Coagulopathy Underlying Rotational Thromboelastometry Derangements in Trauma Patients: A Prospective Observational Multicenter Study

This study is conducted as part of a prospective multicenter trial termed Activation of Coagulation and Inflammation in Trauma (United Kingdom Clinical Research Network Study Portfolio, ID: 5637). The Activation of Coagulation and Inflammation in Trauma study is a long-term observational cohort study that investigates early changes in coagulation and inflammation after trauma, performed in six European Level I trauma centers in London, United Kingdom; Oslo, Norway; Copenhagen, Denmark; Oxford, United Kingdom; Cologne, Germany; and Amsterdam, The Netherlands. For the analysis of this subset of patients, samples collected between January 2008 and November 2016 were used. Patients 18 yr or older admitted with a full trauma team activation and that arrived within 2 h after injury were screened for inclusion. Patients were not eligible if they received greater than 2 l of clear intravenous fluids before hospital admission (in order to exclude patients with a dilutional coagulopathy and capture patients with trauma-induced coagulopathy), suffered from burns covering more than 5% of the body or if they had known liver failure, were taking anticoagulant medication other than aspirin, or had pre-existing bleeding disorders. Furthermore, patients were excluded for data analysis if there were no EXTEM plus FIBTEM ROTEM variables available. This study was performed in accordance with the Statement of the Declaration of Helsinki and conducted after approval by the East London and The City Research Ethics Committee (07/Q0603/29) and the national ethics committees of participating centers. Written informed consent was obtained from each patient or from a legal representative in the case of incapacitated patients.

Data were collected to a centralized database and included patient demographics, mode of injury, injury severity score, vital signs, and laboratory test results at admission, as well as 12-h and 28-day mortality. In addition, the amount of fluids as well as amount of transfusions of red blood cells, fresh frozen plasma, and platelet units received between time of injury and baseline sampling at the emergency department were collected.

Blood samples were drawn in citrated tubes upon emergency department arrival. After blood samples were drawn, ROTEM was performed within 1 h at 37°C by trained study personnel on the ROTEM Delta (TEM International GmbH, Germany). Intrarater and interrater variability of this test have previously been published.¹⁵ 

Samples were analyzed with an automated analyzer (Sysmex CA-CS2100i System, Siemens, AG, Germany) to measure coagulation factor activity of factor [normal range]: II [78 to 117%], V [66 to 114%], VII [50 to 150%], VIII [50 to 150%], IX [58 to 138%], X [50 to 150%] and XIII [70 to 140%], with intra- and interassay coefficients of variation less than 5%.¹⁶  Clauss fibrinogen was processed by the central hospital laboratory. Enzyme-linked immunosorbent essays were used to quantify levels of D-dimer, α2-antiplasmin, plasmin antiplasmin complexes, and tissue plasminogen activator. Coagulation assays have been performed batchwise and without previous freeze-thaw cycles, thereby decreasing intertest variability.

In this study, both the EXTEM and FIBTEM tracings of ROTEM are used. In the EXTEM tracing, coagulation is activated by tissue factor derived from rabbit brain. In the FIBTEM tracing, coagulation is also activated by tissue factor with the addition of cytochalasin D to inhibit platelet function in order to assess the effect of fibrinogen on clot formation. These tracings were chosen since EXTEM and FIBTEM derangements are relevant and frequent issues in trauma induced coagulopathy. Furthermore, FIBTEM is more sensitive to fibrinolysis compared to EXTEM.^17–19  The following ROTEM cutoff values were used to define hypocoagulability: EXTEM clotting time greater than 80 s, EXTEM clot amplitude at 5 min less than 40 mm, EXTEM lysis index at 30 min less than 85%, FIBTEM clot amplitude at 5 min less than 10 mm, and FIBTEM lysis index at 30 min less than 85%. The cutoff value for clotting time was chosen based on the recommendation by a consensus group on viscoelastic test–based transfusion guidelines for early trauma resuscitation as well as the ROTEM reference value, since current evidence is insufficient to recommend an exact clotting time threshold for transfusion.²⁰  The choice for the clot amplitude at 5 min and lysis index at 30 min thresholds was based on previous work in this same patient cohort demonstrating that these thresholds have a good performance to predict transfusion requirement or mortality.²¹ 

For the analyses, patients were categorized into eight groups based on admission ROTEM profiles. Group 1 contained patients with a normal ROTEM profile. Group 2 contained patients with an isolated EXTEM clotting time prolongation (EXTEM clotting time greater than 80 s). Group 3 contained patients with a decrease in EXTEM clot amplitude at 5 min (EXTEM clot amplitude at 5 min less than 40 mm). Group 4 contained patients with a decrease in FIBTEM clot amplitude at 5 min (FIBTEM clot amplitude at 5 min less than 10 mm). Group 5 contained patients with a normal EXTEM clotting time (80 s or less) but with a decreased EXTEM and FIBTEM clot amplitude at 5 min (EXTEM clot amplitude at 5 min less than 40 and FIBTEM clot amplitude at 5 min less than 10). Group 6 contained patients with a prolonged EXTEM clotting time, decreased EXTEM clot amplitude at 5 min, and decreased FIBTEM clot amplitude at 5 min (EXTEM clotting time 80 s or less, EXTEM clot amplitude at 5 min less than 40 mm, and EXTEM clot amplitude at 5 min less than 10 mm). Group 7 contained patients with a normal EXTEM clotting time, but with decreased EXTEM and FIBTEM clot amplitude at 5 min in combination with lysis index at 30 min less than 85 in either the EXTEM or FIBTEM tracing (EXTEM clotting time 80 s or less, EXTEM clot amplitude at 5 min less than 40 mm, FIBTEM clot amplitude at 5 min less than 10 mm, EXTEM lysis index at 30 min or FIBTEM lysis index at 30 min less than 85%). Group 8 contained patients with a prolonged EXTEM clotting time, decreased EXTEM and FIBTEM clot amplitude at 5 min, and lysis index at 30 min less than 85% in either the EXTEM or FIBTEM tracing (EXTEM clotting time greater than 80 s, EXTEM clot amplitude at 5 min less than 40 mm, FIBTEM clot amplitude at 5 min less than 10 mm, EXTEM or FIBTEM lysis index at 30 min less than 85%).

A data analysis and statistical plan were performed as part of a secondary analysis of the Activation of Coagulation and Inflammation in Trauma study; therefore, the sample size of this descriptive observational cohort study was based on the available data, and no a priori statistical power calculation was conducted. A statistical analysis plan was made before accessing the data. No minimum clinically meaningful effect size was defined before data access. Previous analyses of this cohort have been used to construct a ROTEM-based treatment algorithm for trauma-induced coagulopathy, related to clinical outcome.²¹  Data are presented as median with interquartile range for numerical variables and as percentages for categorical variables. ROTEM profile group is the independent variable with levels of coagulation factors and fibrinolysis markers as dependent variables. Levels of coagulation factors and fibrinolysis markers were the primary outcome, and mortality was the secondary outcome. Statistical significance between the previously defined groups was computed using a Kruskal-Wallis test with Bonferroni correction for multiple testing with pairwise comparison with a Mann-Whitney U test. Categorical variables were compared with a chi-square test. A two-tailed P value < 0.05 was considered statistically significant. Missing data were likely at random as they appeared in all participating centers and across all injury severity scores and not imputed, although this study in severely injured trauma patients carries a risk for selection bias. Furthermore, in response to peer review, an adjusted association between mortality and ROTEM derangement was added as a post hoc sensitivity model, which can be found in the Supplemental Digital Content (https://links.lww.com/ALN/C863, https://links.lww.com/ALN/C864, https://links.lww.com/ALN/C865). To do this, a binary logistic regression was performed to correct for age, sex, lactate as a marker for shock, and injury severity as possible confounders for mortality. This resulted in a model with mortality as outcome variable and age, sex, lactate, injury severity score, and ROTEM derangement group as predictor variables. These predictor variables were selected based on their clinical association with mortality after trauma. Statistical analysis was performed with SPSS version 26.0 (IBM, USA), and graphs were made with GraphPad version 9.1.0 (USA).

Of 1,855 eligible patients, 27 patients could not be stratified into one of the eight predefined subgroups, leaving 1,828 patients for analysis (baseline demographics in table 1). Tranexamic acid was administered in hospital as part of the major hemorrhage protocol, and 105 patients (5.7%) received tranexamic acid before baseline sampling. A deranged ROTEM profile was present in 732 (40%) patients and most often consisted of a combined decrease in EXTEM and FIBTEM clot amplitude at 5 min, present in 217 patients (11.9%). On the other hand, hyperfibrinolysis (lysis index at 30 min less than 85% in either EXTEM or FIBTEM tracing) was observed in only 35 patients (1.9%). Injury severity score was higher in patients with a combination of ROTEM derangements and was the highest when all parameters were deranged. Moreover, patients with a derangement in all ROTEM parameters were most in shock. When compared to the other groups, platelet count was most markedly decreased in the group where all ROTEM parameters were deranged, although the median level remained above 150 × 10⁹/l. Furthermore, this group was the only one with an increase in prothrombin time and activated partial thromboplastin time above the reference range.

In patients with an isolated clotting time greater than 80 s, none of the factor levels were lower when compared to patients with a normal ROTEM profile (fig. 1). In patients with an EXTEM clot amplitude at 5 min less than 40 mm, fibrinogen was the only factor level that was decreased compared to patients with a normal ROTEM profile, whereas in patients with a FIBTEM clot amplitude at 5 min less than 10 mm, fibrinogen as well as coagulation factors II, V, and XIII were decreased compared to patients with a normal ROTEM profile. In all other ROTEM coagulation groups, coagulation factors I (fibrinogen), II, V, VII, IX, X, and XIII displayed a similar pattern of gradual decrease as more ROTEM parameters became deranged. Of these, fibrinogen and factor V but not the other factors decreased less than 50%. Patients with a combined clotting time greater than 80 s, EXTEM clot amplitude at 5 min less than 40 mm, and FIBTEM clot amplitude at 5 min less than 10 mm derangement had a median value of factor V of 38.3% (25.0 to 71.6%), and patients with a combination of all ROTEM derangements (clotting time, EXTEM and FIBTEM clot amplitude at 5 min, and lysis index at 30 min) had the most severe factor V depletion with a median of 22.8% (8.9 to 33.5%).

Compared to patients with a normal ROTEM profile, a finding of EXTEM or FIBTEM lysis index at 30 min less than 85% was associated with increased levels of D-dimer (78,681 ng/ml vs. 6,658 ng/ml), plasmin antiplasmin complexes (21,587 µg/l vs. 2,753 µg/l), and a decreased level of α2-antiplasmin (32.1 pg/ml vs. 98.9 pg/ml), thus indicating fibrinolysis (fig. 2). Increased fibrinolysis as reflected by a lysis index at 30 min less than 85% was often accompanied by other ROTEM derangements. Of importance, fibrinolysis was also present in the groups with deranged clotting time or clot amplitude at 5 min but with a normal lysis index at 30 min. The group with a lysis index at 30 min derangement without EXTEM clotting time prolongation was small with a high variability in levels of fibrinolysis markers, resulting in an absence of statistical significance compared to the group with a normal ROTEM profile.

Patients with a deranged ROTEM profile received more erythrocyte transfusion after hospital admission and had a higher mortality compared to patients with a ROTEM profile within the reference range. Mortality was highest when fibrinolysis or multiple ROTEM derangements were present. Furthermore, ROTEM profiles remained significantly associated with mortality after correction for confounders (P = 0.009; Supplemental Digital Content table 3, https://links.lww.com/ALN/C866). An exception to this finding were the patients with an isolated clotting time prolongation, in whom outcome was comparable to patients who present with a normal ROTEM profile.

There is a surge of attention to the use of ROTEM in trauma management. This study aims to improve understanding of the coagulopathies that are reflected by ROTEM derangements. We report the following clinically relevant findings. First, in the group with an isolated clotting time prolongation, levels of coagulation factors as well as outcome were similar to patients with a normal ROTEM profile. Second, in patients with a deranged ROTEM profile, levels of fibrinogen and factor V show the most pronounced decrease. Third, FIBTEM clot amplitude at 5 min derangements reflects more severe coagulation factor deficiencies than EXTEM clot amplitude at 5 min derangements. Last, fibrinolysis also occurs when EXTEM and FIBTEM clot amplitude at 5 min are both deranged in the presence of a normal lysis index at 30 min.

An isolated EXTEM clotting time prolongation is thought to reflect abnormalities in clot initiation in the tissue factor–dependent coagulation pathway, in particular deficiencies of coagulation factors I (fibrinogen), II, V, and VII.²²  Therefore, it has been suggested that clotting time derangements may reflect the need for plasma (or coagulation factor concentrate). In line with this, current trauma guidelines advise that a clotting time prolongation may be considered an indication for plasma transfusion but that current evidence thus far is insufficient to recommend specific thresholds.^20,23  Our data show that the hemostasis profile of patients with an isolated clotting time prolongation is similar to patients who present with a normal ROTEM profile. Of importance, an isolated EXTEM clotting time greater than 80 s was not related to adverse outcome. Therefore, we suggest that an isolated EXTEM clotting time greater than 80 s is not the appropriate cutoff level that should trigger plasma or coagulation factor concentrates, because many patients will have adequate coagulation factor levels. This gives rise to the question whether an alternative clotting time cutoff value would provide a better positive predictive value for coagulation disturbance. However, the clinical relevance of this question is debatable considering that in our cohort, none of the patients with an isolated clotting time prolongation died. This suggests that clinically relevant clotting time prolongations are accompanied by other ROTEM derangements, and a clotting time prolongation in combination with other ROTEM abnormalities should trigger plasma transfusion. Of note, a clotting time prolongation can also reflect anticoagulant treatment.^24–26  Since these patients were excluded in this study, our results are not generalizable to patients on oral anticoagulants other than aspirin.

Despite increasing evidence on the association between coagulation factor depletion, trauma-induced coagulopathy, and adverse outcome, it is not known which activity of factor levels is optimal after trauma.^27,28  Although it is assumed that normal hemostasis requires a coagulation factor activity between 20 and 50%, these levels are based on patients with a single factor deficiency.²⁹  The clinical importance of a more modest but simultaneous decrease in activity of several coagulation factors should probably not be underestimated and may also contribute to trauma-induced coagulopathy.³⁰  However, our results suggest differential decreases in factor levels as demonstrated already for cardiac surgery.³¹  In spite of a gradual decrease as more ROTEM parameters become deranged, levels of coagulation factors II, VII, VIII, IX, X, and XIII remain largely within normal range, even in groups with increased mortality. In contrast, levels of fibrinogen and factor V decrease more frequently and more severely after injury. Given the association between fibrinogen levels and mortality, it is recommended to maintain levels of fibrinogen above 1.5 g/l.^23,32  In our cohort, a combination of EXTEM, and FIBTEM clot amplitude at 5 min derangements represents a fibrinogen depletion less than 1.5 g/l and should trigger fibrinogen suppletion. Of interest, factor V depletion in our cohort equaled that of fibrinogen. This finding is consistent with previous studies showing a marked decrease in factor V, while other factor levels remained relatively normal.^30,33  In particular, factor VII does not decrease. We offer the following explanation for these findings: in trauma induced coagulopathy, activated protein C is increased, which inhibits factor V and VIII with ensuing factor X, and fibrinogen, whereas factor VII remains out of this activation loop. Of note, a decrease in factor V activity in our study is associated with increased mortality.³⁴  This begs the question whether substitution of factor V in severely injured patients is rational. It is important to consider that factor V is not present in the currently trialed four factor concentrates (NCT04534751).

A decreased FIBTEM clot amplitude at 5 min reflects more severe depletion of fibrinogen and factor levels than EXTEM clot amplitude at 5 min, as found before.^35,36  This finding may beg the question whether the dosage of substitution of factor levels should be higher in case of low FIBTEM clot amplitude at 5 min compared to low EXTEM clot amplitude at 5 min.

Hyperfibrinolysis as measured with viscoelastic testing is a relevant finding, associated with increased mortality and massive transfusion, with incidences reported as high as 32.5% based on FIBTEM lysis index at 60 minutes less than 85%.^7,37,38  In our cohort, hyperfibrinolysis measured by lysis index at 30 min is rare. This is not unexpected given the low sensitivity of lysis index at 30 min. Of note, EXTEM lysis index at 60 minutes and, even better, FIBTEM lysis index at 60 minutes are more sensitive to detect moderate to mild hyperfibrinolysis.^7,8,39  However, the detection time for lysis index at 60 minutes is long, and untreated fibrinolysis can aggravate bleeding. Therefore, early diagnosis of fibrinolysis with lysis index at 30 min is used in clinical ROTEM algorithms.¹⁴ 

An important finding is that a combination of clotting time and clot amplitude at 5 min derangements also reflects increased fibrinolysis, even when the lysis index at 30 min was normal. A retrospective study in patients undergoing noncardiac surgery was able to demonstrate that low early values of clot firmness in the EXTEM pathway are associated with fibrinolysis.⁴⁰  Accordingly, our results demonstrate that in trauma patients, a combination of EXTEM and FIBTEM clot amplitude at 5 min derangements represent increased D-dimer with a concomitant decrease in α2 antiplasmin, indicating fibrinolysis. Furthermore, when EXTEM and FIBTEM clot amplitude at 5 min derangements are accompanied by clotting time prolongation, fibrinolysis becomes even more outspoken. This indicates that elevated fibrinolysis should not be ruled out in these patients and that combined ROTEM derangements should trigger (repeated) antifibrinolytic therapy.

This study has limitations. Patients with a dilutional coagulopathy and patients on anticoagulant treatment were excluded, we therefore cannot make recommendations about these patient categories. In addition, the study was performed some time ago. Since then, prehospital interventions such as administration of tranexamic acid have become more common. However, we do not think that this influences the implications of our findings, as a deranged ROTEM profile should trigger an intervention regardless of previously administered therapy. Moreover, we cannot rule out selection bias as families of the most severely injured patients are less likely to give consent, possibly leading to smaller samples in the more deranged ROTEM groups. Furthermore, EXTEM and FIBTEM lysis index at 60 minutes results are not available in this study, which resulted in the low incidence of reported hyperfibrinolysis based on ROTEM results. Last, general limitations of ROTEM should be acknowledged. ROTEM measurements are not sensitive to the effect of platelet inhibitors nor the effect of von Willebrand Factor on clot formation, which is particularly relevant in patients with severe injury since von Willebrand Factor can be released by the endothelium as a response to severe trauma.

Our findings are relevant, as the ROTEM derangements appear to be independently associated with outcome. Results enable potential improvement of trauma-induced coagulopathy treatment when bleeding is deemed clinically relevant. We propose to adjust ROTEM-based algorithms by omitting an isolated EXTEM clotting time prolongation as a trigger for intervention in patients not on anticoagulant treatment. Also, the presence of both EXTEM and FIBTEM clot amplitude at 5 min derangements with a normal lysis index at 30 min may justify administration of antifibrinolytic therapy, although this hypothesis needs to be validated prospectively. Furthermore, whether suppletion of factor V during trauma is beneficial is an important subject for future research.

In conclusion, ROTEM derangements are a frequent observation at trauma presentation. An isolated EXTEM clotting time prolongation should not reflexively trigger therapy in patients not on anticoagulant therapy or not bleeding. In contrast, other ROTEM derangements reflect a decrease in fibrinogen and factor V levels, and a combination of EXTEM and FIBTEM clot amplitude at 5 min derangements represents fibrinolysis, even when the lysis index at 30 min is normal. These results should be considered in current ROTEM-based trauma algorithms.

This study is part funded by the European Commission (Brussels, Belgium) under the FP-7 HEALTH-contract No. F3-2013-602771, entitled “Targeted Action for curing Trauma Induced Coagulopathy.” Clinical trial number: ID: 5637, sponsor Queen Mary University of London, REC reference 07/Q0603/29, IRAS reference 71328, Chief investigator: Dr. Davenport.

Dr. Davenport has a financial relationship with Werfen (Barcelona, Spain). Dr. Curry has a financial relationship with CSL Behring (King of Prussia, Pennsylvania) for investigator-led research; Bayer (Leverkusen, Germany) for consultancy; and Sob for an unrestricted research grant. Dr. Maegele received honoraria for consultancy and lectures and advisory boards from Abbott (Abbot Park, Illinois), Astra Zeneca (Cambridge, United Kingdom), Biotest (Dreieich, Germany), Bayer, CSL Behring, IL-Werfen, LFB Biomedicaments (Les Ulis, France), and Portola Inc. (San Francisco, California). Dr. Juffermans has a financial relationship with Octapharma (Lachen, Switzerland). The other authors declare no competing interests.

Supplemental Figure A: Coagulation Factor Levels with Li30 Cutoff < 95%, https://links.lww.com/ALN/C863

Supplemental Figure B: Fibrinolysis Marker Levels with Li30 Cutoff < 95%, https://links.lww.com/ALN/C864

Legends to Supplemental Figures, https://links.lww.com/ALN/C865

Supplemental Tables: Supplementary Patient Characteristics and Statistics, https://links.lww.com/ALN/C866