 1 0 to be positively correlated with MA. Ln(fibrinogen) explained over 50% of the variance in 1 7 9 ln(MA) (adjusted R 2 = 0.567) and the association was significant for patients on heparin, 1 8 0 bivalirudin, and in aggregate (p < 0.001) (Fig 1B). With ln(fibrinogen) in the model, CRP and 1 8 1 ferritin were examined as additional predictors and were not significant.   Receiver operator curves (ROC) based on CRP (Fig 2) and fibrinogen levels (Fig 3) were  85.7%) satisfies the closest-to-(0, 1) criterion for threshold selection. The emergence of COVID-19 as a global pandemic has necessitated the implementation of 1 9 7 treatment protocols based on expert opinion given limited available clinical data. D-dimer has 1 9 8 been cited as correlating with worse clinical outcomes in COVID-19 and coagulopathy has been 1 9 9 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 29, 2020. In this present study we aimed to evaluate the correlation of commonly available measures of Our results demonstrate that in this population of critically ill COVID-19 patients receiving actually inversely related to MA in our study population. This finding was unexpected and the 2 1 0 physiologic basis is uncertain. of critically ill patients with 3,15] Despite these guidelines, local and regional 2 1 5 hospital practices vary widely, with some recommending D-dimer thresholds to aid the decision 2 1 6 regarding initiation and intensity of anticoagulation. Based on our data, we speculate that 2 1 7 commonly noted severe D-dimer aberrations are likely markers of COVID-19 related sepsis 2 1 8 rather than degree of hypercoagulability. In the absence of additional data, it seems prudent to  . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted July 29, 2020. ICU also demonstrated an association between MA and fibrinogen. [10,16] In this study, TEG is not routinely available at many centers and takes specialized training to perform to 2 3 0 achieve accurate results. Fibrinogen is a commonly available laboratory tests in most hospital 2 3 1 environments and results may be easier to obtain. Thus, it may represent a more practical noted in patients with COVID 19 is not well proven. We postulate that an imbalance between 2 4 2 procoagulant factors and anticoagulant factors is logical. How elements such as factor VIII, 2 4 3 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

4 5
Our findings should be interpreted in the context of the study limitations. First, our data is 2 4 6 retrospectively derived from a single tertiary care center. The study was limited to severe patients. Given these considerations, our selected population may limit generalizability and our boluses of heparin and one patient received tissue plasminogen activator in addition to 2 6 0 anticoagulation for repetitive pulmonary emboli. There is limited data regarding the use of should also be explored further in this population. As a result, the effects of anticoagulation and dosing adjustments of these medications on the study outcomes are unknown and may introduce 2 6 5 a source of bias. Of note, all findings were consistent when examined for patients on heparin 2 6 6 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 29, 2020. . https://doi.org/10. 1101/2020