In patients with cirrhosis associated with liver dysfunction, thromboelastographic (TEG) and thromboelastometric (ROTEM) indices of clot firmness have long been reported to decrease in proportion to the degree of liver dysfunction and to indicate hypocoagulability.1–3 In contrast, studies have shown that due to reduced levels of both procoagulant factors and inhibitors, the haemostatic balance is likely to be preserved in patients with cirrhosis, despite a prolonged prothrombin time (PT).4,5 Similarly, it has been suggested that variable thrombocytopenia and defects in platelet function are balanced by increased levels of factor VIII and von Willebrand factor (vWF), the main determinant of platelet adhesion.6,7 In addition, a limited number of studies based on the measurements of individual components of the fibrinolytic system have suggested enhanced fibrinolysis.8 In contrast, a ROTEM-based calculation of fibrinolysis velocity showed no significant differences between cirrhotic patients of all Child groups and healthy controls.9 More recently, thrombin generation has been shown to be indistinguishable in patients with stable cirrhosis from that seen in healthy volunteers when thrombomodulin, a physiological endothelial activator of protein C, was added to the test mixture, provided that the platelet count was higher than 60 x 109 l−1.9–12 Evidence is also available that plasma from patients with cirrhosis possesses a procoagulant imbalance detected in vitro by the ratio of thrombin generation measured in the presence vs. absence of thrombomodulin.9–12 The thrombin generation test (TGT) in the presence of thrombomodulin accounts best for the interplay between pro and anticoagulant factors in the haemostatic process and is strongly associated with the overall in-vivo potential of each individual to generate thrombin.13,14
Clinical data are in line with the concept of preserved or even increased coagulation in patients with cirrhosis.15–22 Various procedures, including liver biopsy, dental extraction and liver transplantation, have been conducted without bleeding or transfusion requirement in several series of cirrhotic patients with prolonged PT.15–18 Cirrhotic patients are not protected from peripheral vein thrombosis or pulmonary embolism even when the international normalised ratio (INR) is above 2.2.19 In addition, the prevalence of portal vein thrombosis is 8 to 25% in candidates for liver transplantation.20 Activation of coagulation and intrahepatic microthrombi influence the progression of liver fibrosis.21,22 As conflicting aspects of the complex haemostatic process have been investigated in separate studies in patients with cirrhosis,1–12 a comprehensive study of haemostasis was designed in patients with various degrees of liver dysfunction associated with cirrhosis, with a special focus on the relevance of the ROTEM tests. Three approaches of haemostasis, including measurement of coagulation factors and inhibitors, ROTEM analysis and TGT assay in the presence of thrombomodulin, were prospectively investigated.
Materials and methods
Ethical approval for this prospective, observational, single-centre study was provided by the Ethics Committee of Ile de France III, Tarnier-Cochin University Hospital, 89 rue d’Assas, 75006 Paris, France (Chairperson: David Simhon) on 8 February 2013 and registered as ‘Ethical Committee: S.C. 3024 – Trial registration: ID RCB: 2012-A01728-35’.
Forty patients admitted consecutively to the Liver Unit of Cochin University Hospital, Paris, France for clinical assessment, from 1 February 2013 to 25 June 2013 were included as part of routine clinical care. According to French regulations, the patients were informed of the study, but written consent was waived. Adult patients, aged 18 years or older, with a diagnosis of cirrhosis, were eligible for the study. Cirrhosis was diagnosed on the basis of clinical, laboratory and ultrasound evidence, or on liver biopsy. The criteria for exclusion were as follows: the use of any drugs known to interfere with blood coagulation; overt bleeding; packed red blood cell and/or fresh frozen plasma and/or platelet transfusion 3 weeks prior to inclusion; any ongoing documented bacterial infection; hepatocellular carcinoma; extrahepatic malignancy; peripheral, pulmonary or portal thrombosis; haemostatic disorders not related to cirrhosis; and cirrhosis resulting from an inflammatory or an immune disorder. The severity of the liver disease was estimated in accordance with the model for end-stage liver disease (MELD) score.23 To calculate the MELD score, we used the modified MELD score as employed by the United Network for Organ Sharing (UNOS) for organ allocation: MELD = [0.957 × log (creatinine) + 0.378 × log (bilirubin) + 1.2 × log (INR) + 0.64] × 10 (available at: http//http://www.unos.org/resources).
Blood was drawn using minimum venous stasis by clean venepuncture and collected in plastic tubes containing 1 : 10 volume of 3.2% citrate sodium solution (9 NC; Becton, Dickinson and Co., Meylan, France). Four millilitres of whole blood was used for ROTEM analysis. The remaining citrated blood was centrifuged at 2500g for 15 min at 20°C before analysis of the inhibitors and activators of coagulation. Platelet-poor plasma was aliquoted and stored at -80°C according to French Group on Haemostasis and Thrombosis (GEHT) recommendations after a second centrifugation until tested for the TGT assays.
Measurement of conventional tests
Plasma assessment of fibrinogen, factor V, factor VIII, vWF, protein C activity, free protein S and antithrombin (AT) activity was carried out on an automated coagulation analyser (STAR; Stago, Asnières, France). The coagulation factors were measured by a clotting assay using Siemens deficient reagent for factors V and VIII (Siemens Healthcare GmbH, Erlangen, Germany). vWF and free protein S were determined by turbidimetric assays using the STA Liatest vWF:AG (Stago) and STA Liatest Free Protein S (Stago). When the vWF values were above the upper limit of quantification, that is 420%, the plasmas were diluted in buffer for assessment. The AT and protein C levels were determined by chromogenic assays with STA Stachrom ATIII (Stago) and Biophen PC5 (Hyphen Biomed., Neuville sur Oise, France), respectively.
Thrombin generation tests
Thrombin generation was measured with a calibrated automated thrombography system in accordance with Hemker et al.24 as previously described. Briefly, the procedure was carried out by means of an automated fluorometer (Fluoroskan Ascent; ThermoLab System, Helsinki, Finland). Readings from the fluorometer were automatically recorded and calculated using dedicated software (Thrombinoscope; Thrombinoscope BV, Maastricht, The Netherlands), which displays thrombin generation curves (generated thrombin vs. time) and calculates the parameters described below. To assess the normal value ranges, the same batch of frozen pooled normal plasma (Cryochek pooled normal plasma, Cryopep, France) was systematically tested in each run of TGT assay. In addition, normal values without thrombomodulin addition were confirmed using samples from controls. The control population comprised 30 apparently healthy volunteers from our hospital laboratory, who were not using drugs known to affect the coagulation system and without history of venous thromboembolism or bleeding disorders. Measurements were conducted on plasma with final concentrations of 5 pmol tissue factor (TF) and 5 nmol thrombomodulin. This concentration of thrombomodulin inhibits thrombin generation in normal pooled plasma by 50%.
Rotational thromboelastometry assays
Blood samples were analysed at 37°C using the four-channel ROTEM delta analyser (Tem Innovations GmbH, Munich, Germany). The details of the ROTEM analysis are described elsewhere.25 Briefly, a runtime of 60 min was applied and regular quality control tests and ROTEM tests were run in accordance with the manufacturer's instructions. For the present study, three tests were carried out using reagents provided by the manufacturer: ellagic acid activated intrinsic pathway (INTEM), which evaluates the so-called intrinsic pathway; TF triggered extrinsic pathway (EXTEM), which evaluates the so-called extrinsic pathway; and cytochalasin D, which is a platelet inhibitor and evaluates the contribution of fibrinogen to clot formation (FIBTEM).
In a study based on TGT assessment, Brummel et al.13 reported relevant information by studying 20 patients. Pending the number of studied variables, a sample twice the size of sample by Brummel et al.13 was believed to be appropriate and 40 patients were used in the present study. Recorded variables were collected in three main areas. Patient characteristics: age; sex; cause of liver disease and MELD score; haemostatic variables: platelet count; factor V; vWF; factor VIII; AT and protein C activity; and free protein S antigen; TGT variables: lag-time (min) [defined as the time elapsing from activation of coagulation until generation of thrombin generation]; peak-thrombin (nanomol); time needed to reach the thrombin peak (time-to-peak, min); endogenous thrombin potential without (ETP) and with thrombomodulin addition [ETP(TM)] [defined as the area under the thrombin generation curve expressed as nanomol thrombin × minutes (nanomol min−1)]; and ROTEM-maximum clot firmness (ROTEM-MCF) from EXTEM, INTEM and FIBTEM analysis. Similarly as in previous studies conducted in patients with cirrhosis, cancer and procoagulant disorders, the ratios between ETP values obtained with-to-without thrombomodulin were calculated
. These ratios are regarded as reflecting the efficiency of thrombomodulin in the activation of protein C and are taken as in-vitro indices of the procoagulant imbalance (the greater the ratios, the higher the procoagulant imbalance).9–11,26,27
, MELD scores, coagulation factors and inhibitors and ROTEM-MCF values were tested. All of the results are given as the mean (SD) or median (range) after testing for normality. Throughout the study, factor V, which is not dependent on vitamin K input, was regarded as reflecting liver function. PT results, which are dependent on vitamin K input, were not reported. Correlations were tested using Spearman's rank correlation. A correlation coefficient (ρ) value of at least 0.5 was considered to be a strong correlation. When ρ values were within the 0.4 to 0.5 range, the correlation was regarded as weak. The Student's t-test was used whenever necessary after assessing normality. All data were analysed using JMP 8.0 software (SAS Institute, Grégy-sur-Yerres, France) with significance defined as a P value of 0.05 or less.
Demographic data and liver tests are summarised in Table 1. All degrees of liver dysfunction, as assessed by a wide range in MELD score values, were recorded.
Rotational thromboelastometry analysis
ROTEM-MCF values in EXTEM, INTEM and FIBTEM assays are displayed as scatter plots in Fig. 1. There was a wide distribution of MCF values within the normal and hypocoagulation ranges. The MCF values did not exceed the upper limit of the normal range and never suggested hypercoagulation.
Coagulation factors and inhibitors
Coagulation factors and platelet counts are displayed as scatter plots in Fig. 1. Out of 40 assessments, 34 protein C activity values, 36 AT values, as well as 32 platelet counts, were below normal ranges; 34 fibrinogen values were variously distributed within the normal ranges; 20 factor V values were below the normal range; 38 vWF values and 27 factor VIII values were above normal values; and most of the free protein S values were within the reference values.
Thrombin generation test assays
TGT variables are summarised in Table 2 and the distribution of ETP (TM) is displayed in Fig. 2. ETP and peak thrombin significantly decreased and time to peak and lag time significantly increased in the presence of thrombomodulin. ETP(TM) values were distributed within and above the normal ranges, indicating preserved coagulation and/or suggesting a potential for increased coagulation, further confirmed by the ratios
Correlations between coagulation factors and inhibitors, ROTEM-MCF in EXTEM, INTEM, FIBTEM assays and TGT variables are summarised in Table 3. Factor V was strongly correlated to platelet count, all coagulation factors and inhibitors synthesised in the liver (except for protein C), and MCF values. There was a strong inverse correlation between factor V and MELD scores. These results support the concept of preserved haemostasis and show that ROTEM-MCF values strongly reflect the degree of liver dysfunction in patients with cirrhosis. vWF (which is not synthesised in the liver) concentration and factor VIII concentration (which varies according to vWF) were not correlated to any variables. A weak to strong inverse correlation was found between
ratios and other variables, vWF and factor VIII excepted, suggesting that a procoagulant potential could be inversely correlated to the degree of liver dysfunction. The FIBTEM test and
ratios are not dependent on platelet counts and therefore correlations were not calculated. ETP (TM) was not correlated to any variables.
In patients with stable cirrhosis, this prospective investigation of haemostasis provided three pieces of primary information. The ROTEM-based clot strength measurement showed a hypocoagulation correlated to the degree of liver dysfunction. In contrast, the TGT measured with and without thrombomodulin showed preserved or even increased thrombin generation, indicating that coagulation is preserved and/or even suggesting a potential for hypercoagulability inversely correlated to the degree of liver dysfunction. Both the strong correlation between the levels of pro and anticoagulant proteins synthesised in the liver and the increase in vWF and factor VIII levels support the concept of preserved haemostasis.
In patients with cirrhosis, in view of the conflicting results previously provided in separate studies, on the one hand by TEG and ROTEM assays,1–3 and on the other hand by TGT and coagulation factor levels,4–12 an overall investigation of haemostasis was deemed necessary with a special focus on the limitations of ROTEM assessment. Currently, these three assessments of the haemostatic process do not investigate haemostasis at the same times or to the same extent.13,14,28 The ROTEM instrument includes independent measurement channels allowing the so-called extrinsic and intrinsic coagulation pathways to be studied, and finally the conversion of fibrinogen to fibrin at the very early phase of thrombin formation to be investigated.13,14,28 However, overall coagulation is no longer restricted to the concept of separate extrinsic and intrinsic pathways resulting in the thrombin-dependent conversion of fibrinogen to fibrin.9,13,14,28 The overall in-vivo coagulation process includes numerous feedback systems and interactions between cells, pro and anticoagulant plasmatic factors and thrombomodulin released by the endothelium.13,14,28 In particular, in order to exert its full activity, the natural anticoagulant protein C pathway must be activated by protein C complexed with thrombomodulin released by the endothelium.12–14,28 Furthermore, the assessment of plasmatic factor levels explores only specific stages of the clotting cascade in plasma outside of their biological environment.4,6,13,14 In contrast, TGT best mimics the in-vivo balance of pro- and anticoagulant proteins in plasma and the dynamics of thrombin generated in vivo.13,14,28 In TGT, similar to in-vivo conditions, TF and phospholipids trigger the haemostatic cascade and the addition of thrombomodulin fully activates the protein C anticoagulant system.13,14,28 The resulting thrombogram initially shows a short lag time followed by a burst in thrombin release (peak-thrombin), which precedes a decrease in thrombin level due to thrombin neutralisation.13,14,28 Importantly, the conversion of fibrinogen to fibrin, accounted for by ROTEM analysis, takes place during the lag time when more than 95% of thrombin, the key enzyme of the haemostatic process, remains to be generated, accounting for a limitation of ROTEM analysis.13,14,28
The findings of the present study are supported by previous clinical investigations.15–22,29–31 No relevant correlation has been reported between haemostatic disturbances and intraoperative bleeding in cirrhotic patients undergoing abdominal paracentesis, insertion of an intracranial pressure monitor, liver biopsy, minor surgery such as dental extraction and liver transplantation, a procedure that can nowadays be conducted without transfusion requirement.15–18 In patients with cirrhosis, spontaneous bleeding has never been related to any haemostatic disturbances.29–31 Spontaneous bleeding is regarded as being caused by haemodynamic alterations secondary to portal hypertension, endothelial dysfunction and endogenous heparinoid release with bacterial infections, or renal failure.29–31
Our findings confirm the limits of these three coagulation approaches and have direct implications for clinical practice. In patients with stable cirrhosis associated with liver dysfunction, ROTEM is a reliable index of the degree of liver failure and indicates hypocoagulation correlated to the degree of liver impairment,1–3 which is not confirmed by TGT assay in the presence of thrombomodulin.9,10–12 Consequently, ROTEM is not appropriate for assessing the coagulation status in cirrhotic patients with hypoprothrombinemia related to liver dysfunction. There is no evidence for a cut-off value for PT, coagulation factors or ROTEM indices to implement transfusion of fresh frozen plasma or coagulation factors in stable cirrhotic patients undergoing invasive procedures.10–12
Our findings have some limitations. Our ex-vivo data are not supported by an in-vivo study of periprocedural bleeding. This limitation is recorded in most studies of coagulation in patients with cirrhosis,1–12 partly because cirrhotic patients with liver dysfunction undergo invasive procedures reluctantly.32 ROTEM analysis was conducted on whole blood, while the TGT assays were conducted on platelet-poor plasma. In the present study, nine out of 40 investigated patients had platelet counts below 60 x 109 l−1. Indeed, one might argue that the investigation of such blood samples with low platelet count and/or potential platelet dysfunction could have altered our results, as previously suggested in patients with cirrhosis.11 This is not a limitation for our findings. ETP, assessed on platelet-poor plasmas following activation by TF and phospholipids and in the presence of thrombomodulin, has provided contributive information on thrombin capacity in patients with cirrhosis and in normal controls.10–12,27,28,33 Some additional comments are warranted. Weak correlations were recorded between protein C activity and most investigated variables in contrast with the strong correlations recorded with factor V and AT values. Indeed, factor V and AT are synthesised in the liver and are not influenced by confounding variables.4,5 In contrast, protein C synthesis is influenced by vitamin K absorption, inflammation and the activity of enzymes involved in its γ-carbonylation, which are sometimes altered in cirrhotic patients.4,5,34,35 Finally, patients were not checked for unusual prothrombotic profiles such as factor V Leiden, G20210A prothrombin gene mutation and myeloproliferative and antiphospholipid syndromes.20–22 These circumstances associated with thrombosis are unusual even in patients with liver cirrhosis, and therefore are not likely to have altered our results.20–22
Our results are supported by relevant methods. Stable cirrhotic patients with a wide range of liver dysfunction, resulting in a wide range of recorded haemostatic values allowing significant statistical correlations, were prospectively investigated. We excluded inflammatory and immunologic liver diseases, together with the combination of cancer, infection, recent red cell, platelet or factor transfusions and overt thrombosis. These circumstances have been associated with alterations in the levels of fibrinogen, factors V and VIII, vWF, platelet count and ROTEM and TGT tests.1,3,26,27,35 Patients were investigated under standard conditions, which were shown by Dargaud et al.33 to offer consistency in TGT results. The normal ranges for ETP values were assessed on normal pooled plasma and confirmed on healthy controls, as previously reported in similar investigations.9,10 The ETP ratios with-to-without thrombomodulin reflect the efficiency of thrombomodulin in the activation of protein C and were taken as in-vitro indexes of the procoagulant imbalance, in a manner similar to previous investigations conducted in cirrhotic patients with low protein C levels and/or resistance to protein C and in patients with cancer or carrying prothrombotic phenotypes.9,10–12,26,27 Lastly, ETP(TM) values were not correlated with any investigated variables because complete activation of the protein C system resulted in equalisation of ETP(TM) values.9,12,14
Finally, standard coagulation tests including platelet count and PT are not predictive of perioperative blood transfusion requirement in cirrhotic patients.15–18 Neither are ROTEM tests contributive in this respect or to assess when activated fibrinolysis starts to contribute to active bleeding.1,2,5,9 In contrast, TGT indicates that (a) thrombin generation is similar in patients with cirrhosis and in healthy controls and (b) a potential for thrombosis, correlated to the degree of liver dysfunction, is observed in cirrhotic patients even when the PT is prolonged provided that platelet count is above 50 x 109 l−1 and thrombomodulin is added to the assay.11 Consequently, clinical evaluation of intraoperative bleeding in the operative field remains a valid practice in patients with cirrhosis. As soon as bleeding is not obviously surgically related, local guidelines concerning administration of coagulation factors, platelets and antifibrinolytic drugs together with probabilistic antibiotic administration are warranted.31 Importantly, transfusion of fresh frozen plasma does not consistently increase coagulation factors in patients with cirrhosis.36 The potential for thrombosis correlated to the degree of liver dysfunction should not alter bleeding care.
In conclusion, in cirrhotic patients with various degrees of liver dysfunction, the correlated decrease in both coagulation factors and inhibitors supports the concept of preserved haemostasis towards normality. TGT assay conducted in the presence of thrombomodulin revealed preserved thrombin release and even suggested a potential for hypercoagulability inversely correlated to the degree of liver dysfunction. In contrast, measurement of ROTEM-based clot strength showed hypocoagulation correlated to the level of liver impairment and may lead to unnecessary coagulation factor transfusion. ROTEM may not be appropriate for haemostasis assessment in patients with cirrhosis. Practically, clinical evaluation of bleeding remains a valid practice in patients with cirrhosis and an adequate trigger to implement transfusion of coagulation factors and platelets, and for antifibrinolytic drug administration.
Acknowledgements relating to this article
Assistance with the study: none.
Financial support and sponsorship: this study was funded by the ARFAAR (Association for Supporting Research in Anaesthesia, Analgesia and Intensive Care). TEM Innovations (GmbH, Munich, Germany) provided some of the ROTEM reagents.
Conflict of interest: CMS provided some expert reports for Haemonetics and TEM Innovation with no relationship with the study.
Presentation: this work was presented in an abstract form at the 2014 meeting of the European Society of Anaesthesiology, Stockholm, Sweden, 31 May to 3 June 2014.
Comment from the editor: CMS is an Associate Editor of the European Journal of Anaesthesiology.
1. Ben-Ari Z, Panagou M, Patch D, et al. Hypercoagulability in patients with primary biliary cirrhosis and primary sclerosing cholangitis evaluated by thromboelastography. J Hepatol
2. Tripodi A, Primignani M, Chantarangkul V, et al. The coagulopathy of cirrhosis assessed by thromboelastometry and its correlation with conventional coagulation parameters. Thromb Res
3. Thalheimer U, Triantos CK, Samonakis DN, et al. A comparison of kaolin- activated versus nonkaolin-activated thromboelastography in native and citrated blood. Blood Coagul Fibrinolysis
4. Orloff MJ, Halasz NA, Lipman C, et al. The complications of cirrhosis of the liver. Ann Intern Med
5. Tripodi A, Mannucci M. The coagulopathy of chronic liver disease. N Engl J Med
6. Ordinas A, Escolar G, Cirera I, et al. Existence of a platelet-adhesion defect in patients with cirrhosis independent of hematocrit: studies under flow conditions. Hepatology
7. Lisman T, Bongers TN, Adelmeijer J, et al. Elevated levels of von Willebrand factor in cirrhosis support platelet adhesion despite reduced functional capacity. Hepatology
8. Rijken DC, Kock EL, Guimara AHC, et al. Evidence for an enhanced fibrinolytic capacity in cirrhosis as measured with two different global fibrinolysis tests. J Thromb Haemost
9. Kleinegris MC, Bos MH, Roest M, et al. Cirrhosis patients have a coagulopathy that is associated with decreased clot formation capacity. J Thromb Haemost
10. Tripodi A, Salerno F, Chantarangkul V, et al. Evidence of normal thrombin generation in cirrhosis despite abnormal conventional coagulation tests. Hepatology
11. Tripodi A, Primignani M, Chantarangkul V, et al. Thrombin generation in patients with cirrhosis: the role of platelets. Hepatology
12. Tripodi A, Primignani M, Lemma L, et al. Evidence that low protein C contributes to the procoagulant imbalance in cirrhosis. J Hepatol
13. Brummel KE, Paradis SG, Butenas S, Mann KG. Thrombin functions during tissue factor-induced blood coagulation. Blood
14. Dielis AW, Castoldi E, Spronk HM, et al. Coagulation factors and the protein C system as determinants of thrombin generation in a normal population. J Thromb Haemost
15. Massicotte L, Beaulieu D, Thibeault L, et al. Coagulation defects do not predict blood product requirements during liver transplantation. Transplantation
16. Goldaracena N, Méndez P, Quiñonez E, et al. Liver transplantation without perioperative transfusions single-center experience showing better early outcome and shorter hospital stay. J Transplant
17. Ewe K. Bleeding after liver biopsy does not correlate with indices of peripheral coagulation. Dig Dis Sci
18. Perdigao JP, de Almeida PC, Rocha TD, et al. postoperative bleeding after dental extraction in liver pretransplant patients. J Oral Maxillofac Surg
19. Dabbagh O, Oza A, Prakash S, et al. Coagulopathy does not protect against venous thromboembolism in hospitalized patients with chronic liver disease. Chest
20. Amitrano L, Guardascione MA, Brancaccio V, et al. Risk factors and clinical presentation of portal vein thrombosis in patients with liver cirrhosis. J Hepatol
21. Abe W, Ikejima K, Lang T, et al. Low molecular weight heparin prevents hepatic fibrogenesis caused by carbon tetrachloride in the rat. J Hepatol
22. Mahrshak N, Halfon P, Deutsch V, et al. Increased fibrosis progression rates in hepatitis C patients carrying the prothrombin G20210A mutation. World J Gastroenterol
23. Onaca NN, Levy MF, Sanchez EQ, et al. A correlation between the pretransplantation MELD score and mortality in the first two years after liver transplantation. Liver Transpl
24. Hemker HC, Giesen P, Al Dieri R, et al. Calibrated automated thrombin generation measurement in clotting plasma. Pathophysiol Haemost Thromb
25. Gronchi F, Perret A, Ferrari E, Marcucci CM, et al. Validation of rotational thromboelastometry during cardiopulmonary bypass: a prospective, observational in-vivo study. Eur J Anaesthesiol
26. Debaugnies F, Azerad MA, Noubouossié D, et al. Evaluation of the procoagulant activity in the plasma of cancer patients using a thrombin generation assay. Thromb Res
27. Dargaud Y, Trzeciak MC, Bordet JC, et al. Use of calibrated automated thrombinography ± thrombomodulin to recognise the prothrombotic phenotype. Thromb Haemost
28. Machlus KR, Colby EA, Wu JR, et al. Effects of tissue factor, thrombomodulin and elevated clotting factor levels on thrombin generation in the calibrated automated thrombogram. Thromb Haemost
29. Garcia-Tsao G, Bosch J. Management of varices and variceal haemorrhage in cirrhosis. N Engl J Med
30. Israelsen ME, Gluud LL, Krag A. Acute kidney injury and hepatorenal syndrome in cirrhosis. J Gastroenterol Hepatol
31. Montalto P, Vlachogiannakos J, Cox DJ, et al. Bacterial infection in cirrhosis impairs coagulation by a heparin effect: a prospective study. J Hepatol
32. Northup PG, Wanacker RC, Lee VD, et al. Model for end-stage liver disease (MELD) predicts nontransplant surgical mortality in patients with cirrhosis. Ann Surg
33. Dargaud Y, Luddington R, Gray E, et al. Effect of standardization and normalization on imprecision of calibrated automated thrombography: an international multicenter study. Br J Haematol
34. Aryal B, Jeong J, Rao VA. Doxorubicin-induced carbonylation and degradation of cardiac myosin binding protein C promote cardiotoxicity. Proc Natl Acad Sci U S A
35. Falanga A, Panova-Noeva M, Russo L. Procoagulant mechanisms in tumour cells. Best Pract Res Clin Haematol
© 2016 European Society of Anaesthesiology
36. Bartelmaos T, Chabanel A, Léger J, et al. Plasma transfusion in liver transplantation: a randomized, double-blind, multicenter clinical comparison of three virally secured plasmas. Transfusion