Secondary Logo

Journal Logo


Modified ROTEM for the detection of rivaroxaban and apixaban anticoagulant activity in whole blood

A diagnostic test study

Pailleret, Claire; Jourdi, Georges; Siguret, Virginie; Gouin-Thibault, Isabelle; Gandrille, Sophie; Stepanian, Alain; Curis, Emmanuel; Golmard, Jean-Louis; Gaussem, Pascale; Le Bonniec, Bernard; Samama, Charles M.

Author Information
European Journal of Anaesthesiology: June 2019 - Volume 36 - Issue 6 - p 449-456
doi: 10.1097/EJA.0000000000000903



Rivaroxaban and apixaban are oral anticoagulants with direct factor Xa (FXa) inhibition and are given at a fixed dose which results in mean peak therapeutic plasma concentrations of 100 to 200 ng ml−1 without the need for routine coagulation monitoring.1,2 Like any anticoagulant, FXa inhibitors are associated with an increased risk of spontaneous and peri-operative bleeding complications.3 Thus, managing clinical situations such as urgent surgery, major bleeding or performing thrombolysis in patients suffering from stroke, is a challenge with these agents. The problem is twofold: no antidote has been approved yet,4 and no test allows ultrarapid evaluation of the anticoagulant status of a patient.5

Currently, measurement of FXa inhibitor concentration is obtained through target-specific assay (anti-Xa activity).5–10 These are precise titrations but are not readily available. In addition, these assays require a relatively long turnaround time rendering them unsuitable in the context of an emergency. Other tests, such as thrombin generation assay (TGA) and clot waveform analysis triggered in platelet-poor plasma (PPP), enable detection of as little as 5 to 10 ng ml−1 FXa inhibitor.11–13 These are triggered with much lower amounts of tissue factor (TF) than prothrombin time, possibly explaining their higher sensitivity. Nevertheless, these methods are only performed in specialised laboratories, and the blood samples require centrifugation to create a PPP sample; hence, the delay in obtaining results exceeds the time scale required for emergencies.

Rotational thromboelastometry (ROTEM) and thromboelastography (TEG) are viscoelastic point-of-care devices which assess haemostasis in whole blood, therefore not requiring preparation of PPP plasma.14–16 Performed under low shear conditions, these machines reveal changes of the viscoelastic properties of blood in a forming clot, at all stages of its maturation, including the subsequent retraction/fibrinolysis. ROTEM or TEG is readily available in a number of operating rooms and intensive care or emergency units. On the contrary, with the commercially available test reagents [EXTEM (a test of external haemostatic function, similar to prothrombin time) and INTEM (a measure of the internal haemostatic system, similar to the activated partial thromboplastin time)] reagents, they are ineffective for FXa inhibitor detection.12,17–24 Studies show that although clotting time (CT) triggered with EXTEM is significantly affected by FXa inhibitors,25 values observed below 200 ng ml−1 of inhibitor overlap with those observed in the absence of any drug, precluding the use of this assay for FXa inhibitor detection.18,19,22,24 Part of the limitation seems to originate from the 15% interassay variability.26 Lack of sensitivity undoubtedly originates too from the TF content of the EXTEM reagent. Adelmann et al.27 have shown with ROTEM that decreasing the TF amount up to a final concentration of 0.35 pmol l−1 (without phospholipid complementation) dramatically improves the sensitivity to FXa inhibitors with however a concurrent increase of the interassay variability, which compromised reliable detection of FXa inhibitors.

First, this study aimed to assess the effect of whole blood components on FXa inhibitor detection with viscoelasticity measurements, and second to evaluate whether a modified ROTEM assay, triggered with a low amount of TF as in TGA or clot waveform analysis, can reliably detect FXa inhibitors in whole blood.

Materials and methods

Study population

For the in vitro experiments, blood samples from 13 healthy volunteers, who gave their written informed consent, were collected at the local blood bank (convention C CPSL UNT no. 12/EFS/038 Établissement Français du Sang, Paris, France). The in vitro study was started in November 2014 and ended in November 2015. For the ex vivo experiments, citrated blood samples from patients on rivaroxaban (N=30) or apixaban (N=17) were screened between 2 March and 29 April 2016. Blood samples from 19 patients who were not taking anticoagulant treatment formed the control group. Exclusion criteria included constitutional and acquired abnormalities of haemostasis and the current use of any other antithrombotic drug. Blood samples were collected by venepuncture in tubes containing (1 to 10) 0.105 mol l−1 buffered trisodium citrate.

Ethics committee approval

Ethical approval for the clinical part of the study (Ethical Committee BB-0033-00064) was provided by the local Ethical Committee of Lariboisière University Hospital (Centre de Ressources Biologiques; Chairperson Prof JM Launay), Paris, France. Lariboisière Haematology Laboratory has a general consent to use patient blood samples for various haemostatic assays. All these patients gave their written informed consent. Our study was conducted on 1-ml blood samples from FXa-inhibitor-treated patients: the blood was obtained to perform other assays totally independent of our study.

Drugs and reagents

Apixaban and rivaroxaban were kindly provided by Bristol-Myers Squibb/Pfizer (Princeton, New Jersey, USA) and Bayer HealthCare AG (Leverkusen, Germany), respectively. Apixaban and rivaroxaban were dissolved in 100% dimethyl sulfoxide (DMSO) and stored at −80 °C. Just before use, aliquots were rapidly diluted 1 to 100 in H2O: the solution was then stabilised for at least 24 h. Further dilutions were performed in 1% DMSO-containing 0.15 mol l−1 NaCl. Effective concentrations were determined by anti-Xa activity measurement in PPP on a STA-R (Stago, Asnières, France), using the STA-Liquid anti-Xa assay and the specific set of calibrators for rivaroxaban and apixaban (Stago). An amount of 20 ng ml−1 was the lower limit of quantification. Phospholipid vesicles were prepared by sonication (2 min in pulse mode 0.15 s−1, 80 W, 4 °C) of a 1-mg ml−1 mixture of L-α-phosphatidylcholine (66%, w/w) with L-α-phosphatidylserine (33%, w/w), both from Avanti Polar Lipids (Alabaster, Alabama, USA) as previously described.28 Recombinant human TF (Innovin) was purchased from Dade Behring (Marburg, Germany). Its concentration was estimated by comparing its activity to the 5 pmol l−1 TF with 4 μmol l−1 phospholipid vesicles PPP-Reagent (Stago) as reference. Fibrinogen (Clottafact) was obtained from LFB Biotechnologies (Courtaboeuf, France) and was concentrated by ultrafiltration to 34 mg ml−1 according to its absorbance at 280 nm (A280, ε% = 15.1).

Preparation of platelet-poor plasma, platelet-rich plasma and red blood cells

Pooled normal PPP (Cryocheck) was purchased from Cryopep (Montpellier, France), and a single lot was used in this study. PPP was supplemented with purified fibrinogen (1/10; v/v) when required. Whole blood from healthy volunteers was used to prepare platelet-rich plasma (PRP) and red blood cells (RBC) by centrifugation at 216 g for 11 min and 2400 g for 10 min, respectively. Artificial 60% haematocrit was prepared by adding PPP to packed RBC (4/6; v/v); other artificial haematocrit levels were obtained by further PPP addition.

Rotational thromboelastometry

ROTEM was performed on a ROTEM delta in 340-μl cups (Tem International GmbH, Munich, Germany). Final DMSO content was 0.05% in all assays which was shown not to interfere with coagulation.13,29 Typically, 50 μl of a 1% DMSO solution containing or not containing FXa inhibitor were added to 950 μl of PPP, or PRP or whole blood from healthy volunteers. Assays performed on whole blood from patients on anticoagulant treatment were done without added DMSO. Samples were processed at least 1 h after and within 6 h of collection.30 Modified ROTEM tests were triggered with the addition of a saturating amount of phospholipid vesicles (10 μmol l−1) and a low amount of TF (5 pmol l−1 in PPP or PRP and adjusted to 2.5 pmol l−1 in whole blood, considering an average haematocrit about 45%) such as in TGA or in clot waveform analysis. We saturated the assay with phospholipid vesicles to minimise the platelet contribution. To each ROTEM cup, 20 μl of a mixture of TF and phospholipid vesicles were added together with 20 μl of 0.2 mol l−1 of calcium chloride. Assays were triggered by adding 300-μl prewarmed PPP, PRP or whole blood and rapid mixing by pipetting up and down three times. Viscoelasticity data were recorded at 37 °C for 45 min. CT, clot formation time (CFT), alpha angle and maximum clot firmness (MCF) were provided by the ROTEM software analysis. Raw data were extracted using the DyCoDerivAU software (Tem International GmbH).

Study protocol

The sensitivity of commercialised EXTEM reagent to FXa inhibitor was first assessed in spiked whole blood of six healthy volunteers. Samples were spiked with DMSO only (0.05%) as a control, and various amounts of rivaroxaban (25, 100 or 200 ng ml−1). Each sample was run in triplicate.

To evaluate the specific effect of whole blood (WB) components on FXa inhibitor detection by viscoelasticity measurement, ROTEM assays were performed in various samples including pooled normal PPP (Supplemented or not with purified fibrinogen,, PRP, PPP with packed RBC and healthy volunteer whole blood. Six experiments for each condition and each FXa inhibitor amount were performed. In the ex vivo study, each patient sample was run in duplicate.

Data collection

Data collection was planned before ROTEM analysis and anti-Xa assay were performed. Both ROTEM analysis and anti-Xa assay results were available to the performers or readers of each test.

Statistical analysis

Data were expressed as mean ± SD or median with interquartile range [Q1 to Q3]. CT, CFT, alpha angle or MCF ratios (with/without inhibitor) in the rivaroxaban or apixaban (25, 100 and 200 ng ml−1) or control groups were compared all together by an analysis of variance on repeated data (repeated analysis of variance). Change induced by xaban 25 ng ml−1 was considered as the primary criterion and comparisons between results from 25 ng ml−1 and no drug were performed without adjustment. Comparisons between 100 or 200 ng ml−1 and no drug were assessed by Dunnett tests. In the artificially adjusted haematocrit experiments, values of CT between the rivaroxaban and the control groups were compared by the Kruskal–Wallis test followed by Mann–Whitney tests. Receiver operator characteristic (ROC) curve analysis was used to determine the optimal CT cut-off value (best compromise between sensitivity and specificity), allowing distinction between xaban treated and untreated patients. Sensitivity and specificity are given with their two-sided 95% confidence intervals (95% CIs). Computation was performed with the SAS V9.3 (Statistical Analysis Software version 9.3, Cary, North Carolina, USA) statistical software (SAS, Cary, North Carolina), except ROC curves that were built using R (http:/ and package Proc (from Comprehensive R Archive Network: https:/ Statistical significance was accepted for P values below 0.05.


Limitation of commercialised rotational thromboelastometry tests in factor Xa inhibitor detection in whole blood (WB)

Clot formation was triggered with EXTEM reagent in whole blood (from six healthy volunteers) spiked with various amounts of rivaroxaban. On the basis of the upper limit of the reference ranges of CT values for the EXTEM test defined by Lang et al.,26 we confirmed that conventional ROTEM analysis in whole blood detects only amounts higher than 200 ng ml−1 rivaroxaban reliably (Supplementary Figure,

Detection of rivaroxaban and apixaban in platelet-poor plasma

Various ROTEM triggering conditions were examined in PPP. To minimise intra-assay variability, we choose to saturate assays with 10 μmol l−1 of phospholipid vesicles. Triggering ROTEM with 5 pmol l−1 TF and 10 μmol l−1 phospholipid vesicles in PPP resulted in a reproducible CT with a median value of 92 [89 to 95] s in the absence of FXa inhibitors.

Then we investigated whether modified ROTEM could reliably detect low amounts of FXa inhibitors in PPP. CT value increased 1.4-fold (P = 0.02) with 25 ng ml−1 rivaroxaban and 1.6-fold (P < 0.0001) with 25 ng ml−1 apixaban (Table 1). CT was the most relevant parameter to detect the FXa inhibitors. Although alpha angle decreased significantly, its published interassay variability precluded its use.26 MCF was little affected, if at all, and at least 100 ng ml−1 of the drug were required to increase CFT significantly.

Table 1
Table 1:
Detection of rivaroxaban and apixaban in platelet-poor plasma

Transposing detection of factor Xa inhibitor in platelet-poor plasma to reconstituted whole blood

Neither changes of fibrinogen concentration nor variations in platelet or RBC content preclude rivaroxaban detection. Neither fibrinogen nor platelets modified CT significantly in the presence of FXa inhibitor (Table 2).

Table 2
Table 2:
Influence of fibrinogen and platelets on rivaroxaban detection

CT lengthened only 1.14-fold when the haematocrit increased from 30 to 60% in the absence of FXa inhibitors: median CT value was 154 [146 to 166] versus 176 [172 to 179] s (P = 0.03), respectively. Adding 25 ng ml−1 rivaroxaban significantly increased CT (P < 0.001, Fig. 1) irrespective of the haematocrit: median CT value was 230 [207 to 252] versus 165 [152 to 170] s. The overall ratio (with/without rivaroxaban) was 1.47 (P < 0.001) for haematocrits between 30 and 60%.

Fig. 1
Fig. 1:
Detection of rivaroxaban by modified rotational thromboelastometry in reconstituted blood, box-and-whisker plot of the clotting time in seconds, measured in reconstituted blood having haematocrit artificially adjusted to 30, 40, 50 or 60% (six estimations each) in relation to the rivaroxaban content in platelet-poor plasma. Clot formation was triggered by adding 2.5 pmol l−1 tissue factor and 10 μmol l−1 phospholipid vesicles (with respect to whole blood volume). Each box represents 24 measurements of clotting time (six estimations at each haematocrit level). The bottom and top of the box indicate first and third quartiles, line within each box the median value, and top and bottom of the whiskers the lowest and highest values observed. Statistical significance: *** P less than 0.001 (versus control without factor Xa inhibitor). Unambiguous detection of 25 ng ml−1 rivaroxaban or more was achieved irrespective of the haematocrit.

Detection of factor Xa inhibitors in patients treated with these drugs

A total of 66 patients [24 men, 42 women, mean age 60 (range 18 to 89) years] were included in the study of whom 30 were treated with rivaroxaban, 17 with apixaban and 19 without treatment.

The measured anti-Xa activity ranged from 20 (lower limit of quantification) to 483 ng ml−1 for rivaroxaban or 434 ng ml−1 for apixaban. The optimal cut-off of the modified ROTEM CT value for discriminating untreated patients from those receiving either rivaroxaban or apixaban was assessed using ROC curve analysis (Fig. 2). A CT value more than 197 s provided the highest sum of sensitivity and specificity, 85% (95% CI, 94 to 72%) and 100% (95% CI, 100 to 82%), respectively, with an area under the ROC curve of 0.96 (95% CI, 0.99 to 0.92), allowing distinction between samples containing FXa inhibitors or not. Sensitivity and specificity for this CT cut-off value allowing detection of FXa inhibitor concentrations more than 30 ng ml−1 and more than 100 ng ml−1 were also calculated. A CT value more than 197 s was able to predict rivaroxaban and apixaban concentrations more than 30 ng ml−1, with a sensitivity of 90% (95% CI, 97 to 76%) and specificity of 85% (95% CI, 96 to 65%). A 96% sensitivity (95% CI, 100 to 81%) and 64% specificity (95% CI, 79 to 47%) were obtained for rivaroxaban and apixaban concentration exceeding 100 ng ml−1 with CT value more than 197 s determined by modified ROTEM assay.

Fig. 2
Fig. 2:
Receiver operating characteristic curve of modified rotational thromboelastometry clotting time for factor Xa inhibitor detection in patients’ whole blood. A total of 66 patients were included, 30 treated with rivaroxaban, 17 with apixaban and 19 with neither. Actual factor Xa inhibitor concentration was estimated by anti-Xa assay (in platelet-poor plasma). Clot formation was triggered by adding 2.5 pmol l−1 tissue factor and 10 μmol l−1 phospholipid vesicles. Receiver operating characteristic curve of modified rotational thromboelastometry clotting time for the discrimination between untreated and rivaroxaban or apixaban-treated patients. The area under curve was 0.96. A cut-off of clotting time value more than 197 s provided the highest sum of sensitivity (85%) and specificity (100%) allowing identification of patients treated with rivaroxaban or apixaban.


Increasingly, FXa inhibitors such as apixaban and rivaroxaban are prescribed for oral anticoagulant therapy. A challenge persists in detecting the activity of these drugs in an emergency. Given that TGA or clot waveform analysis is adequately sensitive to FXa inhibitors in PPP, we hypothesised that the effects of these drugs should also be detectable through viscoelasticity measurements triggered in whole blood in the same conditions – a low amount of TF and a saturating amount of phospholipid vesicles. The data we present here suggest that a modified ROTEM assay has a considerably improved sensitivity to FXa inhibitors compared with conventional ROTEM analysis.

The lack of a readily available test to detect the anticoagulant status of a patient treated with direct FXa inhibitors represents a major concern to clinicians in the context of an emergency. ROTEM and TEG are point of care devices assessing haemostasis in WB.14–16 On the contrary, previous studies have shown that they lacked sensitivity in monitoring the effects of FXa inhibitors: patients treated with FXa inhibitors frequently show CT values within the normal range.12,17–24 Our findings show that a lengthening of CT values above 197 s, as measured with a modified ROTEM, could distinguish between untreated patients and those receiving FXa inhibitors, with a good sensitivity and high specificity. However, we must acknowledge that with a limited number of samples tested, erroneous diagnoses cannot be ruled out for 15% of patients. Although the safe haemostatic level of FXa inhibitors is not yet clearly defined, current literature considers concentrations of rivaroxaban and apixaban below 30 ng ml−1 as well tolerated for surgery, and below 100 ng ml−1 as well tolerated for thrombolysis in acute ischaemic stroke.31,32 Based on these guidelines, we analysed the ability of a modified ROTEM to predict values above or below these thresholds. We observed that a modified ROTEM CT more than 197 s could detect samples containing more than 30 ng ml−1 of rivaroxaban or apixaban with a good sensitivity and specificity. However, modified ROTEM analysis could not exclude a concentration above 30 ng ml−1 in 10% of cases, suggesting that results should be interpreted with caution. Considering a sensitivity more than 95% as safe in clinical practice, a modified ROTEM CT more than 197 s enables reliable detection of FXa inhibitor amounts above 100 ng ml−1 with 96% sensitivity (95% CI, 100 to 81%) but with a moderate specificity of 64% (95% CI, 79 to 47%).

The potential of viscoelasticity measurement is that it is feasible in WB, possibly opening an avenue for point-of-care monitoring of FXa inhibitors. At least three components dramatically influence the rheology of clot formation in WB, namely platelets, fibrinogen and RBC.33–39 Our study showed that those components would probably not preclude detection of low concentrations of FXa inhibitor when coagulation was triggered by a fixed amount of TF. We observed that CT was not significantly modified when comparing PPP containing various amounts of fibrinogen. Consistent with our results, Fenger-Eriksen et al.40 and Tanaka et al.41 reported that hypofibrinogenaemia does not influence clot initiation. We chose to saturate WB with procoagulant phospholipid vesicles as an alternative to cytochalasin D inhibition,42 to minimise the platelet contribution to viscoelasticity. Adelmann et al.27 demonstrated that a low amount of TF allows detection of FXa inhibitors through ROTEM analysis, yet interassay variability precluded safe estimation as CT values in the absence of inhibitor overlapped with concentrations of drug below 200 ng ml−1. These authors used highly diluted TF (Innovin) without supplementing phospholipid content. Thus, clot formation relied upon the procoagulant surface formed through platelet activation. That clotting depends on platelet activation possibly explains the variability observed. In support of this hypothesis, Butenas et al.43 reported that high amounts of phospholipid vesicles minimise the platelet contribution in thrombin generation. In addition, the volume occupied by RBC introduces uncertainty in the effective concentration of the trigger. Normal haematocrit varies between 35 and 55%, with the result that reagent concentration according to WB volume could be between 0.65 and 0.45 of the amount measured in the corresponding PPP. Indeed, adding 5 pmol l−1 TF to PPP results in 10 pmol l−1 effective TF in WB when haematocrit is 50%, but only 6 pmol l−1 effective TF when haematocrit is 30%. Thus, effective concentration of the trigger is proportional to the haematocrit. As a point-of-care monitor is designed to simplify the test and should not require a correction according to haematocrit, the challenge became to evaluate how accurate FXa inhibitor detection could be in WB with an unvarying TF amount, taking into account the inherent variability of samples having haematocrits between 30 and 60%. Our results suggest that RBC count did not preclude reliable FXa inhibitor detection when clot formation was triggered with 2.5 pmol l−1 TF amount.

Several limitations of our study have to be considered including, first, the single-centre nature and the small size of our study. Further studies with more patients are needed to confirm our results. A reference range of modified ROTEM CT values should be established: this will require a large-scale evaluation beyond the scope of the current study. The effect of WB components on FXa inhibitor detection was evaluated only using rivaroxaban, whereas the evaluation of our modified ROTEM was performed in patients’ samples treated with either rivaroxaban or apixaban. Therefore, the lack of effect of changes of fibrinogen concentration, platelet or RBC content on apixaban detection was not verified but is unlikely to be different, given the reliable apixaban detection achieved in patient WB samples with the modified ROTEM in the current study. As multiple ROTEM assays could not be performed simultaneously immediately following WB collection, we used citrated blood samples with recalcification for ROTEM analysis. Those ROTEM analyses were delayed by 1 h after sampling, to achieve a stable coagulant status comparable with that obtained from noncitrated fresh blood samples.29 Meesters et al.44 have described a progressive acceleration of blood coagulation revealed by ROTEM analysis over the 30 to 60 min after recalcification of citrated blood samples. Another limitation is that the CT value in the ROTEM assay may be affected by other factors, including inflammatory syndromes, coagulopathies or other anticoagulant treatments: it should be kept in mind that an increased CT does not only imply that the sample contains FXa inhibitor.45–47 Ali et al.48 found that TEG has limited clinical utility to detect preinjury anticoagulation in acute trauma patients. Furthermore, viscoelasticity assays still await proceedings of standardisation and quality control to ensure reliable results. The challenge for the ROTEM to become a reliable point-of-care device implies overcoming these hurdles.

Our study aimed to use a modified ROTEM as a point-of-care test, allowing reliable detection of FXa inhibitors in an emergency context. We conclude that the modified ROTEM fulfils this specification with good sensitivity and specificity. Its ready availability can make this approach an added value in critical situations, in which a rapid decision is required. However, a large-scale evaluation of the modified ROTEM assay is still needed to confirm our results.

Acknowledgements relating to this article

Assistance with the study: none.

Financial support and sponsorship: the authors would like to thank the generous donors of the CONNY-MAEVA Charitable Foundation. The authors also thank Bristol-Myers Squibb/Pfizer for providing apixaban and Bayer HealthCare Pharmaceuticals France for rivaroxaban.

Conflicts of interest: CMS, VS and IG-T have received honoraria for participating in expert meetings on apixaban (Bristol-Myers Squibb/Pfizer) and together with PG on rivaroxaban (Bayer Healthcare AG).

Presentation: preliminary data for this study were presented as a poster at the French Society of Hematology (SFH), 23 to 25 March 2016, Paris and at the International Society of Thrombosis and Haemostasis (ISTH), 8 to 13 July 2017, Berlin.


1. Mueck W, Eriksson BI, Bauer KA, et al. Population pharmacokinetics and pharmacodynamics of rivaroxaban – an oral, direct factor Xa inhibitor in patients undergoing major orthopaedic surgery. Clin Pharmacokinet 2008; 47:203–216.
2. Frost C, Song Y, Barrett YC, et al. A randomized direct comparison of the pharmacokinetics and pharmacodynamics of apixaban and rivaroxaban. Clin Pharmacol 2014; 6:179–187.
3. Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011; 365:883–891.
4. Levy JH, Ageno W, Chan NC, et al. Subcommittee on Control of Anticoagulation. When and how to use antidotes for the reversal of direct oral anticoagulants: guidance from the SSC of the ISTH. J Thromb Haemost 2016; 14:623–627.
5. Gosselin RC, Adcock DM. The laboratory's 2015 perspective on direct oral anticoagulant testing. J Thromb Haemost 2016; 14:886–893.
6. Baglin T, Hillarp A, Tripodi A, et al. Measuring oral direct inhibitors (ODIs) of thrombin and factor Xa: a recommendation from the Subcommittee on Control of Anticoagulation of the Scientific and Standardisation Committee of the International Society on Thrombosis and Haemostasis. J Thromb Haemost 2013; 11:756–760.
7. Francart SJ, Hawes EM, Deal AM, et al. Performance of coagulation tests in patients on therapeutic doses of rivaroxaban. A cross-sectional pharmacodynamic study based on peak and trough plasma levels. Thromb Haemost 2014; 111:1133–1140.
8. Hillarp A, Gustafsson KM, Faxälv L, et al. Effects of the oral, direct factor Xa inhibitor apixaban on routine coagulation assays and anti-FXa assays. J Thromb Haemost 2014; 12:1545–1553.
9. Gouin-Thibault I, Flaujac C, Delavenne X, et al. Assessment of apixaban plasma levels by laboratory tests: suitability of three anti-Xa assays. A multicentre French GEHT study. Thromb Haemost 2014; 111:240–248.
10. Schellings MW, Boonen K, Schmitz EM, et al. Determination of dabigatran and rivaroxaban by ultra-performance liquid chromatography-tandem mass spectrometry and coagulation assays after major orthopaedic surgery. Thromb Res 2016; 139:128–134.
11. Dinkelaar J, Molenaar PJ, Ninivaggi M, et al. In vitro assessment, using thrombin generation, of the applicability of prothrombin complex concentrate as an antidote for Rivaroxaban. J Thromb Haemost 2013; 11:1111–1118.
12. Herrmann R, Thom J, Wood A, et al. Thrombin generation using the calibrated automated thrombinoscope to assess reversibility of dabigatran and rivaroxaban. Thromb Haemost 2014; 111:989–995.
13. Jourdi G, Siguret V, Martin AC, et al. Association rate constants rationalise the pharmacodynamics of apixaban and rivaroxaban. Thromb Haemost 2015; 114:78–86.
14. Zuckerman L, Cohen E, Vagher JP, et al. Comparison of thrombelastography with common coagulation tests. Thromb Haemost 1981; 46:752–756.
15. Luddington RJ. Thrombelastography/thromboelastometry. Clin Lab Haematol 2005; 27:81–90.
16. Levi M, Hunt BJ. A critical appraisal of point-of-care coagulation testing in critically ill patients. J Thromb Haemost 2015; 13:1960–1967.
17. Dias JD, Norem K, Doorneweerd DD, et al. Use of thromboelastography (TEG) for detection of new oral anticoagulants. Arch Pathol Lab Med 2015; 139:665–673.
18. Casutt M, Konrad C, Schuepfer G. Effect of rivaroxaban on blood coagulation using the viscoelastic coagulation test ROTEM. Anaesthesist 2012; 61:948–953.
19. Eller T, Busse J, Dittrich M, et al. Dabigatran, rivaroxaban, apixaban, argatroban and fondaparinux and their effects on coagulation POC and platelet function tests. Clin Chem Lab Med 2014; 52:835–844.
20. Körber MK, Langer E, Ziemer S, et al. Measurement and reversal of prophylactic and therapeutic peak levels of rivaroxaban: an in vitro study. Clin Appl Thromb Hemost 2014; 20:735–740.
21. Escolar G, Arellano-Rodrigo E, Lopez-Vilchez I, et al. Reversal of rivaroxaban-induced alterations on hemostasis by different coagulation factor concentrates – in vitro studies with steady and circulating human blood. Circ J 2015; 79:331–338.
22. Chojnowski K, Górski T, Robak M, et al. Effects of rivaroxaban therapy on ROTEM coagulation parameters in patients with venous thromboembolism. Adv Clin Exp Med 2015; 24:995–1000.
23. Fontana P, Alberio L, Angelillo-Scherrer A, et al. Impact of rivaroxaban on point-of-care assays. Thromb Res 2017; 153:65–70.
24. Seyve L, Richarme C, Polack B, et al. Impact of four direct oral anticoagulants on rotational thromboelastometry (ROTEM). Int J Lab Hematol 2018; 40:84–93.
25. Schenk B, Würtinger P, Streif W, et al. Ex vivo reversal of effects of rivaroxaban evaluated using thromboelastometry and thrombin generation assay. Br J Anaesth 2016; 117:583–591.
26. Lang T, Bauters A, Braun SL, et al. Multicentre investigation on reference ranges for ROTEM thromboelastometry. Blood Coagul Fibrinolysis 2005; 16:301–310.
27. Adelmann D, Wiegele M, Wohlgemuth RK, et al. Measuring the activity of apixaban and rivaroxaban with rotational thrombelastometry. Thromb Res 2014; 134:918–923.
28. Le Bonniec BF, Guinto ER, Esmon CT. The role of calcium ions in factor X activation by thrombin E192Q. J Biol Chem 1992; 267:6970–6976.
29. Camici GG, Steffel J, Akhmedov A, et al. Dimethyl sulfoxide inhibits tissue factor expression, thrombus formation and vascular smooth muscle cell activation: a potential treatment strategy for drug-eluting stents. Circulation 2006; 114:1512–1521.
30. Camenzind V, Bombeli T, Seifert B, et al. Citrate storage affects thrombelastograph analysis. Anesthesiology 2000; 92:1242–1249.
31. Pernod G, Albaladejo P, Godier A, et al. Management of major bleeding complications and emergency surgery in patients on long-term treatment with direct oral anticoagulants, thrombin or factor-Xa inhibitors: proposals of the working group on perioperative haemostasis (GIHP) – March 2013. Arch Cardiovasc Dis 2013; 106:382–393.
32. Steiner T, Bohm M, Dichgans M, et al. Recommendations for the emergency management of complications associated with the new direct oral anticoagulants (DOACs), apixaban, dabigatran and rivaroxaban. Clin Res Cardiol 2013; 102:399–412.
33. Bolliger D, Szlam F, Molinaro RJ, et al. Finding the optimal concentration range for fibrinogen replacement after severe haemodilution: an in vitro model. Br J Anaesth 2009; 102:793–799.
34. Shams Hakimi C, Fagerberg Blixter I, Hansson EC, et al. Effects of fibrinogen and platelet supplementation on clot formation and platelet aggregation in blood samples from cardiac surgery patients. Thromb Res 2014; 134:895–900.
35. Zentai C, Solomon C, Meijden PE, et al. Effects of fibrinogen concentrate on thrombin generation, thromboelastometry parameters, and laboratory coagulation testing in a 24-hour porcine trauma model. Clin Appl Thromb Hemost 2016; 22:749–759.
36. Winstedt D, Thomas OD, Nilsson F, et al. Correction of hypothermic and dilutional coagulopathy with concentrates of fibrinogen and factor XIII: an in vitro study with ROTEM. Scand J Trauma Resusc Emerg Med 2014; 22:73.
37. Shibata J, Hasegawa J, Siemens HJ, et al. Hemostasis and coagulation at a hematocrit level of 0.85: functional consequences of erythrocytosis. Blood 2003; 101:4416–4422.
38. Nagler M, Kathriner S, Bachmann LM, et al. Impact of changes in haematocrit level and platelet count on thromboelastometry parameters. Thromb Res 2013; 131:249–253.
39. Gersh KC, Nagaswami C, Weisel JW. Fibrin network structure and clot mechanical properties are altered by incorporation of erythrocytes. Thromb Haemost 2009; 102:1169–1175.
40. Fenger-Eriksen C, Jensen TM, Kristensen BS, et al. Fibrinogen substitution improves whole blood clot firmness after dilution with hydroxyethyl starch in bleeding patients undergoing radical cystectomy: a randomized, placebo-controlled clinical trial. J Thromb Haemost 2009; 7:795–802.
41. Tanaka KA, Taketomi T, Szlam F, et al. Improved clot formation by combined administration of activated factor VII (NovoSeven) and fibrinogen (Haemocomplettan P). Anesth Analg 2008; 106:732–738.
42. Schlimp CJ, Solomon C, Ranucci M, et al. The effectiveness of different functional fibrinogen polymerization assays in eliminating platelet contribution to clot strength in thromboelastometry. Anesth Analg 2014; 118:269–276.
43. Butenas S, Branda RF, van’t Veer C, et al. Platelets and phospholipids in tissue factor-initiated thrombin generation. Thromb Haemost 2001; 86:660–667.
44. Meesters MI, Koch A, Kuiper G, et al. Instability of the nonactivated rotational thromboelastometry assay (NATEM) in citrate stored blood. Thromb Res 2015; 136:481–483.
45. Ruttmann TG, Jamest MF, Lombard EH. Haemodilution-induced enhancement of coagulation is attenuated in vitro by restoring antithrombin III to predilution concentrations. Anaesth Intensive Care 2001; 29:489–493.
46. Spiel AO, Mayr FB, Firbas C, et al. Validation of rotation thromboelastography in a model of systemic activation of fibrinolysis and coagulation in humans. J Thromb Haemost 2006; 4:411–416.
47. Rugeri L, Levrat A, David JS, et al. Diagnosis of early coagulation abnormalities in trauma patients by rotation thromboelastography. J Thromb Haemost 2007; 5:289–295.
48. Ali JT, Daley MJ, Vadiei N, et al. Thromboelastogram does not detect preinjury anticoagulation in acute trauma patients. Am J Emerg Med 2017; 35:632–636.

Supplemental Digital Content

© 2019 European Society of Anaesthesiology