Proefschrift Muller

Coagulopathy and plasma transfusion

in critically ill patients

Marcella Müller

Coagulopathy and plasma transfusion in critically ill patients

Marcella M

üller

UITNODIGINGVoor het bijwonen van de

openbare verdediging van het proefschrift


door

Marcella Müller

op donderdag 4 september 2014

om 12:00 uur

in de Agnietenkapel van de Universiteit van AmsterdamOudezijds Voorburgwal 231

te Amsterdam

Na afloop bent u van hartewelkom op de receptie.

ParanimfenBoukje van DijkNelien Hylkema

[email protected]

Marcella MüllerJohannes Bildersstraat 35

2596 EE Den Haag06-41810633

[email protected]


Marcella Catharina Antoinetta Müller

Muller.indd 1 25-7-2014 14:40:17

Coagulopathy and plasma transfusion in critically ill patients,

Dissertation, University of Amsterdam, the Netherlands

This dissertation was prepared at the Department of Intensive Care Medicine and the

Laboratory of Experimental Intensive Care and Anesthesiology (L.E.I.C.A.), Academic Medical

Center, Amsterdam, the Netherlands

Copyright © M.C.A. Müller 2014

No part of this thesis may be reproduced, stored or transmitted in any form or by any means,

without prior permission of the author.

Cover artwork: “Onvoltooid” by Corinne Frencken-van den Berg (1942–1997)

Lay-out: Nicole Nijhuis, Gildeprint, Enschede, the Netherlands

Printed by: Gildeprint, Enschede, the Netherlands

ISBN/EAN: 9789461087263

Financial support by the Dutch Heart Foundation for the publication of this thesis is gratefully

acknowledged.

The publication of this thesis was also generously supported by the University of Amsterdam,

Stichting Gilles Hondius Foundation, Sanquin, Chipsoft B.V., TEM international and CSL

Behring.

Muller.indd 2 25-7-2014 14:40:17


ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctor

aan de Universiteit van Amsterdam

op gezag van de Rector Magnificus

prof. dr. D.C. van den Boom

ten overstaan van een door het college voor promoties ingestelde

commissie, in het openbaar te verdedigen in de Agnietenkapel

op donderdag 4 september 2014, te 12:00 uur

door Marcella Catharina Antoinetta Müller

geboren te Groningen

Muller.indd 3 25-7-2014 14:40:17

PromotiecommissiePromotor: Prof. dr. M.B. VroomCopromotor: Dr. N.P. Juffermans

Overige leden: Prof. dr. M.J. Schultz Prof. dr. M.W. Hollmann Prof. dr. J.C. Goslings Prof. dr. P.W. Kamphuisen Prof. dr. E. de Jonge Dr. S.S. Zeerleder

Faculteit der Geneeskunde

Muller.indd 4 25-7-2014 14:40:17

Voor mijn ouders

Muller.indd 5 25-7-2014 14:40:17

Muller.indd 6 25-7-2014 14:40:17

7

Contents

Chapter 1 General introduction and outline of this thesis 11

Part I Diagnosing coagulopathy in the critically ill: thromboelastometry

Chapter 2 The utility of thromboelastometry (ROTEM) or thromboelastography (TEG) in non-bleeding ICU patients 29

Chapter 3 Utility of thromboelastography and/or thromboelastometry in adults with sepsis: a systematic review 41

Chapter 4 Thromboelastometry and multiple organ failure in trauma patients 65

Part II Efficacy of plasma transfusion in critically ill patients

Chapter 5 Transfusion of fresh-frozen plasma in critically ill patients with a coagulopathy before invasive procedures: a randomized clinical trial 89

Chapter 6 Fresh frozen plasma transfusion fails to alter the hemostatic balance in critically ill patients with a coagulopathy 111

Chapter 7 Correlation of thromboelastometry with conventional hemostatic tests in critically ill patients with a coagulopathy treated with fresh frozen plasma 127

Chapter 8 Effect of transfusion of fresh frozen plasma on parameters of endothelial condition and inflammatory status in non-bleeding critically ill patients 143

Muller.indd 7 25-7-2014 14:40:18

Contents

8

Chapter 9 Clinicians’ attitude towards prophylactic transfusion of fresh frozen plasma – an evaluation of a multicenter randomized clinical trial 157

Part III Risks of transfusion in the critically ill: TRALI

Chapter 10 Risk factors for transfusion-related acute lung injury in ICU patients 171

Chapter 11 Contribution of damage-associated molecular patterns to transfusion-related acute lung injury in cardiac surgery 183

Chapter 12 Transfusion-related acute lung injury: a preventable syndrome? 201

Chapter 13 Prevention of immune-mediated transfusion-related acute lung injury; from bloodbank to patient 225

Chapter 14 Low risk TRALI donor strategies and the impact on the onset of transfusion-related acute lung injury; a meta-analysis 247

Chapter 15 Methylprednisolone fails to attenuate lung injury in a mouse model of transfusion-related acute lung injury 277

Chapter 16 The effect of C1-inhibitor in a murine model of transfusion-related acute lung injury 289

Chapter 17 Summary and general discussion 301

Chapter 18 Dutch summary 319

Muller.indd 8 25-7-2014 14:40:18

Contents

9

AppendicesAuthors and Affiliations 330Publications 334Acknowledgements 337Research portfolio 340Curriculum Vitae 342

Muller.indd 9 25-7-2014 14:40:18

Muller.indd 10 25-7-2014 14:40:18

Chapter 1

General introduction and outline of this thesis

Marcella C.A. Müller and Nicole P. Juffermans

Muller.indd 11 25-7-2014 14:40:18

Chapter 1

12

Background

Due to the high prevalence of coagulopathy among patients at the intensive care unit (ICU) [1], administration of Fresh Frozen Plasma (FFP) to critically ill patients is common practice. Thirteen per cent of ICU patients receives FFP during their stay at the ICU [2] and this can be up to 60% in specific ICU patient populations [3]. How-ever, knowledge regarding identifying coagulopathy as well as efficacy and adverse effects of FFP in the critically ill is scarce.This thesis aims to study the risk benefit balance of FFP transfusion. In order to opti-mize the risk benefit decision to transfuse, we aimed to assess whether identifica-tion of patients with a coagulopathy and increased bleeding risk could be improved. In addition, we studied the effectiveness of FFP transfusion to correct coagulopathy in critically ill patients and thereby mitigate the risk of bleeding. To further improve risk benefit assessment, we studied pathophysiology, treatment and prevention of transfusion related acute lung injury, which is the main adverse outcome of FFP transfusion.

Coagulopathy in critically ill patients

Pathophysiology, prevalence and outcomeCoagulopathy is often referred to as “a condition in which the blood’s ability to clot is impaired” [4], however this description is a simplification of reality. The coagulation system consists of three main components. The pro-coagulant elements include the endothelium, thrombocytes, individual coagulation factors and fibrinogen. The anti-coagulant system includes proteins C and S and antithrombin. The third component of coagulation is the fibrinolytic system (figure 1). In critical illness, these different components can be deranged in various ways, often resulting in an imbalance due to enhanced activation of coagulation and impaired inhibition of coagulation and fibrinolysis (figure 2). The result is a variable clinical picture with patients with an increased bleeding tendency (“consumption coagulopathy”) and those having dis-seminated intravascular coagulation (DIC) with the formation of (micro) vascular thrombi, associated with multiple organ failure.

Muller.indd 12 25-7-2014 14:40:18


13

Figure 1: Schematic and simplified overview of coagulation cascade, including the anticoagu-lant and fibrinolysis pathways.

The coagulation cascade (indicated in black) starts with endothelial injury, resulting in the formation of the tissue factor (TF) - factor VII (FVIIa) complex. Subsequently the conversion of prothrombin to thrombin results from the activation of factors X (FXa) and V (FVa). The whole process is enhanced via an amplification loop, thrombin is the key activator of this loop. Ultimately thrombin converses fibrinogen into fibrin resulting in a formation of a stable clot. The anticoagulant pathways are indicated in dashed lines. Activated protein C (APC) inactivates factors Va, while antithrombin (AT) blocks the action of multiple coagulation factors (including Xa and thrombin). Tissue factor pathway inhibitor (TFPI) inhibits the TF-FVIIa complex and hereby the stepwise activation of the coagulation cascade. The fibrinolytic system, responsible for clot degradation with the formation of fibrin degradation products (FDP), is indicated in grey. Thrombin activatable fibrinolysis inhibitor (TAFI) and plasminogen activator inhibitor (PAI-1) are the main inhibitors of fibrinolysis.

Muller.indd 13 25-7-2014 14:40:18

Chapter 1

14

Figure 2: Factors influencing the hemostatic balance in critically ill patients.

VWF = von Willebrand factorADAMTS-13 = a disintegrin and metalloproteinase with thrombospondin type 1, motif 13TFPI = tissue factor pathway inhibitorPAI-1 = plasminogen activator inhibitor type ITAFI = thrombin activatable fibrinolysis inhibitor

Table 1: Main causes of coagulopathy in critically ill patients

Causes of coagulopathy in critically ill patientsSepsisMultiple traumaBrain injuryMajor blood loss (e.g. gastro-intestinal, obstetric)Liver diseaseDisseminated intravascular coagulation Use of vitamin K antagonists before ICU admissionVitamin K deficiencyRenal failureCardiac surgeryThrombotic micro-angiopathies

Muller.indd 14 25-7-2014 14:40:18


15

Studies using conventional coagulation tests indicate that the prevalence of coagu-lation abnormalities in critically ill patients is high. Up to 30-66% of patients have an International Normalized Ratio (INR) of >1.5 or a prothrombin (PT) ratio of >1.5 [1,5] and 8 to 45% of patients develops thrombocytopenia during their intensive care unit (ICU) stay [6-9]. The most common causes of deranged coagulation are summarized in table 1 [4]. Of note, irrespective of the underlying disease, the occur-rence of both prolonged PT/INR [1] and thrombocytopenia [7,9] are independently associated with increased mortality in critically ill patients. Furthermore, in addition to a hypocoagulable state, up to 20% of critically ill patients develops DIC and the resulting hypercoagulable state is associated with multiple organ failure and high mortality (45-78%) [10].

Assessing coagulation status in critically ill patientsMost commonly used tests to assess coagulation status include assessment of plate-let count, coagulation times (activated partial thromboplastin time (aPTT), PT and INR) and levels of fibrinogen and d-dimers. Results from these tests only reflect a limited part of the haemostatic process. Therefore they fail to reflect in vivo hae-mostatic potential and hereby cannot reliably predict potential bleeding risk [11]. In addition, there is a lack of available tests for clinical use to detect a defective natural anticoagulant system and the same applies to markers of fibrinolysis [12]. Moreover, conventional coagulation assays lack the ability to detect a hypercoagulable state. To date, the ISTH DIC score [13], a composite of platelet count, PT prolongation, D-dimer and fibrinogen levels, is the only clinical way to diagnose DIC accompanied by a hypercoagulable state.In contrast to conventional coagulation tests, rotational thromboelastography assesses the whole process of clot formation and degradation. The resulting throm-boelastogram represents initiation of clot formation, fibrin formation and clot deg-radation [14]. In the past decade, evidence has shown that thromboelastography has additional value to rapidly detect depletion of coagulation factors and fibrino-gen in massive bleeding [15-17]. In critically ill patients with a suspected coagulop-athy, thromboelastography could improve identification of patients who have an increased bleeding risk and are in need of FFP transfusion.

Muller.indd 15 25-7-2014 14:40:18

Chapter 1

16

Also, enhanced fibrinolysis or hyperfibrinolysis, in particular after trauma, can be diagnosed using thromboelastography [18-20]. Moreover, it has well been shown that this bedside test is capable of detecting a hypercoagulable state [21-23]. Of note, there is no uniform definition for hypercoagulability, which renders the inter-pretation of data on clinical relevance of thromboelastography detected hyperco-agulability challenging. In addition, data from critically ill patients are scarce and reference standards for this population, with a high prevalence of coagulation abnormalities, are lacking. Hereby, despite its potential advantages, the additional value of this test in critically ill patients remains to be established.

Plasma transfusion in the critically ill

Current indications and use of plasma in critically ill patientsFresh frozen plasma (FFP) effectively corrects multiple clotting factor deficiencies and guidelines recommend its use in severe bleeding [24,25]. However, a number of audits have shown that a substantial amount of FFP is administered prophylactically to patients who have a coagulopathy, but lack signs of active bleeding [26-28]. Of note, evidence that prophylactic administration prevents bleeding complications, in particular after interventions in patients with a coagulopathy, is absent [29,30]. Despite the reported lack of evidence on effectiveness, inappropriate use of FFP is widespread. In the Netherlands, 80.000 FFP units are issued annually (www.san-quin.nl). Of note, critical care physicians exert the majority of these requests [26,27]. The following paradigm responsible for inappropriate use of FFP has been postu-lated: clinicians assume that (1) elevated PT/INR predicts an enhanced bleeding risk in the setting of a procedure, (2) pre-procedural FFP administration improves PT/INR values and (3) prophylactic FFP transfusion indeed results in fewer bleeding compli-cations (figure 3) [11,31-33].

Efficacy of FFP to correct coagulopathy in critically ill patientsA few small clinical trials have studied the efficacy of FFP in critically ill patients with a coagulopathy. These studies are hampered by different doses used and no assessment of the effect of FFP administration on occurrence of bleeding complica-

Muller.indd 16 25-7-2014 14:40:18


17

Figure 3: Paradigm on the use of fresh frozen plasma to treat or prevent bleeding [11,33].

PT = prothrombin timeINR = international normalized ratio

tions [34,35]. Judgement of results is further hampered by the fact that both actively bleeding and patients without bleeding were included.In assessment of efficacy of FFP, the dose is of importance. Interestingly, for pro-phylactic FFP transfusion the reported doses used in clinical practice are often less than 10 ml/kg [27,36]. These doses of FFP fail to correct marginally elevated INR values [37]. Despite these findings, physicians consider a small dose of FFP effec-tive to reduce risk of bleeding in coagulopathic patients (figure 3). Altogether, evi-dence supporting the efficacy of FFP to correct coagulopathy in critically ill patients is scarce.In addition to a lack of evidence on effectiveness, some reports suggest detrimental effects of FFP. In critically ill patients, FFP transfusion was shown to be an independ-ent risk factor for the occurrence of lung injury [38-41] and was associated with an increased ICU length of stay [40-42]. Also, transfusion of FFP has been linked to occurrence of infectious complications [43,44]. Taken together, the widespread use of FFP together with a lack of knowledge on effectiveness of FFP in the critically ill and the potential detrimental effects of FFP transfusion in these patients resulted in an expressed need for more and better evidence supporting the use of FFP in ICU patients [32,45].

Muller.indd 17 25-7-2014 14:40:19

Chapter 1

18

Risks of plasma transfusion in the critically ill: transfusion related acute lung injury

BackgroundDespite obvious clinical benefit in bleeding, FFP is repeatedly associated with the occurrence of lung injury in trauma, postoperative and critically ill patients [46-48]. We investigated the problem of lung injury following FFP transfusion in more detail in the last section of this thesis.Transfusion related acute lung injury (TRALI) is the leading cause of transfusion related morbidity and mortality [49,50] and is defined as the onset of acute lung injury (ALI) occurring within six hours after a blood transfusion. The ensuing pulmo-nary edema is characterized by worsening oxygenation and bilateral infiltrates on a chest radiograph [49]. TRALI most frequently occurs in at risk populations, such as critically ill and cardiac surgery patients, with incidences of 5-8% [39,48,51]. Attrib-utable morbidity and mortality is considerable. Up to 70% of TRALI patients requires mechanical ventilation [52] and up to a fifth of patients die [53]. In addition, in criti-cally ill patients, TRALI prolongs duration of mechanical ventilation, ICU length of stay and increases hospital and long-term mortality [39,51,54].

PathophysiologyPathophysiology of TRALI has not been fully elucidated yet. However, data acquired in the past decade have resulted in the concept of a ‘two-event’ model [55]. This model might explain the relatively high incidence of TRALI in the critically ill, as the first event composes of the patients’ underlying condition. An underlying inflamma-tory condition, such as sepsis, induces priming of pulmonary neutrophils and vascu-lar endothelium. The second event is the transfusion of a blood product containing antibodies or other substances. Subsequently, primed neutrophils are activated with the release of inflammatory cytokines, resulting in endothelial damage and the development of lung injury [55]. The second event can be immune mediated or non-immune mediated. Immune mediated TRALI results from blood products contain-ing human anti-neutrophil (HNA) or anti-leukocyte antibodies (HLA) that cross-react with recipients’ cognate antigens. Up to two-thirds of TRALI cases is thought to be immune mediated. In particular in plasma rich products, antibodies are frequently

Muller.indd 18 25-7-2014 14:40:19


19

Figure 4: The occurrence and severity of a TRALI reaction depends on the degree of predis-position of the recipient (grey box) and the strength of the neutrophil priming activity of the transfused product (white box). Edited from Bux et al. [60] and Vlaar [69].

present. A non-immune mediated TRALI reaction results from bioactive substances, which accumulate during storage of cell containing blood products. Implicated sub-stances include lysophosphatidylcholines (LysoPCs), non-polar lipids, inflammatory cytokines and soluble CD40 [56-59]. However, the precise pathways that lead to activation of primed neutrophils and subsequent endothelial damage are still largely unknown.In addition to the ‘two-event’ model, a threshold model has been postulated [60]. According to this hypothesis, the severity of patients’ underlying condition (‘first event’), determines the threshold for the second event [60,61]. This means that smaller amounts of antibodies or bioactive substances can induce a TRALI reaction in critically ill patients compared to more healthy individuals needing a transfusion (figure 4). Thereby, ICU patients may be at increased risk of acquiring TRALI. This underlines the need of a careful risk benefit assessment of the need of transfusion of FFP or any blood product in this population.

Prevention and treatmentAs about two-thirds of TRALI cases are immune-mediated, a substantial amount of reactions can be prevented if donors harbouring HLA or HNA antibodies are excluded from donation of plasma containing blood products. The most important

Muller.indd 19 25-7-2014 14:40:19

Chapter 1

20

reasons for occurrence of HLA or HNA antibodies within the donor population are previous pregnancies and a history of transfusion [62]. In particular multiple preg-nancies induces antibody formation and up to a third of multipara women develops HLA antibodies [62,63]. Based on this knowledge, various donor risk reduction strat-egies have been implemented by different blood suppliers. Strategies range from active screening of allo-exposed donors and deferral of antibody positive donors [64], deferral of all allo-exposed donors [65-67] and deferral of all female donors [53,68]. The effect of the abovementioned strategies on the onset of TRALI varies, while they also differ in impact on the loss of potential donors and workload for blood suppliers.Although preventive strategies have contributed to a reduction in occurrence of TRALI, cases still occur, in particular in the most severely ill patients, as they have the lowest threshold to develop a reaction. Despite the reported attributable morbid-ity and mortality, a pharmacological treatment is not available. Therefore, research into therapeutic strategies is highly warranted.

Outline of this thesis

This thesis focuses on the risk benefit balance of FFP transfusion. Improving iden-tification of patients with a coagulopathy and increased bleeding risk may aid in an improved risk benefit decision to transfuse. Therefore, we studied in depth the assessment of coagulopathy in critically ill patients. The second part focuses on the ability of FFP transfusion to correct coagulopathy in critically ill patients and thereby mitigate the risk of bleeding. To further improve risk benefit assessment, the occur-rence of TRALI as the main adverse outcome of FFP transfusion was investigated in detail using clinical and experimental studies on the pathophysiology, treatment and prevention of TRALI.

Part I investigates the ability of thromboelastometry to detect coagulopathic derangements in critically ill patient populations. Chapter 2 gives an overview of available literature on the use of thromboelastometry in the intensive care setting. Chapter 3 is a systematic review of literature on the diagnostic and prognostic value of thromboelastometry in patients with sepsis. These chapters are followed by large

Muller.indd 20 25-7-2014 14:40:19


21

cohort study on the value of coagulopathic profiles as detected by thromboelasto-metry to assess outcome after major trauma (chapter 4). Other hemostatic tests, which assess coagulopathy, in addition to INR, are studied in part II. The efficacy of FFP in non-bleeding critically ill patients with a coagulopathy is another focus of part II. First the results of a multicentre randomized clinical trial on the ability of FFP to prevent bleeding complications in critically ill patients with a coagulopathy who need to undergo an intervention (TOPIC trial) are discussed (chapter 5). Chapters 6 and 7 describe the ability of hemostatic tests, including thromboelastometry, on assessment of coagulopathy and the in vivo effects of FFP transfusion on the coagu-lation system of critically ill patients. Effect of FFP on levels of individual coagulation factors, natural anticoagulants, thrombin generation and the fibrinolysis system are also described. In addition the effect of FFP transfusion on thromboelastometry is discussed. Furthermore, the effect of FFP transfusion on the inflammatory response and endothelial function in critically ill patients is discussed in chapter 8. Chapter 9 contains an evaluation of the TOPIC trial, as inclusion targets were not achieved in this trial despite a reported need for more evidence on the use of FFP in patients with a coagulopathy. The aim of this chapter was to identify constraints in conduct-ing intervention trials on the effectiveness of FFP in critically ill patients and to make recommendations for further research in this field.In part III, several aspects on the pathophysiology, treatment and prevention of TRALI are discussed. Chapter 10 reviews the risk factors for TRALI in critically ill patients. Chapter 11 is an observational study on the contribution of damage-asso-ciated molecular pattern molecules (DAMPs), which are thought to mediate host response to both infectious and non-infectious stimuli, to the development of TRALI after cardiac surgery. A general overview of potential strategies to reduce the risk of TRALI is given in chapter 12, while measures to specifically prevent immune-medi-ated TRALI are discussed in chapter 13. In addition, we performed a meta-analysis on the impact of low risk TRALI donor strategies for plasma containing blood prod-ucts on the onset of TRALI (chapter 14).Finally, chapters 15 and 16 describe the effects of administration of methylpredni-solone and C1 inhibitor in a murine model of TRALI.The results from all parts are summarized and discussed in chapter 17 (English) and chapter 18 (Dutch).

Muller.indd 21 25-7-2014 14:40:19

Chapter 1

22

References

1. Walsh TS, Stanworth SJ, Prescott RJ, et al.: Prevalence, management, and outcomes of critically ill patients with prothrombin time prolongation in United Kingdom intensive care units. Crit Care Med 2010;38: 1939-46

2. Stanworth SJ, Walsh TS, Prescott RJ, et al.: A national study of plasma use in critical care: clinical indications, dose and effect on prothrombin time. Crit Care 2011;15: R108

3. Reiter N, Wesche N, Perner A: The majority of patients in septic shock are transfused with fresh-frozen plasma. Dan Med J 2013;60: A4606

4. Hunt BJ: Bleeding and coagulopathies in critical care. N Engl J Med 2014;370: 847-595. Chakraverty R, Davidson S, Peggs K, et al.: The incidence and cause of coagulopathies in

an intensive care population. Br J Haematol 1996;93: 460-36. Williamson DR, Albert M, Heels-Ansdell D, et al.: Thrombocytopenia in critically ill

patients receiving thromboprophylaxis: frequency, risk factors, and outcomes. Chest 2013;144: 1207-15

7. Williamson DR, Lesur O, Tetrault JP, et al.: Thrombocytopenia in the critically ill: preva-lence, incidence, risk factors, and clinical outcomes. Can J Anaesth 2013;60: 641-51

8. Stanworth SJ, Walsh TS, Prescott RJ, et al.: Thrombocytopenia and platelet transfusion in UK critical care: a multicenter observational study. Transfusion 2013;53: 1050-8

9. Hui P, Cook DJ, Lim W, et al.: The frequency and clinical significance of thrombocytope-nia complicating critical illness: a systematic review. Chest 2011;139: 271-8

10. Singh B, Hanson AC, Alhurani R, et al.: Trends in the incidence and outcomes of dissemi-nated intravascular coagulation in critically ill patients (2004-2010): a population-based study. Chest 2013;143: 1235-42

11. Segal JB, Dzik WH: Paucity of studies to support that abnormal coagulation test results predict bleeding in the setting of invasive procedures: an evidence-based review. Trans-fusion 2005;45: 1413-25

12. Levi M, Meijers JC: DIC: which laboratory tests are most useful. Blood Rev 2011;25: 33-713. Taylor FB, Jr., Toh CH, Hoots WK, et al.: Towards definition, clinical and laboratory crite-

ria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost 2001;86: 1327-30

14. Reikvam H, Steien E, Hauge B, et al.: Thrombelastography. Transfus Apher Sci 2009;40: 119-23

15. Rourke C, Curry N, Khan S, et al.: Fibrinogen levels during trauma hemorrhage, response to replacement therapy, and association with patient outcomes. J Thromb Haemost 2012;10: 1342-51

16. Rugeri L, Levrat A, David JS, et al.: Diagnosis of early coagulation abnormalities in trauma patients by rotation thrombelastography. J Thromb Haemost 2007;5: 289-95

17. Holcomb JB, Minei KM, Scerbo ML, et al.: Admission rapid thrombelastography can replace conventional coagulation tests in the emergency department: experience with 1974 consecutive trauma patients. Ann Surg 2012;256: 476-86

Muller.indd 22 25-7-2014 14:40:19


23

18. Schochl H, Frietsch T, Pavelka M, et al.: Hyperfibrinolysis after major trauma: differen-tial diagnosis of lysis patterns and prognostic value of thrombelastometry. J Trauma 2009;67: 125-31

19. Theusinger OM, Wanner GA, Emmert MY, et al.: Hyperfibrinolysis diagnosed by rota-tional thromboelastometry (ROTEM) is associated with higher mortality in patients with severe trauma. Anesth Analg 2011;113: 1003-12

20. Raza I, Davenport R, Rourke C, et al.: The incidence and magnitude of fibrinolytic activa-tion in trauma patients. J Thromb Haemost 2013;11: 307-14

21. Kashuk JL, Moore EE, Sabel A, et al.: Rapid thrombelastography (r-TEG) identifies hypercoagulability and predicts thromboembolic events in surgical patients. Surgery 2009;146: 764-72

22. Park MS, Martini WZ, Dubick MA, et al.: Thromboelastography as a better indicator of hypercoagulable state after injury than prothrombin time or activated partial thrombo-plastin time. J Trauma 2009;67: 266-75

23. Taura P, Rivas E, Martinez-Palli G, et al.: Clinical markers of the hypercoagulable state by rotational thrombelastometry in obese patients submitted to bariatric surgery. Surg Endosc 2014; 28: 543-51

24. O’Shaughnessy DF, Atterbury C, Bolton MP, et al.: Guidelines for the use of fresh-frozen plasma, cryoprecipitate and cryosupernatant. Br J Haematol 2004;126: 11-28

25. De Backer D, Vandekerckhove B, Stanworth S, et al.: Guidelines for the use of fresh frozen plasma. Acta Clin Belg 2008;63: 381-90

26. Stanworth SJ, Grant-Casey J, Lowe D, et al.: The use of fresh-frozen plasma in England: high levels of inappropriate use in adults and children. Transfusion 2011;51: 62-70

27. Tinmouth A, Thompson T, Arnold DM, et al.: Utilization of frozen plasma in Ontario: a provincewide audit reveals a high rate of inappropriate transfusions. Transfusion 2013;53: 2222-9

28. Vlaar AP, in der Maur AL, Binnekade JM, et al.: A survey of physicians’ reasons to trans-fuse plasma and platelets in the critically ill: a prospective single-centre cohort study. Transfus Med 2009;19: 207-12

29. Stanworth SJ, Brunskill SJ, Hyde CJ, et al.: Is fresh frozen plasma clinically effective? A systematic review of randomized controlled trials. Br J Haematol 2004;126: 139-52

30. Yang L, Stanworth S, Hopewell S, et al.: Is fresh-frozen plasma clinically effective? An update of a systematic review of randomized controlled trials. Transfusion 2012;52: 1673-86; quiz

31. Watson DM, Stanworth SJ, Wyncoll D, et al.: A national clinical scenario-based survey of clinicians’ attitudes towards fresh frozen plasma transfusion for critically ill patients. Transfus Med 2011;21: 124-9

32. Puetz J: Fresh frozen plasma: the most commonly prescribed hemostatic agent. J Thromb Haemost 2013;11: 1794-9

33. Tinmouth A: Evidence for a rationale use of frozen plasma for the treatment and pre-vention of bleeding. Transfus Apher Sci 2012;46: 293-8

Muller.indd 23 25-7-2014 14:40:19

Chapter 1

24

34. Chowdhury P, Saayman AG, Paulus U, et al.: Efficacy of standard dose and 30 ml/kg fresh frozen plasma in correcting laboratory parameters of haemostasis in critically ill patients. Br J Haematol 2004;125: 69-73

35. Tinmouth A, Chatelain E, Fergusson D, et al.: A Randomized Controlled Trial of High and Standard Dose Fresh Frozen Plasma Transfusions in Critically Ill Patients. Transfusion 2008;48: 26A-7A

36. Hall DP, Lone NI, Watson DM, et al.: Factors associated with prophylactic plasma trans-fusion before vascular catheterization in non-bleeding critically ill adults with pro-longed prothrombin time: a case-control study. Br J Anaest 2012;109: 919-27

37. Holland LL, Brooks JP: Toward rational fresh frozen plasma transfusion: The effect of plasma transfusion on coagulation test results. Am J Clin Pathol 2006;126: 133-9

38. Khan H, Belsher J, Yilmaz M, et al.: Fresh-frozen plasma and platelet transfusions are associated with development of acute lung injury in critically ill medical patients. Chest 2007;131: 1308-14

39. Gajic O, Rana R, Winters JL, et al.: Transfusion-related acute lung injury in the critically ill: prospective nested case-control study. Am J Respir Crit Care Med 2007;176: 886-91

40. Gajic O, Rana R, Mendez JL, et al.: Acute lung injury after blood transfusion in mechani-cally ventilated patients. Transfusion 2004;44: 1468-74

41. Rana R, Fernandez-Perez ER, Khan SA, et al.: Transfusion-related acute lung injury and pulmonary edema in critically ill patients: a retrospective study. Transfusion 2006;46: 1478-83

42. Lauzier F, Cook D, Griffith L, et al.: Fresh frozen plasma transfusion in critically ill patients. Crit Care Med 2007;35: 1655-9

43. Sarani B, Dunkman WJ, Dean L, et al.: Transfusion of fresh frozen plasma in critically ill surgical patients is associated with an increased risk of infection. Crit Care Med 2008;36: 1114-8

44. Bochicchio GV, Napolitano L, Joshi M, et al.: Blood product transfusion and ventilator-associated pneumonia in trauma patients. Surg Infect 2008;9: 415-22

45. Blajchman MA, Glynn SA, Josephson CD, et al.: Clinical trial opportunities in Transfu-sion Medicine: proceedings of a National Heart, Lung, and Blood Institute State-of-the-Science Symposium. Transfus Med Rev 2010;24: 259-85

46. Dara SI, Rana R, Afessa B, et al.: Fresh frozen plasma transfusion in critically ill medical patients with coagulopathy. Crit Care Med 2005;33: 2667-71

47. Park PK, Cannon JW, Ye W, et al.: Transfusion strategies and development of acute res-piratory distress syndrome in combat casualty care. J Trauma Acute Care Surg 2013;75: S238-46

48. Vlaar AP, Hofstra JJ, Determann RM, et al.: The incidence, risk factors, and outcome of transfusion-related acute lung injury in a cohort of cardiac surgery patients: a prospec-tive nested case-control study. Blood 2011;117: 4218-25

49. Goldman M, Webert KE, Arnold DM, et al.: Proceedings of a consensus conference: towards an understanding of TRALI. Transfus Med Rev 2005;19: 2-31

Muller.indd 24 25-7-2014 14:40:19


25

50. Holness L, Knippen MA, Simmons L, et al.: Fatalities caused by TRALI. Transfus Med Rev 2004;18: 184-8

51. Vlaar AP, Binnekade JM, Prins D, et al.: Risk factors and outcome of transfusion-related acute lung injury in the critically ill: a nested case-control study. Crit Care Med 2010;38: 771-8

52. Popovsky MA, Moore SB: Diagnostic and pathogenetic considerations in transfusion-related acute lung injury. Transfusion 1985;25: 573-7

53. Chapman CE, Stainsby D, Jones H, et al.: Ten years of hemovigilance reports of transfu-sion-related acute lung injury in the United Kingdom and the impact of preferential use of male donor plasma. Transfusion 2009;49: 440-52

54. Li G, Kojicic M, Reriani MK, et al.: Long-term survival and quality of life after transfusion-associated pulmonary edema in critically ill medical patients. Chest 2010;137: 783-9

55. Silliman CC: The two-event model of transfusion-related acute lung injury. Crit Care Med 2006;34: S124-S31

56. Silliman CC, Paterson AJ, Dickey WO, et al.: The association of biologically active lipids with the development of transfusion-related acute lung injury: a retrospective study. Transfusion 1997;37: 719-26

57. Silliman CC, Voelkel NF, Allard JD, et al.: Plasma and lipids from stored packed red blood cells cause acute lung injury in an animal model. J Clin Invest 1998;101: 1458-67

58. Silliman CC, Moore EE, Kelher MR, et al.: Identification of lipids that accumulate during the routine storage of prestorage leukoreduced red blood cells and cause acute lung injury. Transfusion 2011;51: 2549-54

59. Khan SY, Kelher MR, Heal JM, et al.: Soluble CD40 ligand accumulates in stored blood components, primes neutrophils through CD40, and is a potential cofactor in the devel-opment of transfusion-related acute lung injury. Blood 2006;108: 2455-62

60. Bux J, Sachs UJ: The pathogenesis of transfusion-related acute lung injury (TRALI). Br J Haematol 2007;136: 788-99

61. Vlaar AP, Wolthuis EK, Hofstra JJ, et al.: Mechanical ventilation aggravates transfu-sion-related acute lung injury induced by MHC-I class antibodies. Intensive Care Med 2010;36: 879-87

62. Triulzi DJ, Kleinman S, Kakaiya RM, et al.: The effect of previous pregnancy and transfu-sion on HLA alloimmunization in blood donors: implications for a transfusion-related acute lung injury risk reduction strategy. Transfusion 2009;49: 1825-35

63. Densmore TL, Goodnough LT, Ali S, et al.: Prevalence of HLA sensitization in female apheresis donors. Transfusion 1999;39: 103-6

64. Kleinman S, Grossman B, Kopko P: A national survey of transfusion-related acute lung injury risk reduction policies for platelets and plasma in the United States. Transfusion 2010;50: 1312-21

65. Strong D.M SLK: Association Bulletin #06-07: Transfusion Related Lung Injury. Bethesda (MD):AABB, 2006.

66. Jutzi M, Levy G, Taleghani BM: Swiss Haemovigilance Data and Implementation of Meas-ures for the Prevention of Transfusion Associated Acute Lung Injury (TRALI). Transfus Med Hemother 2008;35: 98-101

Muller.indd 25 25-7-2014 14:40:19

Chapter 1

26

67. Eder AF, Herron RM, Jr., Strupp A, et al.: Effective reduction of transfusion-related acute lung injury risk with male-predominant plasma strategy in the American Red Cross (2006-2008). Transfusion 2010;50: 1732-42

68. Wiersum-Osselton JC, Middelburg RA, Beckers EA, et al.: Male-only fresh-frozen plasma for transfusion-related acute lung injury prevention: before-and-after compara-tive cohort study. Transfusion 2010;51: 1278-83

69. Vlaar AP: Transfusion-related acute lung injury in the critically ill; Faculty of Medicine. Amsterdam, the Netherlands, University of Amsterdam, 2010.

Muller.indd 26 25-7-2014 14:40:20

Part I Diagnosing coagulopathy in the criti cally ill:

thromboelastometry

Muller.indd 27 25-7-2014 14:40:20

Muller.indd 28 25-7-2014 14:40:20

Chapter 2

The utility of thromboelastometry (ROTEM) or

thromboelastography (TEG) in non-bleeding ICU patients

Kirsten Balvers, Marcella C.A. Müller, Nicole P. Juffermans

Annual Update in Intensive Care and Emergency Medicine 2014; 583-591

Muller.indd 29 25-7-2014 14:40:20

Chapter 2

30

Introduction

A hypocoagulable state is highly prevalent in critically ill patients. An International Normalized Ratio (INR) of >1.5 occurs in 30% of patients, associated with increased mortality [1]. Also, of critically ill patients, up to 40% develops thrombocytopenia during their intensive care unit (ICU) stay [2-4], associated with increased length of stay, need for transfusion of blood products and increased mortality [5]. A hyper-coagulable state is also associated with adverse outcome, as well as with increased thrombo-embolic events [6]. Disseminated intravascular coagulation (DIC) develops in 10 to 20% of ICU patients. A hypercoagulable state contributes to organ failure and is associated with a high mortality, ranging from 45% to 78% [7].Coagulopathy is thought to result from an imbalance between activation of coagula-tion and impaired inhibition of coagulation and fibrinolysis. Activation is triggered by tissue factor, which is expressed in reaction to cytokines or endothelial damage. Impaired inhibition of coagulation is the consequence of reduced plasma levels of antithrombin (AT), depressed activity of the protein C system and decreased levels of tissue factor pathway inhibitor (TFPI). A decrease in the fibrinolytic system is due to increased levels of plasminogen activator inhibitor type 1 (PAI-1) [8,9]. This dis-turbance between components of the coagulation system leads to a variable clinical picture, ranging from patients with an increased bleeding tendency (hypocoagulable state) to those with DIC with (micro-) vascular thrombosis (hypercoagulable state).Assessment of coagulation status in patients is complex. Global coagulation tests, including activated partial thromboplastin time (aPTT) and prothrombin time (PT), are used clinically. However, these tests are of limited value and their ability to accu-rately reflect in vivo hypocoagulable state is questioned [10]. Also, aPTT/PT reflects a part of the coagulation system and does not provide information on the full balance between coagulation, anti-coagulation and fibrinolysis. Hypercoagulable state can be assessed by increased levels of d-dimers, but specificity is limited [10]. Impaired function of the anticoagulant system can be diagnosed by measuring plasma levels of naturally occurring anticoagulant factors AT, protein C and TFPI. However, these are not readily available for clinical use. Apart from the DIC score, there are no diag-nostic tests which evaluate a hypercoagulable state. Also, markers of the activity of the fibrinolytic system are not used at the bedside [10].

Muller.indd 30 25-7-2014 14:40:20

Thromboelastography or thromboelastometry in non-bleeding ICU patients

31

TEG/ROTEM tests

Rotational thromboelastography (TEG/ROTEM) is a point of care test, which evalu-ates whole clot formation and degradation. The thromboelastogram arises through movement of the cup (TEG) or the pin (ROTEM). As fibrin forms between the cup and the pin, this movement is influenced and converted to a specific trace. The trace reflects different phases of the clotting process. Major parameters are R (reaction/clotting) time, the period from the initiation of the test until the beginning of clot formation. K-time is the period from the start of the clot formation until the curve reaches an amplitude of 20 mm. Kinetics of fibrin formation and cross-linking is expressed by the α-angle, which is the angle between the baseline and the tangent to the TEG/ROTEM curve. Clot strength is represented by the maximal amplitude (MA) of the trace. The degree of fibrinolysis is reflected by the difference between the maximal amplitude and the amplitude measured after 30 and/or 60 minutes. To describe these visco-elastic changes, both systems have their own terminol-ogy (table 1). Both generate similar data. The technique is developed in the 1940s, but until recent, clinical application has been limited. However, technical develop-ments have led to standardization and improved reproducibility of the method [11]. Also, the availability for bedside evaluation and a changing view regarding the use of blood and haemostatic therapy in massive bleeding, have both contributed to a renewed interest in this technique.

Table 1: TEG and ROTEM parameters

ROTEM TEGTime to initial fibrin formation CT RClot strengthening, rapidity of fibrin build up CFT K

a aClot strength, represents maximum dynamics of fibrin and platelet bonding

MCF MA

Clot breakdown,fibrinolysis at fixed time (min)

LI30, LI45, LI60

CL30, CL60

TEG/ROTEM may also facilitate diagnosis of clotting abnormalities in the critically ill. Detecting a hypocoagulable state, TEG/ROTEM may be a useful tool in the assess-ment of the risk of bleeding peri-operatively or prior to an invasive procedure. This

Muller.indd 31 25-7-2014 14:40:20

Chapter 2

32

could lead to a more tailored transfusion strategy, with an efficient use of blood products.Also, TEG/ROTEM may diagnose a hypercoagulable state. With TEG/ROTEM, a hyper-coagulable state can be detected by high maximal amplitude (MA), shortened reac-tion time, increased alfa angle and total cloth strength G (defined as (5000xMA)/(100-MA), table 2). Assessment of a hypercoagulable state could lead to prognos-tication of multiple organ failure (MOF) and risk for thrombo-embolic events. Also, another potential advantage could be a more tailor made administration of thera-pies that interfere with the coagulation system. Difficulties in identifying responders from non-responders may in part have contributed to conflicting results from trials evaluating the effect of strategies that interfere with the coagulation system [12-15].

Table 2: Changes in hypercoagulable state and hypocoagulable state of ROTEM and TEG parameters

Parameters Hypercoagulable state Hypocoagulable stateReaction time, R or CT Shortened ProlongedClot formation time, K or CFT Shortened ProlongedAlpha angle, Angle or α Increased DecreasedMaximum amplitude, MA or MCF Increased Decreased

Utility of TEG/ROTEM to detect sepsis-induced coagulopathyROTEM clearly demonstrates a hypercoagulable state during endotoxemia [16]. In vitro, endotoxin-induced hypercoagulability was demonstrated with TEG. In experi-ments where LPS was infused in healthy volunteers, a hypercoagulable state meas-ured by TEG had a strong correlation with plasma levels of prothrombin fragments F1+2 [17,18]. In sepsis patients however, TEG/ROTEM measurements have shown dif-ferential results. Several studies observed no changes in parameters [19-22], other studies reported a hypercoagulable [23] or hypocoagulable state [24]. A few stud-ies also reported patients showing both a hyper- and hypocoagulable state [25-28]. Taken together, results are heterogeneous. Also, there is a lack of clarity on interpre-tation of the test results.To date, no studies have compared conventional coagulation tests such as PT/aPTT to TEG/ROTEM in sepsis patients. However, the utility of thromboelastography to detect DIC has been evaluated. It seems that thromboelastography can predict

Muller.indd 32 25-7-2014 14:40:20


33

DIC. Patients with DIC present with hypocoagulable state [26]. This may be due to a decrease in coagulation factors used for formation of micro thrombi. In line with this, sepsis patients who met the ISTH DIC criteria showed a hypocoagulable state when compared to healthy controls, while patients without DIC showed a non-signif-icant trend towards hypercoagulation [25]. Also, patients with an underlying disease known to be associated with DIC and ISTH DIC scores ≥5 had significantly prolonged reaction and K times and decreased alpha-angle and MA (signs of a hypocoagula-ble state) compared to patients with low ISTH DIC scores. The authors developed a score, defined as the total number of parameters (R, K, MA, and alpha) that were deranged in the direction of a hypocoagulable state. With this score, the discrimina-tory value of thromboelastometry to detect DIC improved [29]. Impaired fibrinolysis in sepsis may contribute to a hypercoagulable state. Inhibition of the fibrinolytic system was found to discriminate sepsis from postoperative controls [19,28,30].In terms of prognostication, a hypercoagulable state was not found to be a predic-tor of outcome. In contrast, the finding of a hypocoagulable state was repeatedly shown to be associated with a poor outcome. The TEG MA value is an independent predictor for 28-day mortality on admission [27]. Hospital mortality was predicted by a hypocoagulable state due to a deficit in thrombin generation [30]. A hypoco-agulable state measured with TEG is found to be associated with a pro-inflamma-tory response [19,24]. Also, the degree of a hypocoagulable state is associated with severity of organ failure in sepsis [19,22].Taken together, results are heterogeneous. Timing of measurements may be rel-evant to these observations, as a hypocoagulable state may be more outspoken in the acute phase of sepsis and return to normal values towards discharge of ICU, or even to enhanced clot formation.

Use of TEG/ROTEM to guide anticoagulant treatment in sepsis patientsIn sepsis, activation of coagulation is a crucial step in the pathophysiological cascade of sepsis, with concomitant low levels of circulating natural anticoagulants [8,9]. From this perspective, various treatment modalities that interfere with the coag-ulation system have been studied (e.g. rhAPC, AT and heparin) [12-15]. However, efficacy has been questioned. It can be hypothesized that TEG/ROTEM may help to identify patients likely to respond to therapies that target coagulopathy. To date,

Muller.indd 33 25-7-2014 14:40:20

Chapter 2

34

there are no studies which have addressed this question. Only a few small patient series evaluated TEG/ROTEM measurement during anticoagulant medication. ROTEM parameters did not change during anticoagulant medication. Also, treat-ment with antithrombin did not induce changes in the ROTEM measurements [23].

Use of TEG/ROTEM in patients with induced hypothermiaInduced hypothermia is a common therapy in survivors of a cardiac arrest [31-33]. However, hypothermia is associated with coagulopathy, prolongation of aPTT and PT [33,34] and an increased risk of bleeding [35]. A test that reliably detects hypo-thermia induced coagulopathy would be helpful in identifying patients who have an increased bleeding risk while being cooled and sedated. Unfortunately, little is known about the value of TEG/ROTEM in these patients.Spiel et al. observed that ROTEM measurements showed a prolonged CT at 1 hour after infusion of 4°C cold crystalloid solution. All other parameters remained within reference values. An important limitation of this study is that all measurements were performed at 37°C [33]. TEG parameters were evaluated also in patients after cardiac arrest. On the contrary, the TEG was performed at isothermal conditions and a hypocoagulable state was detected by TEG [36].

Use of TEG/ROTEM in patients with brain injuryAfter severe traumatic brain injury and neurosurgery, up to 45% of patients develop a coagulopathy [37-39]. Given the serious consequences of intracranial bleeding, instant assessment of coagulation status is desirable. Two small trials have stud-ied the value of TEG to detect coagulopathy in these patients, which mostly found test results within reference values. However, the functional response of platelets as measured in a platelet mapping™ (TEG-PM) assay, was significantly lower in brain injury patients than in control groups, with a particular low response in those patients who developed bleeding complications [40]. Furthermore, a hypocoagula-ble state on admission to the ICU is associated with worse outcome in patients with traumatic brain injury and intracranial bleeding [41].

Muller.indd 34 25-7-2014 14:40:20


35

Utility of TEG/ROTEM to detect a hypercoagulable state and prognosticate organ failure in trauma patientsPatients who survive the acute phase of trauma are prone to develop a hyperco-agulable state with increased risk for thrombo-embolic events and DIC [1]. Conven-tional coagulation tests are not able to detect such a hypercoagulable state. Also, there is debate as to whether the syndrome DIC is applicable to coagulation abnor-malities in trauma. With TEG/ROTEM, a hypercoagulable state can be detected by high maximal amplitude (MA) and shortened reaction time (table 1). Several reports demonstrate a hypercoagulable state in severely injured patients with TEG/ROTEM. In trauma and burn patients admitted to the ICU, TEG was found to be more sensi-tive in detecting a hypercoagulable state than conventional clotting assays [42,43]. A high MA was found to be an independent contributor of mortality in multiple logistic regression analysis [42]. A hypercoagulable state measured by TEG predicted the development of thrombo-embolic events in trauma patients [44] although not all studies have confirmed this finding [45]. It should be noted that the finding of a hypercoagulable state is not specific for venous thrombo embolism.A study on the use of ROTEM to prognosticate the occurrence of multiple organ failure in a cohort of trauma patients is currently underway.

Considerations

In several non-bleeding critically ill patient populations, evidence supporting the use of TEG/ROTEM to diagnose a hypocoagulable or hypercoagulable state is limited at this stage, mostly because of heterogeneity of the included studies in design, use of control groups and chosen endpoints. Heterogeneity of results can also be caused by differences in disease severity, as changes were more outspoken during severe illness. Timing of TEG/ROTEM measurements may greatly influence results, as coagulopathy is a dynamic process, e.g. evolving from subtle activation of coagu-lation to overt DIC in sepsis and from a hypocoagulable to a hypercoagulable state in trauma. Performing sequential measurements will probably provide better insight in the development of coagulation derangements.Another important issue is that no uniform definitions exist of a hypocoagulable and a hypercoagulable state. Reference values for non-bleeding patients with disorders

Muller.indd 35 25-7-2014 14:40:20

Chapter 2

36

of coagulation are not widely assessed and cut off values are often not defined in studies. To compare patient categories and possibly investigate therapeutic inter-ventions in the coagulation system, validated universal reference values and defini-tions are essential. A study on TEG reference intervals has been recently completed (NCT01357928). Presumably, as patients groups are relatively small, evaluation of larger patient groups may yield more clear results.

Conclusion

TEG/ROTEM can detect coagulopathy in the critically ill. Whether these tests are useful as diagnostic tools remains to be investigated when reference values and clear definitions have been established.TEG/ROTEM may be useful for prognostication of outcome. A hypocoagulable status seems to be an independent predictor for organ failure and mortality in sepsis, also after correction for disease severity. In patients with brain injury, a hypocoagulable state on admission to the ICU is also associated with worse outcome. In patients who survive the acute phase of trauma, a hypercoagulable state as detected by TEG/ROTEM is a common finding. These tests could be helpful in identifying those patients at risk for thrombo-embolic complications, as a hypercoagulable state pre-dicted the development of thrombo-embolic events in the majority of studies. Fur-ther research on this topic is forthcoming.

Muller.indd 36 25-7-2014 14:40:20


37

References


2. Crowther MA, Cook DJ, Meade MO, et al.: Thrombocytopenia in medical-surgical criti-cally ill patients: prevalence, incidence, and risk factors. J Crit Care 2005;20: 348-353

3. Strauss R, Wehler M, Mehler K, et al.: Thrombocytopenia in patients in the medical intensive care unit: bleeding prevalence, transfusion requirements, and outcome. Crit Care Med 2002;30: 1765-1771

4. Vanderschueren S, De WA, Malbrain M, et al.: Thrombocytopenia and prognosis in intensive care. Crit Care Med 2000;28: 1871-1876

5. Hui P, Cook DJ, Lim W, et al.: The frequency and clinical significance of thrombocytope-nia complicating critical illness: a systematic review. Chest 2011;139: 271-278

6. Shackford SR, Davis JW, Hollingsworth-Fridlund P, et al.: Venous thromboembolism in patients with major trauma. Am J Surg 1990;159: 365-369

7. Singh B, Hanson AC, Alhurani R, et al.: Trends in the incidence and outcomes of dissemi-nated intravascular coagulation in critically ill patients (2004-2010): a population-based study. Chest 2013;143: 1235-1242

8. Dempfle CE: Coagulopathy of sepsis. Thromb Haemost 2004;91: 213-2249. Levi M, Ten Cate H: Disseminated intravascular coagulation. N Engl J Med 1999;341:

586-59210. Levi M, Meijers JC: DIC: which laboratory tests are most useful. Blood Rev 2011;25:

33-3711. Reikvam H, Steien E, Hauge B, et al.: Thrombelastography. Transfus Apher Sci 2009;40:

119-12312. Bernard GR, Vincent JL, Laterre PF, et al.: Efficacy and safety of recombinant human

activated protein C for severe sepsis. N Engl J Med 2001;344: 699-70913. Afshari A, Wetterslev J, Brok J, et al.: Antithrombin III in critically ill patients: systematic

review with meta-analysis and trial sequential analysis. BMJ 2007;335: 1248-125114. Abraham E, Reinhart K, Opal S, et al.: Efficacy and safety of tifacogin (recombinant

tissue factor pathway inhibitor) in severe sepsis: a randomized controlled trial. JAMA 2003;290: 238-247

15. Jaimes F, De La Rosa G, Morales C, et al.: Unfractioned heparin for treatment of sepsis: A randomized clinical trial (The HETRASE Study). Crit Care Med 2009;37: 1185-1196

16. Schochl H, Solomon C, Schulz A, et al.: Thromboelastometry (TEM) findings in dissemi-nated intravascular coagulation in a pig model of endotoxinemia. Mol Med 2011;17: 266-272

17. Spiel AO, Mayr FB, Firbas C, et al.: Validation of rotation thrombelastography in a model of systemic activation of fibrinolysis and coagulation in humans. J Thromb Haemost 2006;4: 411-416

Muller.indd 37 25-7-2014 14:40:20

Chapter 2

38

18. Zacharowski K, Sucker C, Zacharowski P, et al.: Thrombelastography for the monitor-ing of lipopolysaccharide induced activation of coagulation. Thromb Haemost 2006;95: 557-561

19. Brenner T, Schmidt K, Delang M, et al.: Viscoelastic and aggregometric point-of-care testing in patients with septic shock - cross-links between inflammation and haemosta-sis. Acta Anaesthesiol Scand 2012;56: 1277-1290

20. Durila M, Kalincik T, Jurcenko S, et al.: Arteriovenous differences of hematological and coagulation parameters in patients with sepsis. Blood Coagul Fibrinolysis 2010;21: 770-774

21. Altmann DR, Korte W, Maeder MT, et al.: Elevated cardiac troponin I in sepsis and sep-tic shock: no evidence for thrombus associated myocardial necrosis. PLoS One 2010;5: e9017

22. Daudel F, Kessler U, Folly H, et al.: Thromboelastometry for the assessment of coagula-tion abnormalities in early and established adult sepsis: a prospective cohort study. Crit Care 2009;13: R42

23. Gonano C, Sitzwohl C, Meitner E, et al.: Four-day antithrombin therapy does not seem to attenuate hypercoagulability in patients suffering from sepsis. Crit Care 2006;10: R160

24. Viljoen M, Roux LJ, Pretorius JP, et al.: Hemostatic competency and elastase-alpha 1-proteinase inhibitor levels in surgery, trauma, and sepsis. J Trauma 1995;39: 381-385

25. Sivula M, Pettila V, Niemi TT, et al.:Thromboelastometry in patients with severe sepsis and disseminated intravascular coagulation. Blood Coagul Fibrinolysis 2009;20: 419-426

26. Collins PW, Macchiavello LI, Lewis SJ, et al.: Global tests of haemostasis in critically ill patients with severe sepsis syndrome compared to controls. Br J Haematol 2006;135: 220-227

27. Ostrowski SR, Windelov NA, Ibsen M, et al.: Consecutive thrombelastography clot strength profiles in patients with severe sepsis and their association with 28-day mor-tality: a prospective study. J Crit Care 2013;28: 317-11

28. Adamzik M, Langemeier T, Frey UH, et al.: Comparison of thrombelastometry with sim-plified acute physiology score II and sequential organ failure assessment scores for the prediction of 30-day survival: a cohort study. Shock 2011;35: 339-342

29. Sharma P, Saxena R: A novel thromboelastographic score to identify overt dissemi-nated intravascular coagulation resulting in a hypocoagulable state. Am J Clin Pathol 2010;134: 97-102

30. Massion PB, Peters P, Ledoux D, et al.: Persistent hypocoagulability in patients with sep-tic shock predicts greater hospital mortality: impact of impaired thrombin generation. Intensive Care Med 2012;38: 1326-1335

31. Hypothermia after Cardiac Arrest Study Group: Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346: 549-556

32. Bernard SA, Gray TW, Buist MD, et al.: Treatment of comatose survivors of out-of- hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346: 557-563

Muller.indd 38 25-7-2014 14:40:20


39

33. Spiel AO, Kliegel A, Janata A, et al.: Hemostasis in cardiac arrest patients treated with mild hypothermia initiated by cold fluids. Resuscitation 2009;80: 762-765

34. Reed RL, Bracey AW, Jr., Hudson JD, et al.: Hypothermia and blood coagulation: dissocia-tion between enzyme activity and clotting factor levels. Circ Shock 1990;132: 141-152

35. Nielsen N, Sunde K, Hovdenes J, et al.: Adverse events and their relation to mortality in out-of-hospital cardiac arrest patients treated with therapeutic hypothermia. Crit Care Med 2011;39: 57-64

36. Cundrle I, Jr., Sramek V, Pavlik M, et al.: Temperature corrected thromboelastography in hypothermia: is it necessary? Eur J Anaesthesiol 2013;30: 85-89

37. Sun Y, Wang J, Wu X, et al.: Validating the incidence of coagulopathy and disseminated intravascular coagulation in patients with traumatic brain injury--analysis of 242 cases. Br J Neurosurg 2011;25: 363-368

38. Lustenberger T, Talving P, Kobayashi L, et al.: Time course of coagulopathy in isolated severe traumatic brain injury. Injury 2010;41: 924-928

39. Stein SC, Smith DH: Coagulopathy in traumatic brain injury. Neurocrit Care 2004; 1: 479-488

40. Nekludov M, Bellander BM, Blomback M, et al.: Platelet dysfunction in patients with severe traumatic brain injury. J Neurotrauma 2007;24: 1699-1706

41. Windelov NA, Welling KL, Ostrowski SR, et al.: The prognostic value of thrombelastogra-phy in identifying neurosurgical patients with worse prognosis. Blood Coagul Fibrinoly-sis 2011;22: 416-419

42. Park MS, Salinas J, Wade CE, et al.: Combining early coagulation and inflammatory status improves prediction of mortality in burned and nonburned trauma patients. J Trauma 2008;64: S188-S194

43. Gonzalez E, Kashuk JL, Moore EE, et al.: Differentiation of enzymatic from platelet hypercoagulability using the novel thrombelastography parameter delta (delta). J Surg Res 2010;163: 96-101

44. Schreiber MA, Differding J, Thorborg P, et al.: Hypercoagulability is most prevalent early after injury and in female patients. J Trauma 2005;58: 475-480


46. Kaufmann CR, Dwyer KM, Crews JD, et al.: Usefulness of thrombelastography in assess-ment of trauma patient coagulation. J Trauma 1997;42: 716-720

Muller.indd 39 25-7-2014 14:40:20

Muller.indd 40 25-7-2014 14:40:20

Chapter 3

Utility of thromboelastography and/or thromboelastometry

in adults with sepsis: a systematic review

Marcella C. Müller, Joost C.M. Meijers, Margreeth B. Vroom, Nicole P. Juffermans

Critical Care 2014;18(1):R30

Muller.indd 41 25-7-2014 14:40:20

Chapter 3

42

Abstract

Introduction: Coagulation abnormalities are frequent in sepsis. Conventional coagu-lation assays, however, have several limitations. A surge of interest exists in the use of point-of-care tests to diagnose hypo- and hypercoagulability in sepsis.We performed a systematic review of available literature to establish the value of rotational thromboelastography (TEG) and thromboelastometry (ROTEM) com-pared with standard coagulation tests to detect hyper- or hypocoagulability in sep-sis patients. Furthermore, we assessed the value of TEG/ROTEM to identify sepsis patients likely to benefit from therapies that interfere with the coagulation system.Methods: MEDLINE, EMBASE, and the Cochrane Library were searched from 1 Jan-uary 1980 to 31 December 2012. The search was limited to adults, and language was limited to English. Reference lists of retrieved articles were hand-searched for additional studies. Ongoing trials were searched on www.controlled-trials.com and www.clinicaltrials.gov. Studies addressing TEG/ROTEM measurements in adult patients with sepsis admitted to the ICU were considered eligible.Results: Of 680 screened articles, 18 studies were included, of which two were rand-omized controlled trials, and 16 were observational cohort studies. In patients with sepsis, results show both hyper- and hypocoagulability, as well as TEG/ROTEM values which fell within reference values. Both hyper- and hypocoagulability were to some extent associated with diffuse intravascular coagulation. Compared to conventional coagulation tests, TEG/ROTEM can detect impaired fibrinolysis, which can possibly help to discriminate between sepsis and systemic inflammatory response syndrome (SIRS). A hypocoagulable profile is associated with increased mortality. The value of TEG/ROTEM to identify patients with sepsis who could possibly benefit from thera-pies interfering with the coagulation system could not be assessed, because studies addressing this topic were limited.Conclusion: TEG/ROTEM could be a promising tool in diagnosing alterations in coag-ulation in sepsis. Further research on the value of TEG/ROTEM in these patients is warranted. Given that coagulopathy is a dynamic process, sequential measurements are needed to understand the coagulation patterns in sepsis as can be detected by TEG/ROTEM.

Muller.indd 42 25-7-2014 14:40:20

Thromboelastography and/or thromboelastometry in sepsis

43

Introduction

Coagulopathy is highly prevalent in sepsis patients and is associated with increased mortality [1]. Coagulopathy results from an imbalance between activation of coag-ulation and impaired inhibition of coagulation and fibrinolysis. The disturbance between components of the coagulation system leads to a variable clinical picture, ranging from an increased bleeding tendency due to consumption of coagulation factors and platelets, to hypercoagulopathy with disseminated intravascular coagu-lation (DIC) and (micro-) vascular thrombosis.Assessment of coagulation status in these patients is complex. Global coagulation tests activating partial thromboplastin time (aPTT) and prothrombin time (PT) are used clinically. However, their ability to reflect in vivo hypocoagulability accurately is questioned [2]. Also, aPTT and PT only reflect a part of the coagulation system and do not provide information on the full balance between coagulation and anticoagu-lation. Activation of coagulation can be assessed by thrombin generation, but this test is not widely available. Impaired function of the anticoagulant system can be diagnosed by measuring plasma levels of naturally occurring anticoagulant factors antithrombin (AT), protein C, protein S and tissue factor pathway inhibitor (TFPI). However, these are not readily available for clinical use. The same applies to markers of the activity of the fibrinolytic system [2]. Although activation of the fibrinolytic system can be detected by increased levels of D-dimers and other fibrin degradation products, specificity is limited [2].Rotational thromboelastography (TEG) and thromboelastometry (ROTEM) are point-of-care tests, which evaluate whole clot formation and dissolution. The thromboe-lastogram arises through movement of the cup (TEG) or the pin (ROTEM). As fibrin forms between the cup and the pin, this movement is influenced and converted to a trace reflecting different phases of the clotting process. Major parameters are reaction time (R) or clotting time (CT), which is the period from the initiation of the test until the beginning of clot formation (figure 1). K time or clot formation time (CFT) is the period from the start of the clot formation until the curve reaches an amplitude of 20 mm. Kinetics of fibrin formation and cross-linking are expressed by the α-angle, which is the angle between the baseline and the tangent to the TEG/ROTEM curve amplitude. Clot strength is represented by the maximal amplitude of

Muller.indd 43 25-7-2014 14:40:20

Chapter 3

44

Figure 1: ROTEM trace with major parameters

Reference: www.rotem.de

Table 1: Parameters displayed on TEG and ROTEM

TEG ROTEMTime to initial fibrin formation(to 2-mm amplitude)

R CT

Clot strengthening, rapidity of fibrin build up Ka

CFTa

Clot strength, represents maximum dynamics of fibrin and platelet bonding

MA MCF

Clot breakdown, fibrinolysis at fixed time CL30, CL60 LI30, LI45, LI60

Reference: www.rotem.de and www.haemoscope.com

the trace. The degree of fibrinolysis is reflected by the difference between the maxi-mal amplitude and the amplitude measured after 30 and/or 60 minutes (figure 1). To describe these viscoelastic changes, both systems have their own terminology (table 1).The technique was developed in the 1940s, but clinical application has been limited. However, technical developments have led to standardization and improved repro-ducibility of the method [3,4]. TEG/ROTEM may facilitate diagnosis of clotting abnor-malities in sepsis, including hypercoagulable states such as DIC. Other potential advantages could be a more tailor-made administration of therapies that interfere

Muller.indd 44 25-7-2014 14:40:20


45

with the coagulation system [5-8]. These tests may also improve prognostication of sepsis [9,10].The main research questions for this systemic review were as follows. Can TEG/ROTEM detect sepsis-induced coagulopathy? Is TEG/ROTEM of additional value compared with standard coagulation tests to detect hyper- or hypocoagulability in sepsis patients? Can TEG/ROTEM help to identify sepsis patients likely to benefit from therapies that interfere with the coagulation system (for example, activated protein C, antithrombin, heparin)? We defined our population as critically ill adults with sepsis and TEG/ROTEM as the intervention. Standard coagulation tests includ-ing aPTT, PT, INR and ISTH DIC score, functioned as comparisons. Outcomes of inter-est were the detection of a hyper- or hypocoagulable state in these patients and the identification of patients likely to benefit from therapies affecting the coagulation system.

Materials and methods

Data sourcesAn electronic search was conducted in MEDLINE, EMBASE, and the Cochrane Library. In addition, we searched for ongoing trials on www.controlled-trials.com and www.clinicaltrials.gov. We hand-searched the reference lists of retrieved papers, reviews and editorials for additional studies. Language was limited to papers written in Eng-lish and published from 1 January 1980 to 31 December 2012. We did not register our protocol.

Study selectionTwo authors (MCM and NPJ) performed the literature search and selected the relevant articles for inclusion. Differences were resolved in consensus meetings. Predefined eligibility criteria were used. Studies were included if TEG/ROTEM meas-urements were performed in adult patients with sepsis admitted to the ICU. Ran-domized controlled trials, prospective and retrospective cohorts, and case series were all eligible for inclusion. Reviews, correspondences, case reports, expert opin-ions and editorials were excluded. We also excluded all studies conducted outside the ICU or that involved subjects younger than 18 years.

Muller.indd 45 25-7-2014 14:40:21

Chapter 3

46

Data collection processTwo of the authors (MCM and NPJ) independently extracted the data using a pre-defined extraction sheet. Discrepancies were resolved in a consensus meeting. If agreement could not be reached a third author was consulted (MBV) to resolve disagreement. The extracted data were general methodological characteristics, set-ting, characteristics of the study population, used test (ROTEM or TEG), timing of thromboelastography, possible comparison of thromboelastography results with a reference test, administration of therapies interfering with the coagulation system, and main outcomes. Furthermore, possible limitations of each study were listed. No assumptions or simplifications were made.

Assessment of methodological qualityWe used the QUADAS-2 checklist to assess quality of diagnostic studies [11,12]. Studies were assessed for the risk of bias in patient selection, conduction and interpretation of TEG/ROTEM, use and interpretation of a reference standard, and patient flow. For all research questions, methodological aspects, including the use of a comparison of TEG/ROTEM measurement, the individual studies were assessed. Furthermore, studies were judged with respect to patient population and selection (use of definition of sepsis). Details on how TEG/ROTEM and reference tests were conducted and interpreted (for example, timing of the tests, blinded interpreta-tion) were assessed. Subsequently, quality of evidence was judged and described in accordance with the GRADE approach (high, moderate, low and very low). Rating of quality of evidence was based on trial design (for example, randomized clinical trial or not), risk of bias and imprecision (for example, patient selection and patient flow, method of conduct and interpretation of TEG/ROTEM and reference test results). We verified whether results of the retrieved trials and abstracts have been pub-lished.

DefinitionsHypocoagulability can be defined as prolonged CT/R and CFT/K times and/or decreased MCF/MA and alpha angle [4]. Conversely, a hypercoagulable state can be detected by shortened reaction times (CT/R and/or CFT/K) and enhanced clot formation, expressed as increased alpha and/or high maximal amplitude (MCF/MA).

Muller.indd 46 25-7-2014 14:40:21


47

However, no universal definitions of hypo- and hypercoagulability assessed by TEG/ROTEM exist.

Results

Study selectionOf 680 screened articles, we included 18 studies (table 2). An overview of the search is presented in figure 2. Twenty-six studies were excluded, because of inappropriate patient population (N=21 [13-33]), no report of TEG/ROTEM data (N=6 [31,32,34-37]) and one conference poster [38].

Study characteristicsWe included two randomized controlled trials [39,40] and 16 observational studies [9,10,41-54]. Of the observational studies, 14 were prospective and one abstract [45] and one article [42] did not state whether the study was prospective or ret-rospective. Of the included studies, 11 [9,10,41,43,44,48-51,53,54] used the sep-sis and SIRS criteria defined by the Society of Critical Care Medicine Consensus Conference [55]. Seven studies used TEG [10,39,40,42,49,51,53] and 11 used ROTEM [9,41,43-48,50,52,54].

Risk of biasThe risk of bias within studies is summarized in table 3. Quality assessment revealed that risk of bias of patient selection was low. However, included studies were het-erogeneous regarding conduction and interpretation of TEG/ROTEM, and no studies reported whether results of TEG/ROTEM were interpreted with or without knowl-edge of the used reference test, which leaves the possibility for interpretation bias. However, applicability concerns of the chosen reference tests are low.In addition to individual sources of bias related to design and methodology, we iden-tified the lack of information about the conduct and interpretation of TEG/ROTEM test and results as the most important source of bias across all studies (table 3). We verified whether abstracts and trial results had been published. None had been pub-lished at December 31, 2012. Construction of a funnel plot was not feasible because of characteristics of retrieved studies.

Muller.indd 47 25-7-2014 14:40:21

Chapter 3

48

Table 2: S

tudi

es a

sses

sing

TEG/

ROTE

M in

seps

is

Author,

year

Type

of study

Popu

latio

n,

(N)

ROTE

M

or TEG

Timing of

mea

suremen

tCo

mpa

rison

Main RO

TEM/TEG

find

ings

Gon

ano,

20

06 [4

0]Su

bana

lysis

of

rand

omize

d co

ntro

lled

tria

l

Seve

re se

psis

(N=3

3)TE

GAt

dia

gnos

is an

d da

ily

ther

eafte

r.

PT, a

PTT,

AT

All p

atien

ts w

ere

hype

rcoa

gula

ble

(sho

rten

ed R

an

d CT

, inc

reas

ed a

lpha

and

MA)

. Anti

thro

mbi

n tr

eatm

ent d

id n

ot a

ffect

TEG

val

ues.

Ra

iner

i, 20

09

(abs

trac

t) [3

9]

Rand

omize

d co

ntro

lled

tria

l Se

vere

seps

is an

d se

ptic

shoc

k (N

=16)

TEG

Daily

for 2

w

eeks

and

day

17

, 20,

23,

28

PAI-1

In p

atien

ts w

ithou

t tigh

t gly

cem

ic c

ontr

ol (T

GC)

, fib

rinol

ysis

was

dec

reas

ed (i

ncre

ased

lysis

inde

x an

d in

crea

sed

PAI-1

), co

mpa

red

to se

psis

patie

nts

not t

reat

ed w

ith T

GC.

Co

llins

, 20

06 [4

3]Pr

ospe

ctive

ob

serv

ation

alSe

psis

(N=3

8),

heal

thy

con-

trol

s (N

=32)

ROTE

MN

ot s

tate

dPT

, aPT

T fib

rinog

en,

fact

or le

vels

In se

psis

ther

e w

as d

elay

ed a

ctiva

tion

of h

emos

ta-

sis, o

nce

activ

ated

clo

t for

mati

on w

as e

xagg

erat

ed

(incr

ease

d M

CF, a

lpha

ang

le, a

rea

unde

r clo

t firm

-ne

ss c

urve

).Ch

iari,

20

09

(abs

trac

t) [4

7]

Pros

pecti

ve

obse

rvati

onal

Seve

re se

psis

(N=1

5)RO

TEM

Befo

re a

nd 1

st

day

of tr

eat-

men

t with

ac

tivat

ed p

ro-

tein

C

aPTT

, PT

Onl

y CT

sign

ifica

ntly

incr

ease

d w

ith a

ctiva

ted

pro-

tein

C tr

eatm

ent.

Daud

el,

2009

[44]

Pros

pecti

ve

coho

rtSe

psis

(N=3

0)RO

TEM

0-48

hou

rs

after

dia

gnos

is an

d at

dis

-ch

arge

INR,

aPT

T,

fibrin

ogen

, in

divi

dual

fa

ctor

s

All p

aram

eter

s with

in re

fere

nce

valu

es. P

atien

ts

with

SO

FA >

10 h

ad in

crea

sed

coag

ulati

on (r

educ

ed

MCF

and

alp

ha a

nd in

crea

sed

CFT)

.

Schm

it-tin

ger,

2009

(a

bstr

act)

[46]

Pros

pecti

veob

serv

ation

alSe

vere

seps

is (N

=49)

, pos

t-op

erati

ve S

IRS

(N=2

7)

ROTE

MDa

y 1,

4, 7

aft

er a

dmis

-si

on

Non

eAl

l par

amet

ers w

ithin

refe

renc

e va

lues

. Mor

talit

y 58

.3%

in p

atien

ts w

ith si

gns o

f hyp

ocoa

gula

tion

vs.

9.1%

in th

ose

with

sign

s of h

yper

coag

ulab

ility

.

Sivu

la,

2009

[41]

Pros

pecti

ve

obse

rvati

onal

Seve

re se

p-sis

(N=2

8),

heal

thy

con-

trol

s (N

=8)

ROTE

MDa

y 1

aPTT

, AT,

D

-dim

er,

fibrin

ogen

Onl

y se

psis

patie

nts w

ith D

IC w

ere

hypo

coag

ulab

le

com

pare

d to

hea

lthy

cont

rols

. CFT

, alp

ha a

nd M

CF

disc

rimin

ated

wel

l bet

wee

n DI

C an

d no

n- D

IC.

Decr

ease

d fib

rinol

ysis

in a

ll se

psis

patie

nts v

ersu

s co

ntro

ls.

Muller.indd 48 25-7-2014 14:40:21


49

Author,

year

Type

of study

Popu

latio

n,

(N)

ROTE

M

or TEG

Timing of

mea

suremen

tCo

mpa

rison

Main RO

TEM/TEG

find

ings

Adam

zik,

2010

[50]

Pros

pecti

ve

obse

rvati

onal

Seps

is (N

=56)

, po

stop

era-

tive

cont

rols

(N=5

2)

ROTE

MW

ithin

24

hour

s of s

epsis

di

agno

sis

Proc

alci

-to

nin,

IL-6

, CR

P

Incr

ease

d ly

sis in

dex

in se

psis

com

pare

d to

po

stop

erati

ve c

ontr

ols (

97±0

.3%

vs.

92±

0.5%

, p<

0.00

1). C

FT, a

lpha

and

MCF

did

not

diff

er

betw

een

grou

ps. L

ysis

inde

x ha

d be

st a

ccur

acy

for

diag

nosis

seps

is.

Altm

ann,

20

10 [4

8]Pr

ospe

ctive

obse

rvati

onal

Septi

c sh

ock

(N=1

6), s

ever

e se

psis

(N=7

), SI

RS (N

=10)

ROTE

M0,

12,

24,

48

h aft

er in

clus

ion

Non

eAl

l par

amet

ers w

ithin

refe

renc

e va

lues

.

Duril

a,

2010

[49]

Pros

pecti

ve

obse

rvati

onal

Seve

re se

psis

(N=4

4)TE

GN

ot s

tate

dIN

R, a

PTT,

fib

rinog

en,

AT

All p

aram

eter

s with

in re

fere

nce

valu

es.

Adam

zik,

2011

[9]

Pros

pecti

ve

obse

rvati

onal

Se

psis

(N=9

8)RO

TEM

With

in 2

4 ho

urs o

f dia

g-no

sis

INR

39%

of s

epsis

pati

ents

had

nor

mal

CFT

, MCF

and

al

pha

angl

e, v

alue

s in

61%

with

pat

holo

gica

l var

i-ab

le sh

owed

bro

ad d

istr

ibuti

on. H

ypoc

oagu

labl

e pr

ofile

ass

ocia

ted

with

incr

ease

d m

orta

lity

(OR

4.1;

95%

CI 1

.4-1

1.9)

.Co

rteg

iani

, 20

11

(abs

trac

t) [5

1]

Pros

pecti

ve

obse

rvati

onal

Seve

re se

psis

(N=3

1), p

ost

oper

ative

(N

=31)

TEG

With

in 1

2 ho

urs o

f dia

g-no

sis

Non

eSe

psis

patie

nts h

ad lo

wer

alp

ha a

ngle

, oth

er T

EG

para

met

ers d

id n

ot d

iffer

.

Bren

ner,

2012

[54]

Pros

pecti

ve

obse

rvati

onal

Septi

c sh

ock

(N=3

0),

maj

or su

r-ge

ry (N

=30)

, he

alth

y vo

lun-

teer

s (N

=30)

ROTE

MSe

psis:

at d

iag-

nosis

, 24h

, 4,

7, 1

4, 2

8 da

ys

Prot

hrom

-bi

n in

dex,

fa

ctor

le

vels

, IL-

6,

TNF-

alph

a.

In se

psis

patie

nts m

ajor

ity o

f RO

TEM

ana

lysis

w

ithin

refe

renc

e va

lues

, how

ever

seps

is pa

tient

s w

ith D

IC sh

owed

mor

e hy

poco

agul

able

trac

es

com

pare

d to

thos

e w

ithou

t DIC

wer

e m

ore

hype

r-co

agul

able

. Com

pare

d to

surg

ical

and

hea

lthy

con-

trol

s fibr

inol

ysis

was

impa

ired

in se

psis

patie

nts.

Muller.indd 49 25-7-2014 14:40:21

Chapter 3

50

Author,

year

Type

of study

Popu

latio

n,

(N)

ROTE

M

or TEG

Timing of

mea

suremen

tCo

mpa

rison

Main RO

TEM/TEG

find

ings

Duril

a,

2012

[53]

Pros

pecti

ve

obse

rvati

onal

Post

surg

ical

oe

soph

agec

-to

my

(N=3

8),

of th

ese

9 de

velo

ped

seps

is.

TEG

Mor

ning

of

surg

ery

and

daily

day

1-6

po

st o

pera

tive

aPTT

, IN

R,

CRP,

lact

ate,

IL-

6, p

roca

l-ci

toni

n, A

T,

D-d

imer

On

6th

post

oper

ative

day

seps

is pa

tient

s had

hi

gher

lysis

inde

x co

mpa

red

to S

IRS

patie

nts.

Ove

r-al

l TEG

not

hel

pful

in d

iscr

imin

ating

seps

is fr

om

SIRS

.

Mas

sion

, 20

12 [5

2]Pr

ospe

ctive

co

hort

Septi

c sh

ock

(N=3

9)RO

TEM

Adm

issi

on to

da

y 7

aPTT

, PT,

Th

rom

bin

gene

ra-

tion,

fact

or

leve

ls, A

T,

prot

ein

C

Fibr

inol

ysis

was

dec

reas

ed (i

ncre

ased

lysis

in

dexe

s), a

ssoc

iate

d w

ith h

ypoc

oagu

latio

n in

con

-ve

ntion

al c

oagu

latio

n te

sts (

decr

ease

d pr

otei

n C

and

AT).

Oth

er p

aram

eter

s with

in re

fere

nce

valu

es

(CT,

MCF

and

alp

ha).

Non

-sur

vivo

rs w

ere

mor

e hy

poco

agul

able

, but

RO

TEM

val

ues w

ere

not i

nde-

pend

ently

ass

ocia

ted

with

mor

talit

y.O

stro

wsk

i, 20

12 [1

0]Pr

ospe

ctive

ob

serv

ation

alSe

vere

seps

is (N

=13)

and

se

ptic

shoc

k (N

=37)

TEG

Day

1-4

ISTH

DIC

sc

ore,

IN

R, a

PTT,

D

-dim

er,

fibrin

ogen

, CR

P

Acco

rdin

g to

clo

th s

tren

gth

(MA)

48%

of s

epsis

pa

tient

s was

nor

moc

oagu

labl

e, 2

2% h

ypoc

oagu

-la

ble

and

30%

hyp

erco

agul

able

. 50%

of p

atien

ts

with

hyp

ocoa

gula

ble

profi

le h

ad o

vert

DIC

, ver

sus

none

of t

hose

with

a h

yper

coag

ulab

le p

rofil

e.

Hypo

coag

ulab

le p

rofil

e pr

edic

ts 2

8-da

y m

orta

lity

if co

rrec

ted

for S

OFA

, but

not

if c

orre

cted

for S

APS

II sc

ore.

Viljo

en,

1995

[42]

Not

sta

ted

Seps

is (N

=15)

, tr

aum

a (N

=14)

, sur

-ge

ry (N

=21)

, he

alth

y co

n-tr

ol (N

=23)

TEG

Daily

Pl

asm

a el

asta

se-

alph

asub

1

PI

Seps

is pa

tient

s wer

e hy

poco

agul

able

com

pare

d to

su

rger

y an

d co

ntro

ls. S

epsis

pati

ents

had

hig

her

elas

tase

-alp

hasu

b1 p

rote

inas

e in

hibi

tor l

evel

s co

mpa

red

to c

ontr

ols,

with

out a

cor

rela

tion

with

TE

G pa

ram

eter

s.

Um

gelte

r, 20

09

(abs

trac

t) [4

5]

Not

sta

ted

Seps

is (N

=21)

, no

seps

is (N

=23)

ROTE

MN

ot s

tate

dTh

rom

-bi

n tim

e,

D-d

imer

, AT

ROTE

M d

id n

ot d

iscr

imin

ate

betw

een

septi

c an

d no

n-se

ptic

cirr

hotic

pati

ents

.

Muller.indd 50 25-7-2014 14:40:21


51

Figure 2: PRISMA flow diagram

Addi�onal recordsiden�fied through other sources

N=13

Records iden�fied through databasescreenings N=1080

Medline N=454Embase N=485

Cochrane library N=141

Records a�er duplicatesremoved

N=680

Records �tles screenedN=680

Excluded a�er �tle reviewN=461

Abstracts screened for eligibilityN=219

Full-text ar�cles assessed foreligibility

N=44

Excluded a�er review ofabstracts

N=175

Excluded a�er review of full-text ar�cle N=26

Reasons for exclusion:– Pa�ent popula�on N=21

– No ROTEM/TEG performed N=6– Conference report of later

published article N=1

Studies included in qualita�vesynthesis

N=18

Iden

�fica

�on

Scre

enin

gEl

igib

ility

Incl

uded

Muller.indd 51 25-7-2014 14:40:21

Chapter 3

52

Table 3: Summary of risk of bias and applicability concerns for included studies

STUDY Risk of bias Applicability concernsPatient selection

Conduction and inter-pretation of TEG/ROTEM

Use and interpre-tation of reference standard

Patient flow

Patient selec-tion

Reference standard

Gonano et al. [40]

- ? - ? - -

Raineri et al. [39]

? ? - - - -

Collins et al. [43]

? ? - ? - -

Chiari et al. [47]

- ? - ? - -

Daudel et al. [44]

- - - ? - -

Schmittinger et al. [46]

- - NA - - NA

Sivula et al. [41]

- ? ? ? - -

Adamzik et al. [50]

- ? ? - - -

Altmann et al. [48]

- ? NA NA - NA

Durila et al. [49]

- + - - - -

Adamzik et al. [9]

- - - + - -

Cortegiani et al. [51]

? ? NA - - NA

Brenner et al. [54]

- - - - - -

Durila et al. [53]

- ? ? + - -

Massion et al. [52]

- ? - - - -

Ostrowski et al. [10]

- - - + - -

Viljoen et al. [42]

+ ? ? ? + +

Umgelter et al. [45]

+ + NA ? + NA

-, low risk; +, high risk; ?, unclear risk; NA, not applicable

Muller.indd 52 25-7-2014 14:40:21


53

Synthesis of resultsStudies varied widely in the way they were conducted. Main differences were tim-ing of TEG/ROTEM measurements, number of measurements, use of preset refer-ence values or control group to assess derangements in TEG/ROTEM measurements in sepsis, and variable use of different comparison tests. Because of this clinically relevant heterogeneity, we carried out a narrative synthesis of the results of the included studies.

Ability of TEG/ROTEM to detect sepsis-induced coagulopathyIn five studies, all TEG/ROTEM measurements in sepsis were within reference val-ues [44,46,48,49,54]. These studies included a total of 176 sepsis and severe sep-sis patients. In seven studies, TEG/ROTEM revealed pathological changes ranging from distinct hypercoagulability [40], to predominantly hypocoagulable profiles [42,51]. Four prospective observational studies, together including 214 patients, reported heterogeneous results with patients showing hyper- and hypocoagulabil-ity [9,10,41,43]. Impaired fibrinolysis in sepsis patients was demonstrated in five dif-ferent observational studies with a total of 162 patients [41,50,52-54]. Two of these studies reported increased lysis indices as the only abnormal ROTEM parameter in 30 and 39 patients with septic shock [52,54]. In a small cohort of 16 patients with severe sepsis; patients were randomized to tight glycemic control or conventional glucose levels; strict regulation of glucose levels resulted in enhancement of fibrinolysis, as measured by lysis index with ROTEM [39]. Overall, if sepsis-induced coagulopathy was detected, the proportion of sepsis patients with sepsis-induced coagulopathy that was detected by TEG/ROTEM ranged from 43% to 100% [9,10,40,41].Altogether, in the majority of studies, TEG/ROTEM was able to detect sepsis-induced coagulopathy. However, changes in parameters were heterogeneous and study designs varied widely, with a lack of clarity on interpretation of test results. Based on the above, the quality of evidence supporting the use of TEG/ROTEM to detect sepsis-induced coagulopathy is considered low.

Muller.indd 53 25-7-2014 14:40:21

Chapter 3

54

Additional value of TEG/ROTEM in sepsis compared with standard coagulation testsStudies designed to compare conventional coagulation tests directly with TEG/ROTEM for the detection of sepsis-induced coagulopathy were not retrieved.However, two studies assessed the value of thromboelastography in the detection of DIC. A prospective pilot study in 28 sepsis patients showed that CFT, MCF and alpha angle discriminated moderately between overt DIC and no DIC. ROC values were 0.815 (CI 0.624 to 0.935) for CFT, 0.891 (CI 0.715 to 0.975) for MCF and 0.828 (CI 0.639 to 0.943) for alpha angle. Combination of CFT, MCF and alpha resulted in a sensitivity of 100% and a specificity of 75%, with a positive likelihood ratio of 4.0 and negative likelihood ratio of 0.0 for the diagnosis of DIC [41].A recent study showed that ROTEM values were within reference values. However, patients with overt DIC had prolonged CFT and reduced MCF compared with those without DIC. ROC curves for MCF in EXTEM (0.806; sensitivity 79%, specificity 75%) and MCF in INTEM (0.853, sensitivity 86% and specificity 81%) achieved fairly good results [54].Interestingly, TEG/ROTEM failed to detect coagulopathy in two observational stud-ies, whereas conventional coagulation assays were outside normal ranges [44,49]. In 44 sepsis patients, mean INR, D-dimer and fibrinogen levels were increased, whereas mean TEG values were not [49]. A study of 30 sepsis patients revealed similar results, with decreased levels of individual factor levels and increased aPTT tests, whereas ROTEM variables remained within reference values [44].Some studies addressed the question whether TEG/ROTEM was superior in discrim-inating sepsis from non-sepsis patients compared with conventional biomarkers. Indeed, two observational studies demonstrated that the lysis index derived from thromboelastometry could be helpful to discriminate between sepsis and postop-erative inflammatory response [50,54]. Decreased fibrinolytic activity, as reflected by the lysis index, was found to discriminate sepsis from postoperative SIRS patients (ROC-AUC 0.811, sensitivity 93%, specificity 50%), which was comparable to CRP and procalcitonin [54]. In a larger cohort of 56 sepsis patients and 52 postoperative controls, lysis index had even better diagnostic value than procalcitonin (ROC-AUC 0.901, sensitivity 84% and specificity 94%) [50]. In contrast, in cirrhosis patients and

Muller.indd 54 25-7-2014 14:40:21


55

post-esophagectomy patients, thromboelastometry variables failed to discriminate between sepsis and non-sepsis patients [45,53].

Ability of TEG/ROTEM to identify patients likely to benefit from anticoagulant treatment in sepsisWe hypothesized that TEG/ROTEM may help to identify patients likely to respond to therapies that target coagulopathy. However, we did not find any study addressing this question. The only data on TEG/ROTEM and therapy interfering with coagu-lation consist of a few small patient series evaluating TEG/ROTEM measurements during anticoagulant treatment. In an observational study of 15 patients treated with rhAPC, ROTEM measurements did not change during treatment [47]. Gonano [40] showed distinct hypercoagulability in 33 patients with severe sepsis, of whom 17 were treated with antithrombin. In these patients, hypercoagulability was not reversed by the treatment.

Use of TEG/ROTEM in prognostication of outcomeAlthough initially not looked for, we extracted a limited number of studies that addressed the value of TEG/ROTEM in predicting outcome in sepsis and decided post hoc to add these data to the review. In a cohort of 98 sepsis patients using mul-tivariate analysis, a hypocoagulable profile on admission was shown to be an inde-pendent risk factor for 30-day mortality (OR 4.1; 95%CI 1.4 to 11.9) [9]. However, not all studies have unequivocally showed the prognostic value of hypocoagulability with mortality after correcting for disease severity [10,52].In 50 patients with severe sepsis, hypocoagulable TEG MA at admission was an independent predictor for 28-day mortality in a multivariate model including SOFA score (hazard ratio 4.29 (1.35 to 13.65), p=0.014), but not in a model using SAPS II score (hazard ratio 2.32 (0.66 to 8.15), p=0.188) [10]. Of note, in a multivariate model a hypocoagulable profile due to a persistent deficit in thrombin generation was a strong predictor of hospital mortality (p=0.024) as was aPTT (p=0.007) [52]. The presence of hypercoagulability did not predict outcome.Quality of evidence of studies addressing the value of TEG/ROTEM to predict mor-tality is considered of moderate quality. Of note, in these studies most patients had thromboelastography values outside reference ranges.

Muller.indd 55 25-7-2014 14:40:21

Chapter 3

56

Discussion

This systematic review on studies performing TEG/ROTEM measurements in sepsis patients shows that studies were heterogeneous in design, use of control groups, timing of TEG/ROTEM measurements, and chosen endpoints. Internal validity of most studies is limited. Although most studies included sepsis patients according to the ACCM/SCCP definition, external validity is limited because of relatively small patient groups in most studies. Furthermore, standardization of used tests is lim-ited, and most studies had methodological flaws, which may have resulted in bias. Thereby, the overall quality of evidence on the value of TEG/ROTEM in adults with sepsis is considered low.Results of TEG/ROTEM measurements in sepsis vary widely across studies and show both hypo- and hypercoagulability [10,40-43,46,51]. This is consistent with the pathophysiology of ‘consumption coagulopathy’ during DIC, in which microvascular thrombi are formed at the expense of a bleeding tendency because of low levels of platelets and coagulation factors [56]. Of note, heterogeneity of results can also be caused by differences in disease severity, as changes were more obvious in severe sepsis patients [40,41,46,51] than in sepsis patients. In addition, variation in the way studies were conducted has probably contributed to differences in outcome.Interestingly, the degree of hypocoagulation was found to be associated with sever-ity of organ failure [44]. In a study comparing different patient populations, hypoco-agulation measured with TEG was most apparent in sepsis patients and associated with a pro-inflammatory response and organ failure [42]. Timing of measurements may be relevant to these observations, as hypocoagulation was found to be more obvious in the acute phase of sepsis and returned to normal values toward discharge of ICU [44,51].Included studies varied widely with regard to the detection of hypercoagulability, ranging from 30% [10] to 100% [40], which may have resulted from variation in tim-ing as well as in the definition of hypercoagulability. Of note, TEG/ROTEM clearly demonstrated hypercoagulability in models of endotoxemia [57], with a strong cor-relation with plasma levels of prothrombin fragments F1+2 [58,59].In addition to hypo- and hypercoagulability, TEG/ROTEM can detect impairment in fibrinolysis, expressed as increased lysis indices. Hypofibrinolysis has been demon-

Muller.indd 56 25-7-2014 14:40:21


57

strated in several studies in sepsis patients [41,50,52-54], but the clinical relevance of this finding must be determined. Of note, increased lysis indices were shown to be helpful in discriminating sepsis and SIRS patients [50,53,54].TEG/ROTEM has shown to be promising in diagnosing DIC, and in particular, the combination of various parameters (reaction times, maximum amplitude and alpha angle) improves diagnostic value [41,54]. A score to detect DIC with the use of throm-boelastometry has been developed, including prolonged reaction and K times and decreased alpha angle and maximum amplitude. This score was validated in patients with an underlying disease known to be associated with DIC and with an ISTH DIC [60] score of more than 5. However, to date, this score has not been validated in critically ill patients with sepsis, and included studies in this review consisted of relatively small patient groups. Therefore to date, quality of evidence supporting the use of TEG/ROTEM to diagnose DIC is low, and further research is necessary.Several factors in the way TEG/ROTEM measurements were conducted may have affected the results of included studies. First, coagulopathy in sepsis is a dynamic process, evolving from subtle activation of coagulation to overt DIC. Therefore tim-ing may greatly influence TEG/ROTEM results. Performing sequential measurements will probably provide better insight in the development of coagulation derange-ments. Timing and number of measurements in included studies varied widely.Second, no uniform definitions exist of hypo- and hypercoagulability. Reference values for patients with sepsis are not widely assessed and only one study deter-mined cut-off values for a cohort of sepsis patients [9]. Some studies classified patients as hypo- or hypercoagulable when measurements were outside preset reference ranges [10,40,44,46]; others compared patients with healthy individu-als [41,43,50,52,54]; and some compared mean or median values among or within different patient groups [9,39,42,45-51,53,54]. To compare patient categories and possibly investigate therapeutic interventions in the coagulation system, validated universal reference values and definitions are essential. For ROTEM, a multicenter investigation has been undertaken to assess reference values [4]. A study to verify reference intervals of TEG reagents has been recently completed (NCT01357928), and we hope that results will contribute to further standardization of TEG.Third, the included studies differed in the types of reagents used, which may have considerable effects on the results of the studies. Most studies using ROTEM used

Muller.indd 57 25-7-2014 14:40:21

Chapter 3

58

tissue factor, after recalcification of the citrated sample, to enhance coagulation [41,43,44,48,52,54], although some also performed a non activated test (NATEM). Of the studies using TEG only one study reported the use of kaolin activation [49]; the others only recalcified samples before testing. Of note, correlation between non-kaolin-activated and kaolin-activated thromboelastography has shown to be poor [61]. Furthermore, in some studies, a potential heparin effect was blocked by the addition of heparinase [9,40,44,50,52,54], whereas others lacked information on the use of heparinase [10,42].In the current review, we included studies using ROTEM and studies using TEG. None of the included studies studied both devices. Some studies that compared both devices in other patient populations showed differences in test results between ROTEM and TEG [62,63], although not all [64]. Therefore, comparisons of results of studies using different devices should be made cautiously.A hypocoagulable profile detected by TEG/ROTEM seems to be associated with increased mortality among sepsis patients [9,10,46]. One could argue that a hypoco-agulable profile merely reflects severity of disease. However, in the study of Adamzik [9], hypocoagulable TEG/ROTEM remained an independent predictor of mortality after correction for severity of disease. These findings are in line with results in a larger cohort of general intensive care patients, in which a hypocoagulable profile at admission was associated with an increased mortality [26]. This relation questions the role of coagulation in inflammatory processes. We speculate that enhanced coagulation during infection is functional, thereby preventing dissemination of bac-teria. Thereby, hypocoagulability may facilitate enhanced spread of infection and subsequently mortality [65]. The finding that hypocoagulability is associated with organ failure and is an independent risk factor for mortality underlines the need for further research. Currently, two observational prospective trials in sepsis patients are conducted (NCT00994877 and NCT00299949) on the value of TEG/ROTEM to diagnose DIC and to predict organ failure in sepsis. Results of these studies may help to determine whether TEG/ROTEM can be used to select specific patient popula-tions who are likely to benefit from therapies aimed at intervention in the coagula-tion cascade during sepsis.Our review has limitations, which include the lack of a uniform definition of hypo- and hypercoagulability assessed by ROTEM or TEG, different reference values, dif-

Muller.indd 58 25-7-2014 14:40:21


59

ferences in control groups, and the heterogeneous study quality. Another limitation is related to our search, in which we may have missed studies published in languages other then English, as well as unpublished data. In addition, we might have missed studies due to applied date restriction and limitation of our search to three data-bases.

Conclusion

A considerable proportion of sepsis patients have an altered coagulation status. An abnormal TEG/ROTEM, in particular hypocoagulability, is prognostic for mortality in the critically ill. Also, hypocoagulability as detected by TEG/ROTEM may aid in diagnosing DIC and hypofibrinolysis. Despite heterogeneity and limited quality of most included studies, application of TEG/ROTEM seems a promising tool in sepsis. However, given that coagulopathy is a dynamic process, more insight in the kinetics of the coagulation alterations, as diagnosed by TEG/ROTEM is needed before the general use of TEG/ROTEM to detect hyper- or hypocoagulability and DIC can be advocated.

Muller.indd 59 25-7-2014 14:40:21

Chapter 3

60

References


2. Levi M, Meijers JC: DIC: which laboratory tests are most useful. Blood Rev 2011;25: 33-37

3. Reikvam H, Steien E, Hauge B, et al.: Thrombelastography. Transfus Apher Sci 2009;40: 119-123

4. Lang T, Bauters A, Braun SL, et al.: Multi-centre investigation on reference ranges for ROTEM thromboelastometry. Blood Coagul Fibrinolysis 2005;16: 301-310

5. Abraham E, Reinhart K, Opal S, et al.: Efficacy and safety of tifacogin (recombinant tissue factor pathway inhibitor) in severe sepsis: a randomized controlled trial. JAMA 2003;290: 238-247

6. Afshari A, Wetterslev J, Brok J, et al.: Antithrombin III in critically ill patients: systematic review with meta-analysis and trial sequential analysis. BMJ 2007;335: 1248-1251

7. Bernard GR, Vincent JL, Laterre PF, et al.: Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001;344: 699-709

8. Jaimes F, De La Rosa G, Morales C, et al.: Unfractioned heparin for treatment of sepsis: A randomized clinical trial (The HETRASE Study). Crit Care Med 2009;37: 1185-1196

9. Adamzik M, Langemeier T, Frey UH, et al.: Comparison of thrombelastometry with sim-plified acute physiology score II and sequential organ failure assessment scores for the prediction of 30-day survival: a cohort study. Shock 2011;35: 339-342

10. Ostrowski SR, Windelov NA, Ibsen M, et al.: Consecutive thrombelastography clot strength profiles in patients with severe sepsis and their association with 28-day mor-tality: A prospective study. J Crit Care 2013;28: 317

11. Whiting PF, Rutjes AW, Westwood ME, et al.: QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155: 529-536

12. http://www.bris.ac.uk/quadas/quadas-2/ accessed 5 June 201313. Hasegawa H: Estimation of coagulation-fibrinolytic factors in DIC. Bibl Haematol

1983;49: 173-18014. Clayton DG, Miro AM, Kramer DJ, et al.: Quantification of thrombelastographic changes

after blood component transfusion in patients with liver disease in the intensive care unit. Anesth Analg 1995;81: 272-278

15. Mohri M, Suzuki M, Sugimoto E, et al.: Effects of recombinant human soluble throm-bomodulin (rhs-TM) on clot-induced coagulation in human plasma. Thromb Haemost 1998;80: 925-929

16. Baldwin I, Tan HK, Bridge N, et al.: A prospective study of thromboelastography (TEG) and filter life during continuous veno-venous hemofiltration. Ren Fail 2000;22: 297-306

17. Doria C, Mandala L, Smith JD, et al.: Thromboelastography used to assess coagula-tion during treatment with molecular adsorbent recirculating system. Clin Transplant 2004;18: 365-371

Muller.indd 60 25-7-2014 14:40:21


61

18. Dunser MW, Fries DR, Schobersberger W, et al.: Does arginine vasopressin influence the coagulation system in advanced vasodilatory shock with severe multiorgan dysfunction syndrome? Anesth Analg 2004;99: 201-206

19. King DR, Cohn SM, Feinstein AJ, et al.: Systemic coagulation changes caused by pulmo-nary artery catheters: laboratory findings and clinical correlation. J Trauma 2005;59: 853-857

20. Faybik P, Bacher A, Kozek-Langenecker SA, et al.: Molecular adsorbent recirculating sys-tem and hemostasis in patients at high risk of bleeding: an observational study. Crit Care 2006;10: R24

21. Nilsson CU, Hellkvist PD, Engstrom M: Effects of recombinant human activated protein C on the coagulation system: a study with rotational thromboelastometry. Acta Anaes-thesiol Scand 2008;52: 1246-1249

22. Van PY, Cho SD, Underwood SJ, et al.: Thrombelastography versus AntiFactor Xa levels in the assessment of prophylactic-dose enoxaparin in critically ill patients. J Trauma 2009;66: 1509-1515

23. Mineo G, Cascio ND, Canzio D, et al.: Fondaparinux vs low molecular weight heparin as a thromboprophylaxis in critically ill patients. Coagulation status monitored by throm-boelastography [abstract]. Intensive Care Med 2009;Conference: S217


25. White H, Zollinger C, Jones M, et al.: Can Thromboelastography performed on kaolin-activated citrated samples from critically ill patients provide stable and consistent parameters? Int J Lab Hematol 2010;32: 167-173

26. Johansson PI, Stensballe J, Vindelov N, et al.: Hypocoagulability, as evaluated by thromb-elastography, at admission to the ICU is associated with increased 30-day mortality. Blood Coagul Fibrinolysis 2010;21: 168-174

27. Dimitrova-Karamfilova A: Thromboelasometry for the assessment of hypercoagulation in critically ill cardiac patients [abstract]. Transfusion Alternatives in Transfusion Medi-cine 2010;Conference: 37

28. Stravitz RT, Lisman T, Luketic VA, et al.: Minimal effects of acute liver injury/acute liver failure on hemostasis as assessed by thromboelastography. J Hepatol 2012, 56: 129-136

29. Agarwal B, Wright G, Gatt A, et al.: Evaluation of coagulation abnormalities in acute liver failure. J Hepatol 2012;57: 780-786

30. Ryan ML, Thorson CM, King DR, et al.: Insertion of central venous catheters induces a hypercoagulable state. J Trauma Acute Care Surg 2012;73: 385-390

31. Wada H, Sakakura M, Kushiya F, et al.: Thrombomodulin accelerates activated protein C production and inhibits thrombin generation in the plasma of disseminated intravascu-lar coagulation patients. Blood Coagul Fibrinolysis 2005;16: 17-24

32. Picoli-Quaino SK, Alves BE, Aranha FJP, et al.: Evaluation of thrombin generation in the early stages of sepsis in patients with hematological malignancies and febrile neutrope-nia [abstract]. ASH conference proceedings 2011;Conference: 3347

Muller.indd 61 25-7-2014 14:40:21

Chapter 3

62

33. Soshitova NP, Karamzin SS, Balandina AN, et al.: Predicting prothrombotic tendencies in sepsis using spatial clot growth dynamics. Blood Coagul Fibrinolysis 2012;23: 498-507

34. Baudo F, Caimi TM, DeCataldo F, et al.: Antithrombin III (ATIII)) replacement therapy in patients with sepsis and/or postsurgical complications: A controlled double-blind, randomized, multicenter study. Intensive Care Med 1998;24: 336-342

35. Galstyan GM, Krechetova AV, Alexanyan MG, et al.: Thrombin generation in patients (pts) with severe sepsis and septic shock during the treatment with high doses of antithrombin (AT) [abstract]. J Thromb Haemost 2009;Conference: 684

36. Tsen A, Kirschenbaum LA, Larow C, et al.: The effect of anticoagulants and the role of thrombin on neutrophil-endothelial cell interactions in septic SHOCK. Shock 2009;31: 120-124

37. Ogawa Y, Yamakawa K, Ogura H, et al.: Recombinant human soluble thrombomodulin improves mortality and respiratory dysfunction in patients with severe sepsis. J Trauma 2011;72: 1150-1157

38. Durila M, Bronsky J, Harustiak T, et al.: Biochemical and hematological parameters (including thromboelastography) differ in patients with sepsis and SIRS after esophagec-tomy. Crit Care 2011;15 (suppl 1): P443

39. Raineri SM, Cangemi L, Cortegiani A, et al.: Fibrinolysis system, monitored by throm-boelastography (TEG) and PAI-1 activity, in septic patients undergoing tight glycemic control [abstract]. Intensive Care Med 2009;Conference: S162

40. Gonano C, Sitzwohl C, Meitner E, et al.: Four-day antithrombin therapy does not seem to attenuate hypercoagulability in patients suffering from sepsis. Crit Care 2006;10: R160

41. Sivula M, Pettila V, Niemi TT, et al.: Thromboelastometry in patients with severe sepsis and disseminated intravascular coagulation. Blood Coagul Fibrinolysis 2009;20: 419-426

42. Viljoen M, Roux LJ, Pretorius JP, et al.: Hemostatic competency and elastase-alpha 1-proteinase inhibitor levels in surgery, trauma, and sepsis. J Trauma 1995;39: 381-385



45. Umgelter A, Msmer G, Schmid RM, et al.: Thromboelastography and platelet function assay in cirrhosis and sepsis [abstract]. Intensive Care Med 2009;Conference: S160

46. Schmittinger CA, Dunser MW, Torgersen C, et al.: Rotational thromboelastometry to assess the coagulation system in severe sepsis and the systemic inflammatory response syndrome: a prospective, controlled trial [abstract]. Intensive Care Med 2009;Confer-ence: S162

47. Chiairi F, Ickx B, Barvais L, et al.: Does activated protein C influence the coagulation system assessed by the rotative thromboelastometry analysis [abstract]. Intensive Care Med 2009;Conference: S162

Muller.indd 62 25-7-2014 14:40:21


63

48. Altmann DR, Korte W, Maeder MT, et al.: Elevated Cardiac Troponin I in Sepsis and Sep-tic Shock: No Evidence for Thrombus Associated Myocardial Necrosis. Plos One 2010;5: e9017

49. Durila M, Kalincik T, Jurcenko S, et al.: Arteriovenous differences of hematological and coagulation parameters in patients with sepsis. Blood Coagul Fibrinolysis 2010;21: 770-774

50. Adamzik M, Eggmann M, Frey UH, et al.: Comparison of thromboelastometry with pro-calcitonin, interleukin 6, and C-reactive protein as diagnostic tests for severe sepsis in critically ill adults. Crit Care 2010;14: R178

51. Cortegiani A, Marino L, Montalto F, et al.: Use of thromboelastography in severe sepsis: a case-control study [abstract]. Crit Care 2011;15(suppl 1): P444

52. Massion PB, Peters P, Ledoux D, et al.: Persistent hypocoagulability in patients with sep-tic shock predicts greater hospital mortality: impact of impaired thrombin generation. Intensive Care Med 2012;38: 1326-1335

53. Durila M, Bronsky J, Harustiak T, et al.: Early diagnostic markers of sepsis after oesophagectomy (including thromboelastography). BMC Anesthesiol 2012;12: 12

54. Brenner T, Schmidt K, Delang M, et al.: Viscoelastic and aggregometric point-of-care testing in patients with septic shock - cross-links between inflammation and haemosta-sis. Acta Anaesthesiol Scand 2012;56: 1277-1290

55. Bone RC, Balk RA, Cerra FB, et al.: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992;101: 1644-1655

56. Levi M, de JE, van der Poll T: Sepsis and disseminated intravascular coagulation. J Thromb Thrombolysis 2003;16: 43-47

57. Schochl H, Solomon C, Schulz A, et al.: Thromboelastometry (TEM) findings in dissemi-nated intravascular coagulation in a pig model of endotoxinemia. Mol Med 2011;17: 266-272

58. Spiel AO, Mayr FB, Firbas C, et al.: Validation of rotation thrombelastography in a model of systemic activation of fibrinolysis and coagulation in humans. J Thromb Haemost 2006;4: 411-416

59. Zacharowski K, Sucker C, Zacharowski P, et al.: Thrombelastography for the monitor-ing of lipopolysaccharide induced activation of coagulation. Thromb Haemost 2006;95: 557-561

60. Taylor FB, Jr., Toh CH, Hoots WK, et al.: Towards definition, clinical and laboratory crite-ria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost 2001;86: 1327-1330

61. 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 2008;19: 495-501

62. Nielsen VG: A comparison of the Thrombelastograph and the ROTEM. Blood Coagul Fibrinolysis 2007;18: 247-252

Muller.indd 63 25-7-2014 14:40:22

Chapter 3

64

63. Solomon C, Sorensen B, Hochleitner G, et al.: Comparison of whole blood fibrin-based clot tests in thrombelastography and thromboelastometry. Anesth Analg 2012;114: 721-730

64. Venema LF, Post WJ, Hendriks HG, et al.: An assessment of clinical interchangeability of TEG and RoTEM thromboelastographic variables in cardiac surgical patients. Anesth Analg 2010;111: 339-344

65. Massberg S, Grahl L, von Bruehl ML, et al.: Reciprocal coupling of coagulation and innate immunity via neutrophil serine proteases. Nat Med 2010;16: 887-896

Muller.indd 64 25-7-2014 14:40:22

Chapter 4

Thromboelastometry and multiple organ

failure in trauma patients

Marcella C.A. Müller, Kirsten Balvers, Jan M. Binnekade, Nicola Curry, Simon Stanworth, Christine Gaarder, Knut M. Kolstadbraaten, Claire Rourke, Karim Brohi,

J.Carel Goslings, Nicole P. Juffermans

Submitted for publication

Muller.indd 65 25-7-2014 14:40:22

Chapter 4

66

Abstract

Purpose: Data on the incidence of a hypercoagulable state in trauma, as measured by ROTEM, are limited and the prognostic value of hypercoagulability after trauma on outcome is unclear. We aimed to determine the incidence of hypercoagulability after trauma and to assess whether occurrence of hypercoagulability has prognostic value on the occurrence of multiple organ failure (MOF) and mortality.Methods: Prospective observational cohort study in trauma patients who met high-est trauma level team activation. Hypercoagulability was defined as a G value of ≥ 11.7 dynes/cm2 and hypocoagulability as a G value of <5.0 dynes/cm2. ROTEM was performed on admission and 24 hours later.Results: 1010 patients were enrolled and 948 patients were analyzed. Median age was 38 [IQR 26-53], 77% were male, median Injury Severity Score was 13 [IQR 8-25]. On admission, 7% of the patients were hypercoagulable and 8% were hypocoag-ulable. Altogether, 10% of patients showed hypercoagulability within the first 24 hours of trauma. Hypocoagulability, but not hypercoagulability, was associated with higher SOFA scores, indicating more severe MOF. Mortality in patients with hyper-coagulability was 0%, compared to 7% in normocoagulable and 24% in hypocoagu-lable patients (p<0.001). EXTEM CT, alpha and G were predictors for occurrence of MOF and mortality.Conclusion: The incidence of a hypercoagulable state after trauma is 10% up to 24 hours after admission, which is broadly comparable to the rate of hypocoagu-lability. Further work in larger studies should define the clinical consequences of identifying hypercoagulability and a possible role for very early, targeted use of anti-coagulants.

Muller.indd 66 25-7-2014 14:40:22

Thromboelastometry in trauma patients

67

Introduction

Major trauma is among the most common causes of death worldwide [1]. Whereas uncontrolled bleeding accounts for 50-80% of mortality early following trauma [2,3], multiple organ failure (MOF) is the most important cause of late mortality after trauma [2,4]. Traumatic injury induces a hypocoagulable state, as a result of acute traumatic coagulopathy (ATC) accompanied by loss, consumption and dilu-tion of coagulation factors and fibrinolysis. Hypothermia, shock and acidosis fur-ther amplify the derangement of the coagulation system [5]. In addition to reduced hemostatic potential, trauma can also induce a hypercoagulable state [6-8]. Animal experiments have shown that hypercoagulability can arise within hours of the injury [9], a phenomenon confirmed in humans [6,10]. However, uniform definitions of hypercoagulability are lacking [11] and effects of this hypercoagulable state after trauma are not fully elucidated, with studies showing conflicting results. An associa-tion with adverse events such as an increased risk of venous thrombo-embolism has been reported [8,12,13]. However, early hypercoagulability has also been associated with decreased early mortality, which may suggest that hypercoagulability is a func-tional response in order to reduce blood loss [10]. In addition to this endogenous response, the paradigm of treatment of ATC has shifted, including earlier administra-tion of larger amounts of fresh frozen plasma and other hemostatic agents. This was shown to improve outcome [14], presumably by enhancing procoagulant abilities.In sepsis, it has been demonstrated that hypercoagulability, characterized by the formation of micro-thrombi with concurrent protein C deficiency and impaired fibrinolysis, contributes to MOF and adverse outcome [15-17]. Although sepsis and trauma are different entities, the accompanying coagulopathies show similarities and persistent protein C deficiency after trauma is also associated with occurrence of MOF [18,19]. Shock and hypoperfusion can induce activation of the endothe-lium and if the patient survives the initial bleeding episode, this can result in a pro-coagulant state. It is conceivable that therapy of ATC may add to this endogenous response, possibly resulting in an overshoot in coagulation over time, with subse-quent enhancement of hypercoagulability and MOF.Diagnosing hypercoagulability is complex. Thrombin generation tests, or assess-ment of plasma levels of natural anticoagulants as protein C, protein S and tissue

Muller.indd 67 25-7-2014 14:40:22

Chapter 4

68

factor pathway inhibitor are not readily available for clinical use and not validated to detect hypercoagulability. Thromboelastometry (ROTEM) is a bedside available test providing real time information on all aspects of the coagulation system, includ-ing the presence of hypercoagulability [20,21]. The use of thromboelastography to diagnose hypocoagulability in trauma has frequently been explored in recent years [6-8,10,12,13,22,23]. However, reports on the use of ROTEM to detect a hyperco-agulable state are scarce.We aimed to study the incidence of hypercoagulability in multiple trauma patients in the first 24 hours after injury and to establish whether hypercoagulability was associated with the occurrence of MOF and mortality. In addition, as transfusion strategies have shifted, we assessed whether transfusion strategy influenced the occurrence of hypercoagulability.

Methods

Study design and patientsA prospective observational cohort study was conducted in four level-1 trauma centers in London, Oxford, Oslo and Amsterdam. This study is part of the Activa-tion of Coagulation and Inflammation in Trauma (ACIT) study, an ongoing prospec-tive observational multicenter study in trauma patients. Local ethics committees reviewed and approved the study. All procedures have been performed in accord-ance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. Between January 2008 and March 2013, all adult trauma patients (18 years and older) who met the local criteria for highest trauma team level activation were eligible for enrollment in the study. Patients were excluded if arrival at the emergency department (ED) was >2 hours following injury; >2000 ml of intravenous fluid was administered before ED admission; they were transferred from another hospital or if they had burns covering >5% of total body surface area. Patients were retrospectively excluded if they declined to give consent to use data, were receiving anticoagulation (not including aspirin), or had moderate or severe liver disease or a known bleeding diathesis.

Muller.indd 68 25-7-2014 14:40:22


69

Data collectionData were prospectively collected on patient demographics, time from injury to arrival at the ED, mechanism of injury (blunt or penetrating), presence of traumatic brain injury, vital signs on arrival and 24 hours after injury and amount of fluid and blood products within the first 24 hours of injury. Trauma severity was assessed using the Injury Severity Score (ISS) [24]. Outcome measures were MOF, defined as worst Sequential Organ Failure Assessment (SOFA) score during admission, and mortality after 28 days.

ThromboelastometryThromboelastometric variables were measured with ROTEM (Tem International, Munich, Germany). Citrated blood samples were drawn within 1 hour after arrival in the ED and a second sample was collected 24 hours (±2 hours) after admission. All samples were processed within 1 hour. The ROTEM EXTEM assay was carried out to assess tissue factor initiated coagulation and the ROTEM INTEM assay to assess the intrinsic pathway. For EXTEM, 20 μL of 0.2 mol/L CaCl2 (star-tem®) and 20 μL of human recombinant tissue factor (r EXTEM®) were added to a test vial. Subsequently 300 μL of the citrated blood sample was added. For INTEM, 20 μL of 0.2 mol/L CaCl2 (star-tem®), 20 μL of partial thromboplastin made of rabbit brain (in-tem®) and 300 μL of blood were added to the test cuvette.The electronic pipette program guided all test steps. For both assays, clotting time (CT), clot formation time (CFT), maximum clot firmness (MCF) and alpha angle were recorded. Total clot strength was assessed by G as calculated according to the for-mula: (5000xMCF)/100-MCF and expressed as dynes/cm2 [20]. G has a curvilinear relation with MCF and reflects the contribution of enzymatic and platelet compo-nents to the hemostasis, hereby better reflecting hemostatic potential than indi-vidual thromboelastometry parameters [8,25]. G has been shown to be valuable in diagnosing hypo- and hypercoagulability [8,20,25]. Hypercoagulability was defined as a G value of ≥ 11.7 dynes/cm2 and hypocoagulability as a G value of <5.0 dynes/cm2 (values provided by manufacturer).

Muller.indd 69 25-7-2014 14:40:22

Chapter 4

70

Outcome variablesPrimary outcome was the occurrence of MOF, assessed by the SOFA score, which reliably assesses organ failure in trauma patients [26]. The score awards 0 (normal) to 4 (most abnormal) points for each organ system. MOF was defined by a score of 3 points or more [4]. Secondary outcome was 28-day mortality. In addition, effect of transfusion strategy (ratio Red Blood Cells (RBC): Fresh Frozen Plasma (FFP)) on ROTEM profile and occurrence of hypercoagulability was determined.

StatisticsContinuous normally distributed variables are expressed by their mean and stand-ard deviation. Not normally distributed variables are expressed as medians and their interquartile ranges and categorical variables are expressed as n (%). ISS was treated as a continuous variable. Groups are compared by using Student’s t-test or Mann-Whitney U test in case of not normally distributed data.For comparison of categorical variables, the Chi-square test or Fisher’s exact tests are used.The primary analysis focused on modeling the hypothesized relation between ROTEM detected hypercoagulability, MOF and mortality in trauma patients. First, univariate logistic regression analysis was used to select independent factors achieving a p value ≤ 0.10, in addition to factors that were deemed clinically impor-tant (age, time to ED, presence of traumatic brain injury, injury mechanism, ISS, Base Excess, systolic blood pressure) in relation to the outcome variables. Subsequently, selected ROTEM factors were entered in a multivariate logistic regression model. Patients who died on admission were not included in the analyses to assess the value of thromboelastometry to predict MOF, while patients who died later were included when a SOFA score was available. All deceased patients were included in the analyses to assess the value of ROTEM to predict mortality.To compare the effect of transfusion strategies, transfused patients were divided based on RBC:FFP ratio. Statistical significance was considered to be at p 0.05. Analy-ses were performed using R (version 2·3; R Foundation for Statistical Computing, Vienna, Austria). Graphs were created with Prism 5.0 (GraphPad Software San Diego, CA, USA).

Muller.indd 70 25-7-2014 14:40:22


71

Figure 1: Flow diagram of inclusion and occurrence of multiple organ failure and mortality

Muller.indd 71 25-7-2014 14:40:22

Chapter 4

72

Results

During the study period, 1245 patients were screened and 1010 patients were enrolled in the study (figure 1). For 62 of the patients, no data were available on occurrence of MOF or mortality, therefore analyses were performed in the remain-ing 948 patients. Patient characteristics are listed in table 1. The majority of included patients were males experiencing blunt injury. Median age was 38 years and median ISS was 13 (IQR 8-25).

ROTEM profiles and hypercoagulability on admissionBaseline thromboelastometry data were available for 886 patients upon ED admis-sion and for 451 patients when assessed 24 hours after admission. On admission, the G value was increased in 63 (7%) of the patients, while 71 (8%) were hypoco-agulable and the remaining 85% had normal clot strength according to the G value. Patients showing hypercoagulability on admission were more often female (40% vs. 28%, p<0.001), had lower ISS scores (9 vs. 20, p<0.001) and higher base excess values (-1.3 mEq/L vs. -4.3 mEq/L, p<0.001) compared to hypocoagulable patients. Also, they received less RBC, FFP and platelet transfusions compared to hypocoagu-lable patients. In addition, hypocoagulable patients had longer time to arrival at ED (table 1).

ROTEM profiles and hypercoagulability 24 hours after admissionAfter 24 hours, 26 (6%) patients were hypercoagulable and 35 (8%) were hypocoagu-lable (Supplemental file: figure 1). In accordance with the hypercoagulable patients at ED admission, the hypercoagulable patients 24 hours after admission had lower ISS scores (14 vs. 25, p=0.04), higher base excess values (-1.4 mEq/L vs. -6.2 mEq/L, p<0.001) and received less RBC transfusions compared to the hypocoagulable patients. Amount of FFP and platelets transfused did not differ between hyper-, normo- and hypocoagulable patients.Altogether, during the first 24 hours after trauma, 88 (10%) patients were hyperco-agulable at some point.

Muller.indd 72 25-7-2014 14:40:22


73

Table 1: Characteristics of patients with hyper-, hypo- and normocoagulable ROTEM pro-files at admission.

All patientsN=948

Hyper-coagulable1

N=63

Normo-coagulable2

N=752

Hypo-coagulable3

N=71

p value

Age (years) 38 [26-53] 44 [33-62] 38 [25-53] 38 [25-54] <0.05Sex, male % (n) 77 (730) 60 (38) 80 (599) 72 (51) <0.001Time to ED ( minutes)

70 [51-90] 71 [46-86] 70 [53-88] 80 [60-100] 0.05

Trauma mecha-nism, blunt % (n)

82 (777) 81 (51) 82 (619) 86 (61) 0.69

Brain injury, % (n) 28 (265) 23 (14) 27 (193) 38 (26) 0.09ISS 13 [8-25] 9 [5-17] 13 [5-25] 20 [10-39] <0.001Systolic BP, (mmHg)*

130 (30) 136 (28) 131 (29) 122 (34) 0.06

Base Excess (mEq/L)

-1.5 [-4.2 – 0.6]

-1.3 [-3.2-0.2]

-1.2 [-3.7-0.8]

-4.3 [-9.5—0.5]

<0.001

RBC (units) 5 [3-8] 4 [3-5] 4 [3-8] 6 [4-11] <0.05FFP (units) 4 [4-8] 3 [2-4] 4 [4-8] 6 [4-13] 0.001PLT (units) 1 [1-2] 1 [1-1] 1 [1-2] 2 [1-5] <0.01Cryoprecipitate 2 [2-2] NA 2 [2-2] 2.5 [2-5] 0.06

1Hypercoagulable G≥11.7 dynes/cm2

2Normocoagulable G=5-11.7 dynes/cm2

3Hypocoagulable G<5 dynes/cm2

All variables expressed as median and interquartile ranges [IQR].*expressed as mean and standard deviation (SD)ED = emergency departmentISS = injury severity scoreBP = blood pressureRBC = red blood cellFFP = fresh frozen plasma

ROTEM profiles and multiple organ failure41% of trauma patients developed MOF (figure 1). These patients were older, had higher ISS scores, more often had brain injury and received more blood products (Supplemental file: table 1). Of patients who were hyper- or normocoagulable on admission, 40% developed MOF, compared to 53% of the hypocoagulable patients. In patients presenting with hypocoagulability, the worst SOFA scores were higher compared to those who were normo- or hypercoagulable on admission (p=0.003, figure 2). Also, patients who developed MOF had hypocoagulable admission profiles as measured by ROTEM compared to patients who did not develop MOF (table 2).

Muller.indd 73 25-7-2014 14:40:22

Chapter 4

74

Figure 2: Occurrence of multiple organ failure and worst SOFA scores in patients with hypo-, normo- and hypercoagulable profiles at admission and 24 hours after admission. Gray bars indicate occurrence of multiple organ failure and black dots indicate median SOFA scores and interquartile ranges.

*p<0.01**p<0.05MOF = multiple organ failureSOFA = sequential organ failure assessment

The same picture was noted 24 hours after admission. Worst median SOFA scores were highest among patients showing hypocoagulability 24 hours after admission, indicating more severe organ failure in these patients (figure 2).

ROTEM profiles and prediction of MOF and mortalityUnivariate logistic regression analysis with admission ROTEM variables identified INTEM CFT, INTEM alpha, INTEM MCF, EXTEM CT, EXTEM alpha, EXTEM MCF and G to be associated with the occurrence of MOF, as were trauma characteristics and baseline vital parameters. After performing multiple logistic regression analysis with

Muller.indd 74 25-7-2014 14:40:23


75

Table 2: Thromboelastometry results at admission of patients who did and did not develop multiple organ failure.

MOFN=381

No MOFN=549

p value

Intem CT (sec) 138 [115 to 168] 134 [113 to 166] 0.22Intem CFT (sec) 80 [63 to 104] 71 [60 to 89] <0.001Intem alpha° 74 [70 to 77] 76 [73 to 78] <0.001Intem MCF (mm) 60 [56 to 64] 62 [58 to 65] <0.001Extem CT (sec) 59 [49 to 73] 55 [46 to 68] 0.002Extem CFT (sec) 98 [78 to 122] 88 [72 to 105] <0.001Extem alpha° 71 [66 to 75] 73 [69 to 76] <0.001Extem MCF (mm) 60 [56 to 65] 62 [58 to 66] 0.005Extem G (dynes/cm2) 7.5 [6.4 to 9.3] 8.2 [6.9 to 9.7] 0.005

Median and interquartile range [IQR]MOF = multiple organ failureCT = clotting timeCFT = clot formation timeMCF = maximum clot firmness

ROTEM variables, admission EXTEM CT, alpha and G were shown to be predictors for the occurrence of MOF (table 3). The odds ratios for MOF indicated that change of the parameters towards a more hypocoagulable profile resulted in an increased risk for the development of MOF. We did not find any correlation between a hyper-coagulable profile and the occurrence of MOF.EXTEM CFT and G were predictors for MOF 24 hours after admission (table 3).The total mortality was 10% (n=97) (figure 1). Of note, patients who were hyperco-agulable on admission had lower 28-day mortality compared to normo- and hypo-coagulable patients (0% in hypercoagulable patients vs. 7% in normocoagulable and 24% in hypocoagulable patients, p<0.001). Multivariate analysis with ROTEM vari-ables showed that low EXTEM alpha angle on admission was a predictor for mortal-ity (0.95 (0.91-0.98) p<0.01). Every degree increase of the alpha angle results in a 0.95 reduction of mortality risk.

ROTEM profiles, transfusion strategy and occurrence of MOFIn order to assess whether liberal use of FFP affected occurrence of hypercoagu-lability and subsequent MOF, we performed an additional sub-analysis in patients

Muller.indd 75 25-7-2014 14:40:23

Chapter 4

76

Table 3: Prediction of occurrence of multiple organ failure by EXTEM ROTEM variables at admission and 24 hours after admission with multivariate analysis.

OR 95% CI p valueAdmissionCTCFTAlphaMCFG

1.010.990.951.010.94

1.00-1.010.99-1.000.92-0.980.98-1.040.89-0.99

0.050.23

<0.010.620.02

24 hours after admissionCTCFTAlphaMCFG

1.001.031.041.040.92

0.00-2.391.01-1.040.95-1.150.99-1.090.85-1.00

0.86<0.010.370.130.05

CT = clotting timeCFT = clot formation timeMCF = maximum clot firmness

transfused with RBC and FFP. Transfused patients were divided in one group with an RBC:FFP ratio of 1:1 (n=35), one with a ratio of more then 1:1 (n=115) and one with a ratio of less then 1:1 (n=21). These 3 groups did not differ significantly with respect to baseline characteristics (data not shown) and platelet transfusions. ROTEM EXTEM CT, CFT, MCF and alpha did not differ at baseline and after 24 hours, nor did G values (data not shown). After 24 hours, none of the patients transfused with a RBC:FFP ratio <1:1 showed hypercoagulability and of patients transfused with higher ratios of RBC:FFP, only 2 out of 100 patients progressed from a normocoagulable to a hypercoagulable state (Supplemental file: figure 2).Occurrence of MOF was high in all groups, but did not differ between groups with different transfusion ratios (82% in patients with a ratio of 1:1 or higher, and 81% in patients with RBC:FFP <1:1 respectively, p=0.99).

Discussion

The current study shows that a hypercoagulable state as detected by thromboelas-tometry, occurred in 7% at admission and in 10% of patients within 24 hours after trauma. Characteristics associated with a presence of hypercoagulability included lower ISS, higher base excess values, female gender and shorter time to ED arrival.

Muller.indd 76 25-7-2014 14:40:23


77

These rates were not that different to the detected incidence for hypocoagulabil-ity, which has been the focus of considerable research interest, as part of evolving concepts of ATC. In contrast to our hypothesis, hypercoagulability did not appear to predict the occurrence of MOF. Rather, severity of MOF after trauma was associ-ated with a hypocoagulable state. Hypercoagulable patients at admission had lower mortality, consistent with lower ISS and better base excess values. High EXTEM CFT and low G values were predictive for the development of MOF and low EXTEM alpha was predictive for mortality.Hypercoagulability after trauma has been reported using a variety of TEG measure-ments [6,8,10,27,28], however whether TEG and ROTEM results are interchangeable is still under debate [29,30]. Reported incidences of hypercoagulability diagnosed by TEG range from 11 to 80% [6,8,10,12,28]. This wide variation can be ascribed to use of different definitions of hypercoagulability, with studies using individual parameters of the thromboelastographic trace [6,10,28], a combination of param-eters [7,12] or the use of G as a marker of whole clot strength [8]. Also injury severity and timing of measurements differed among these cohorts. However, our findings are in line with those previously reported in a smaller cohort of trauma patients [28].We hypothesized that occurrence of hypercoagulability was associated with MOF. However, we observed an opposite effect. Patients showing hypocoagulability within the first 24 hours of admission developed more severe organ failure and had an increased late mortality. This observation is in line with studies demonstrating that hypocoagulability is associated with adverse outcome after trauma and brain injury [22,25,31]. We showed that, in addition to individual parameters, G values on admission and 24 hours after admission are predictors for the occurrence of MOF. Of note, G is considered to better represent total clot strength than the individual thromboelastography parameters [8,20,25]. ROTEM has been studied extensively in trauma patients, mainly focusing on diagnosing early coagulation abnormalities [32,33], prediction of transfusion requirements [34,35] and correction of hypoco-agulability [36-38]. Our data indicate, contrary to our hypothesis, that hypocoagula-bility detected by ROTEM also has an enhanced risk of adverse outcome, which is in line with previous observations in a smaller cohort [35].In the current study, patients with hypercoagulability had lower mortality and lower SOFA scores. A similar observation was recently reported in a smaller cohort of

Muller.indd 77 25-7-2014 14:40:23

Chapter 4

78

trauma patients [10]. Based on our data, we hypothesize that early hypercoagulabil-ity after trauma is an evolutionary response to prevent exsanguination. Results do not point towards the hypothesis that early hypercoagulability after trauma resem-bles DIC with the formation of micro-thrombi, thereby contributing to organ failure [19,39]. The observation that early hypercoagulability after trauma is more preva-lent in females, is in line with a previous report [6].Limitations of our study include that we did not systematically look for occurrence of venous thrombo-embolism. It is also not possible to rule out a contribution of late hypercoagulability to the development of MOF, as we only assessed ROTEM on admission and after 24 hours. Prolonged hypercoagulability has been linked to increased risk of thrombo-embolic complications [8,12,13]. Also, we did not assess d-dimers and hereby we were not able to correlate ROTEM findings to DIC scores. However, a recent review of pathology samples obtained early after trauma failed to demonstrate micro-thrombi despite the clinical presence of increased DIC scores [40]. Furthermore, we had a number of missing values at 24 hours following trauma. Limited reports have described the coagulopathic changes over time in trauma and a recent small cohort study suggested that hypercoagulability after trauma occurs after 48 hours [41]. Therefore, further research should include serial measurements and a prospective standardized observation of complications after trauma. How-ever, the current data suggest that early hypercoagulability after trauma not only reduces early mortality [10], but also seems to be associated with lower occurrence and severity of MOF and 28 day mortality.Altogether, this study has identified a significant proportion of patients with hyper-coagulability as defined by ROTEM at admission. Further work in larger studies should define the clinical consequences and prognostic value of identifying hyper-coagulability, specifically including thrombo-embolic events, and might assess a role for very early, targeted use of anti-coagulants in selected patients. The role of plasma or other blood components in potentially exacerbating the consequences of hypercoagulability is also an area of further research. In this study, patients with hypocoagulability on admission mostly tended to regress to normal values over time and not to hypercoagulability, irrespective of blood product ratio. This is in contrast with studies showing an association between amount of blood products and MOF [42,43], but is in line with other studies, which suggested that other fluids were more

Muller.indd 78 25-7-2014 14:40:23


79

associated with MOF than blood products [44,45]. Also, there is experimental evi-dence that FFP preserves endothelial integrity in hemorrhagic shock [46].

Conclusion

In a cohort of trauma patients, 10% shows a hypercoagulable state, as defined by ROTEM G value, within the first 24 hours. Occurrence of early hypercoagulability is not associated with development of MOF, moreover it appears to protect against adverse outcome. Admission ROTEM variables indicating hypocoagulability are pre-dictive of the development of MOF and mortality. Liberal use of FFP is not associ-ated with enhanced hypercoagulability.

Muller.indd 79 25-7-2014 14:40:23

Chapter 4

80

References

1. Lozano R, Naghavi M, Foreman K, et al.: Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380: 2095-128

2. Sauaia A, Moore FA, Moore EE, et al.: Epidemiology of trauma deaths: a reassessment. J Trauma 1995;38: 185-93

3. Hoyt DB, Bulger EM, Knudson MM, et al.: Death in the operating room: an analysis of a multi-center experience. J Trauma 1994;37: 426-32

4. Ulvik A, Kvale R, Wentzel-Larsen T, et al.: Multiple organ failure after trauma affects even long-term survival and functional status. Crit Care 2007;11: R95

5. Brohi K, Cohen MJ, Davenport RA: Acute coagulopathy of trauma: mechanism, identifi-cation and effect. Curr Opin Crit Care 2007;13: 680-5

6. Schreiber MA, Differding J, Thorborg P, et al.: Hypercoagulability is most prevalent early after injury and in female patients. J Trauma 2005;58: 475-80

7. Kaufmann CR, Dwyer KM, Crews JD, et al.: Usefulness of thrombelastography in assess-ment of trauma patient coagulation. J Trauma 1997;42: 716-20

8. Kashuk JL, Moore EE, Sabel A, et al.: Rapid thrombelastography (r-TEG) identifies hypercoagulability and predicts thromboembolic events in surgical patients. Surgery 2009;146: 764-72

9. Mulier KE, Greenberg JG, Beilman GJ: Hypercoagulability in porcine hemorrhagic shock is present early after trauma and resuscitation. J Surg Res 2012;174: e31-5

10. Branco BC, Inaba K, Ives C, et al.: Thromboelastogram Evaluation of the Impact of Hypercoagulability in Trauma Patients. Shock 2014;41:200-7

11. Dai Y, Lee A, Critchley LA, et al.: Does thromboelastography predict postoperative thromboembolic events? A systematic review of the literature. Anesth Analg 2009;108: 734-42


13. Cotton BA, Minei KM, Radwan ZA, et al.: Admission rapid thrombelastography predicts development of pulmonary embolism in trauma patients. J Trauma Acute Care Surg 2012;72: 1470-5

14. del Junco DJ, Holcomb JB, Fox EE, et al.: Resuscitate early with plasma and platelets or balance blood products gradually: findings from the PROMMTT study. J Trauma Acute Care Surg 2013;75: S24-30

15. Fourrier F, Chopin C, Goudemand J, et al.: Septic shock, multiple organ failure, and dis-seminated intravascular coagulation. Compared patterns of antithrombin III, protein C, and protein S deficiencies. Chest 1992;101: 816-23

16. Yan SB, Helterbrand JD, Hartman DL, et al.: Low levels of protein C are associated with poor outcome in severe sepsis. Chest 2001;120: 915-22

Muller.indd 80 25-7-2014 14:40:23


81

17. Madoiwa S, Nunomiya S, Ono T, et al.: Plasminogen activator inhibitor 1 promotes a poor prognosis in sepsis-induced disseminated intravascular coagulation. Int J Hematol 2006;84: 398-405

18. Cohen MJ, Call M, Nelson M, et al.: Critical role of activated protein C in early coagulop-athy and later organ failure, infection and death in trauma patients. Ann Surg 2012;255: 379-85

19. Gando S, Kameue T, Matsuda N, et al.: Combined activation of coagulation and inflam-mation has an important role in multiple organ dysfunction and poor outcome after severe trauma. Thromb Haemost 2002;88: 943-9

20. Taura P, Rivas E, Martinez-Palli G, et al.: Clinical markers of the hypercoagulable state by rotational thrombelastometry in obese patients submitted to bariatric surgery. Surg Endosc 2014;28: 543-51

21. Muller MC, Meijers JC, Vroom MB, et al.: Utility of thromboelastography and/or throm-boelastometry in adults with sepsis: a systematic review. Crit Care 2014;18: R30

22. Nystrup KB, Windelov NA, Thomsen AB, et al.: Reduced clot strength upon admission, evaluated by thrombelastography (TEG), in trauma patients is independently associ-ated with increased 30-day mortality. Scand J Trauma Resusc Emerg Med 2011;19: 52

23. Holcomb JB, Minei KM, Scerbo ML, et al.: Admission rapid thrombelastography can replace conventional coagulation tests in the emergency department: experience with 1974 consecutive trauma patients. Ann Surg 2012;256: 476-86

24. Baker SP, O’Neill B, Haddon W, Jr., et al.: The injury severity score: a method for describ-ing patients with multiple injuries and evaluating emergency care. J Trauma 1974;14: 187-96

25. Kashuk JL, Moore EE, Wohlauer M, et al.: Initial experiences with point-of-care rapid thrombelastography for management of life-threatening postinjury coagulopathy. Transfusion 2012;52: 23-33

26. Antonelli M, Moreno R, Vincent JL, et al.: Application of SOFA score to trauma patients. Sequential Organ Failure Assessment. Intensive Care Med 1999;25: 389-94

27. Differding JA, Underwood SJ, Van PY, et al.: Trauma induces a hypercoagulable state that is resistant to hypothermia as measured by thrombelastogram. Am J Surg 2011;201: 587-91

28. Ostrowski SR, Sorensen AM, Larsen CF, et al.: Thrombelastography and biomarker profiles in acute coagulopathy of trauma: a prospective study. Scand J Trauma Resusc Emerg Med 2011;19: 64

29. Aleshnick M, Orfeo T, Brummel-Ziedins K, et al.: Interchangeability of rotational elasto-graphic instruments and reagents. J Trauma Acute Care Surg 2014;76: 107-13

30. Hagemo JS, Naess PA, Johansson P, et al.: Evaluation of TEG((R)) and RoTEM((R)) inter-changeability in trauma patients. Injury 2013;44: 600-5

31. Kunio NR, Differding JA, Watson KM, et al.: Thrombelastography-identified coagulopa-thy is associated with increased morbidity and mortality after traumatic brain injury. Am J Surg 2012;203: 584-8

32. Rugeri L, Levrat A, David JS, et al.: Diagnosis of early coagulation abnormalities in trauma patients by rotation thrombelastography. J Thromb Haemost 2007;5: 289-95

Muller.indd 81 25-7-2014 14:40:23

Chapter 4

82

33. Meyer AS, Meyer MA, Sorensen AM, et al.: Thrombelastography and rotational throm-boelastometry early amplitudes in 182 trauma patients with clinical suspicion of severe injury. J Trauma Acute Care Surg 2014;76: 682-90

34. Schochl H, Cotton B, Inaba K, et al.: FIBTEM provides early prediction of massive trans-fusion in trauma. Crit Care 2011;15: R265

35. Tauber H, Innerhofer P, Breitkopf R, et al.: Prevalence and impact of abnormal ROTEM(R) assays in severe blunt trauma: results of the ‘Diagnosis and Treatment of Trauma-Induced Coagulopathy (DIA-TRE-TIC) study’. Br J Anaesth 2011;107: 378-87

36. Schochl H, Nienaber U, Hofer G, et al.: Goal-directed coagulation management of major trauma patients using thromboelastometry (ROTEM)-guided administration of fibrino-gen concentrate and prothrombin complex concentrate. Crit Care 2010;14: R55

37. Schochl H, Nienaber U, Maegele M, et al.: Transfusion in trauma: thromboelastome-try-guided coagulation factor concentrate-based therapy versus standard fresh frozen plasma-based therapy. Crit Care 2011;15: R83

38. Rourke C, Curry N, Khan S, et al.: Fibrinogen levels during trauma hemorrhage, response to replacement therapy, and association with patient outcomes. J Thromb Haemost 2012;10: 1342-51

39. Gando S: Acute coagulopathy of trauma shock and coagulopathy of trauma: a rebuttal. You are now going down the wrong path. J Trauma 2009;67: 381-3

40. Rizoli S, Nascimento B, Jr., Key N, et al.: Disseminated intravascular coagulopathy in the first 24 hours after trauma: the association between ISTH score and anatomopathologic evidence. J Trauma 2011;71: S441-7

41. Chapman BC, Moore EE, Barnett C, et al.: Hypercoagulability following blunt solid abdominal organ injury: when to initiate anticoagulation. Am J Surg 2013;206: 917-23

42. Johnson JL, Moore EE, Kashuk JL, et al.: Effect of blood products transfusion on the development of postinjury multiple organ failure. Arch Surg 2010;145: 973-7

43. Silverboard H, Aisiku I, Martin GS, et al.: The role of acute blood transfusion in the development of acute respiratory distress syndrome in patients with severe trauma. J Trauma 2005;59: 717-23

44. Park PK, Cannon JW, Ye W, et al.: Transfusion strategies and development of acute res-piratory distress syndrome in combat casualty care. J Trauma Acute Care Surg 2013;75: S238-46

45. Robinson BR, Cotton BA, Pritts TA, et al.: Application of the Berlin definition in PROM-MTT patients: the impact of resuscitation on the incidence of hypoxemia. J Trauma Acute Care Surg 2013;75: S61-7

46. Torres LN, Sondeen JL, Ji L, et al.: Evaluation of resuscitation fluids on endothelial gly-cocalyx, venular blood flow, and coagulation function after hemorrhagic shock in rats. J Trauma Acute Care Surg 2013;75: 759-66

Muller.indd 82 25-7-2014 14:40:23


83

Supplemental files

Table 1: Characteristics of patients who did and did not develop multiple organ failure.

Multiple Organ FailureN = 381

No Multiple Organ FailureN = 549

p value

Age (years) 40 [27-57] 35 [24-49] <0.001Sex, male % (n) 77 (303) 80 (422) 0.41Time to ED (minutes) 78 [60-95] 65 [47-83] <0.001Trauma mechanism, blunt % (n) 87 (327) 79 (430) 0.001Brain injury, % (n) 42 (160) 15 (80) <0.001Injury severity score 24 [13-34] 9 [4-17] < 0.001Systolic blood pressure (mmHg) 124 (32) 134 (27) <0.001Base Excess (mEq/L) -3 [-6.35 - -0.80] -0.45 [-2.43 - 1.13] <0.001Red Blood Cells Patients transfused, % (n) 44 (168) 12 (65) <0.001 Units 6 [4-10] 3 [2-6] <0.001Fresh frozen plasma Patients transfused, % (n) 34 (130) 6 (30) <0.0001 Units 5 [4-8] 4 [2-4] <0.001Platelets Patients transfused, % (n) 24 (92) 3 (16) <0.0001 Units 1 [1-2] 1 [1- 1] 0.05

Data expressed as median [interquartile ranges] or mean (standard deviation). ED = emergency department.

Muller.indd 83 25-7-2014 14:40:23

Chapter 4

84

Figure 1: ROTEM measurements in trauma patients at admission and 24 hours after admis-sion. Profiles classified as hyper-, normo- or hypercoagulable according to G value.

Hypercoagulable G>11.7 dynes/cm2 ; normocoagulable G=5-11.7 dynes/cm2 ; hypocoagulable G<5 dynes/cm2

Muller.indd 84 25-7-2014 14:40:23


85

Figu

re 2: R

OTE

M m

easu

rem

ents

in tr

aum

a pa

tient

s tra

nsfu

sed

with

RBC

and

FFP

at a

dmis

sion

and

24

hour

s afte

r adm

issi

on. P

rofil

es cl

assi

fied

as h

yper

-, no

rmo-

or h

yper

coag

ulab

le a

ccor

ding

to G

val

ue.

Hyp

erco

agul

able

G>1

1.7

dyne

s/cm

2 ; n

orm

ocoa

gula

ble

G=5

-11.

7 dy

nes/

cm2

; hyp

ocoa

gula

ble

G<5

dyn

es/c

m2

Muller.indd 85 25-7-2014 14:40:24

Muller.indd 86 25-7-2014 14:40:24

Part IIEffi cacy of plasma transfusion

in criti cally ill pati ents

Muller.indd 87 25-7-2014 14:40:24

Muller.indd 88 25-7-2014 14:40:24

Chapter 5

Transfusion of fresh-frozen plasma in critically

ill patients with a coagulopathy before invasive

procedures: a randomized clinical trial

Marcella C. Müller, M. Sesmu Arbous, Angelique M. Spoelstra – de Man, Roel Vink, Atilla Karakus, Marleen Straat, Jan M. Binnekade, Evert de Jonge,

Margreeth B. Vroom, Nicole P. Juffermans

Transfusion 2014; DOI 10.1111/trf.12750

Muller.indd 89 25-7-2014 14:40:24

Chapter 5

90

Abstract

Background: Prophylactic use of fresh-frozen plasma (FFP) is common practice in patients with a coagulopathy undergoing an invasive procedure. Evidence that FFP prevents bleeding is lacking, while risks of transfusion-related morbidity after FFP have been well demonstrated. We aimed to assess whether omitting prophylac-tic FFP transfusion in non-bleeding critically ill patients with a coagulopathy who undergo an intervention is non-inferior to a prophylactic transfusion of FFP.Study design and methods: A multicenter randomized open-label trial with blinded endpoint evaluation was performed in critically ill patients with an increased Inter-national Normalized Ratio (INR; 1.5-3.0). Patients undergoing placement of a central venous catheter, percutaneous tracheostomy, chest tube, or abscess drainage were eligible. Patients with clinically overt bleeding, thrombocytopenia, or therapeutic use of anticoagulants were excluded. Patients were randomly assigned to omitting or administering a prophylactic transfusion of FFP (12 ml/kg). Outcomes were occur-rence of postprocedural bleeding complications, INR correction, and occurrence of lung injury.Results: Due to slow inclusion, the trial was stopped before the predefined target enrollment was reached. Eighty-one patients were randomly assigned, 40 to FFP and 41 to no FFP transfusion. Incidence of bleeding did not differ between groups, with a total of one major and 13 minor bleedings (p=0.08 for noninferiority). FFP transfu-sion resulted in a reduction of INR to less than 1.5 in 54% of transfused patients. No differences in lung injury scores were observed.Conclusion: In critically ill patients undergoing an invasive procedure, no difference in bleeding complications was found regardless whether FFP was prophylactically administered or not.

Muller.indd 90 25-7-2014 14:40:24

FFP transfusion in critically ill patients before invasive procedures

91

Introduction

Fresh-frozen plasma (FFP) is effective in correcting multiple clotting factor deficien-cies, and transfusion guidelines recommend its use during massive bleeding and such a deficiency [1, 2]. In the past decades, use of FFP has grown steadily [3,4]. In addition to the use of FFP in actively bleeding patients, a substantial amount of FFP is transfused prophylactically to nonbleeding patients with a coagulopathy [5-9]. Coagulopathy, defined as prolonged prothrombin time (PT) or International Normal-ized Ratio (INR), has a high prevalence in critically ill patients [5]. Thereby, substan-tial amounts of FFP are utilized in the intensive care unit (ICU) [8-10].An important reason for physicians to transfuse FFP is the prevention of bleeding complications in nonbleeding patients with a coagulopathy, in particular in those undergoing an invasive procedure [8,11]. However, currently used diagnostic tests for coagulopathy such as INR and PT poorly represent in vivo hemostatic potential. Therefore, PT has limited value in predicting bleeding risk [12,13]. In addition, retro-spective studies suggest that commonly performed invasive procedures in the criti-cally ill, such as central venous catheter placement, percutaneous tracheostomy, and thoracocentesis, carry a low bleeding risk [13-16].Evidence from randomized controlled trials that support FFP transfusion to correct coagulopathy in order to reduce risk of bleeding before an invasive procedure is lim-ited [17,18]. On the other hand, FFP transfusion is associated with acute lung injury, which may occur in up to 30% of transfused critically ill patients [19], prolonging mechanical ventilation and ICU length of stay [20]. Also, an increased infection risk has been reported following FFP transfusion [21]. Thereby, the risk-benefit balance of prophylactic FFP may be limited. Despite the absence of evidence of a benefit, prophylactic FFP transfusion is common practice in the ICU [8,10,22,23]. We con-ducted a multi-center randomized controlled clinical trial in critically ill patients with a coagulopathy who needed to undergo an invasive procedure. Using noninferiority analysis, we aimed to determine whether FFP transfusion could be safely omitted in these patients.

Muller.indd 91 25-7-2014 14:40:24

Chapter 5

92


The Institutional Review Board of the Academic Medical Center - University of Amsterdam, Amsterdam, The Netherlands, approved the study protocol. Before entry in the study, written informed consent was obtained from the patient or legal representative in accordance with the Declaration of Helsinki. The study protocol was registered with trial identification numbers NTR2262 and NCT01143909 [24].

Setting and patientsConsecutive patients of 18 years and older admitted to the intensive care with an INR of at least 1.5 and not more than 3.0 and the need to undergo an invasive pro-cedure were considered eligible. Defined invasive procedures were insertion of a central venous catheter, thoracocentesis, percutaneous tracheotomy, drainage of abscess, or fluid collection. Patients with clinically overt bleeding (defined as either a decrease in hemoglobin (Hb) >1.6 g/dL or a need for transfusion or hemodynamic instability due to bleeding at the time of the procedure) or with a thrombocytopenia of not more than 30 x 109/L were excluded from participation. Also, patients were excluded when treated with vitamin K antagonists, activated protein C, abciximab, tirofiban, ticlopidine or prothrombin complex concentrates. In addition, patients with a history of congenital or acquired coagulation factor deficiency or bleed-ing diathesis were excluded. Patients treated with low molecular weight heparin (LMWH) or heparin in therapeutic dose were eligible if medication was discontinued for an appropriate period. Patients were enrolled at four sites in the Netherlands: two university hospitals (Academic Medical Center, Amsterdam and Leiden Univer-sity Medical Center, Leiden) and two large teaching hospitals (Tergooiziekenhuizen, Hilversum and Diakonessenhuis, Utrecht).Clinical practices, such as using ultrasound guidance when inserting a central venous catheter or a chest tube, were applied according to each center’s specific expertise and routine, to minimize interference of the trial intervention with normal clinical practice. Tracheostomy was performed percutaneously under direct sight using bronchoscopy. Use of low molecular weight heparin in a prophylactic dose was standard care in all patients. Selective digestive tract decontamination was part of standard care in three centers. The decision to transfuse red blood cells (RBCs), FFP

Muller.indd 92 25-7-2014 14:40:24


93

or platelets (PLTs) in case of postprocedural bleeding was made by the treating phy-sician in accordance with national guidelines.

Study designSince manufacturing a completely matched placebo in full compliance with the cur-rent good manufacturing practice standards was considered not possible, a pro-spective, randomized, open-label, blinded endpoint evaluation (PROBE) design was chosen. Patients admitted to the ICU were prospectively screened between May 2010 and June 2013 for INR prolongation. Patients fulfilling inclusion criteria were randomly assigned to receive or not to receive a single dose of 12 ml/kg FFP.The randomization procedure was Web-based, using permuted blocks and was stratified by study center and type of invasive procedure. Patients could only be randomized once (e.g., for one procedure). Fresh-frozen quarantine plasma, manu-factured by Sanquin, the Dutch National Bloodbank, was used. Transfusion amount was rounded to whole units. After randomization and transfusion or not, the sched-uled intervention was carried out and patients were observed for 24 hours.

Study end pointsThe primary outcome of the study was procedure-related bleeding, occurring within 24 hours after the procedure. Bleeding was assessed using a tool validated in the critically ill (HEME)[25]. Major bleeding was defined as bleeding accompanied by any of the following: a decrease in Hb by more than 2 g/dL in the absence of another cause of bleeding, transfusion of 2 or more units RBCs without an increase in Hb, a decrease in systolic blood pressure by more than 20 mmHg, an increase in heart rate by 20 beats/min or more, or wound-related bleeding requiring an intervention. Minor bleeding was defined as prolonged bleeding at the site of insertion or increase in size of subcutaneous hematoma. The potential bleeding site was assessed by a physician blinded to the intervention who filled out a predefined bleeding score form consist-ing of blood pressure, heart rate, Hb level, and occurrence of procedure-related bleeding with or without the need for intervention or transfusion. Subsequently this blinded physician assigned a score of major bleeding, minor bleeding or no bleeding at 1 and 24 hours after the intervention. The effects of FFP on correction of INR and additional transfusion requirements were assessed. Development of lung injury was

Muller.indd 93 25-7-2014 14:40:24

Chapter 5

94

assessed by calculating the lung injury score 24 hours after the intervention, which includes radiographic infiltrates on chest X-ray, a hypoxemia score (PaO2/FiO2), positive end expiratory pressure level (cm H2O), and respiratory system compliance score when available [26]. Ventilator associated pneumonia (VAP) was defined as newly diagnosed (infiltrate on chest X-ray and clinical signs of pneumonia), culture proven pneumonia, while on mechanical ventilation.

Statistical considerationsThe sample size and power calculation of this study was based on studies that showed that the occurrence of major bleeding in patients with a coagulopathy undergo-ing invasive procedures was less than 1%. Group size calculation was focused on demonstrating noninferiority. With a sample size in each group of 198, a one-sided Z test with continuity correction (pooled) achieved 80% power to reject the null hypothesis that the proportion of bleeding patients in the experimental group (no FFP transfusion) was higher, that is, inferior to the proportion in the control group (FFP transfusion) with a margin of 0.03. It was assumed that the expected difference in proportions is zero and the proportion in the control group is 0.01. The one-sided significance level of the test was targeted at 0.05. Therefore, we intended to enroll 200 patients per treatment arm [24].

Statistical analysisVariables are expressed by their mean and standard deviation if normally distrib-uted. If not normally distributed they are expressed by their medians and interquar-tile ranges. Categorical variables are expressed as number (%). The primary endpoint of procedure-related major bleeding only occurred once, rendering planned non-inferiority analysis impossible. Therefore, a post hoc noninferiority analysis for all types of bleeding complications (major and minor) was performed according to the method of Dunnett and Gent [27]. The null hypothesis stated that the difference between the control treatment (FFP transfusion) and the intervention (no FFP trans-fusion) equals a delta of 0.03. The alternative hypothesis stated that the difference between the control treatment and the new treatment is smaller than a delta of 0.03. A significant result (p≤0.05) would indicate that omitting FFP transfusion is noninferior to administering FFP before an intervention. In addition, occurrence of

Muller.indd 94 25-7-2014 14:40:24


95

bleeding between both groups was compared by Fisher exact test. Because of the noninferiority design of the trial, both an intention-to-treat analysis and a per-pro-tocol analysis was performed and results of the per-protocol analysis are reported.The secondary outcome variables were analyzed using the two-group t-test or Mann-Whitney test, simple linear regression and chi-square test, when appropriate. In case of paired data the paired t-test was used, or in case of not normally distrib-uted data the Wilcoxon signed rank test was used. Correlations were determined by Spearman’s Rho. In all analyses statistical uncertainties are expressed in 95% confi-dence limits, with a significance level of 0.05. Statistical analyses were carried out with R statistical computing (Vienna, Austria, http://www.R-project.org).

Results

Patients and interventionsOwing to slow inclusion, the trial was stopped before the predefined target enroll-ment was reached. Between May 2010 and June 2013, a total of 13,163 INR values were screened in 6825 patients. A total of 1478 patients had an INR of at least 1.5 and not more than 3.0. Of these, 615 patients did not fulfill inclusion criteria, leaving 263 patients with an INR of at least 1.5 and not more than 3.0 scheduled to undergo a predefined intervention. Of these, 65 patients declined informed consent. An addi-tional 83 patients were missed and 34 patients did not participate due to other rea-sons, including refusal from treating physicians to include a specific patient (3.8%). This left 81 coagulopathic patients who underwent randomization (figure 1). Five patients did not undergo an intervention despite randomization and were there-fore excluded from further analysis. The results of the per-protocol analysis are reported. Half of the included patients had sepsis. Prevalence of disseminated intra-vascular coagulation and liver disease was high (table 1). Groups were balanced with respect to demographic variables and disease severity scores. However, in patients randomly assigned to the nontransfused group, liver disease was more frequent (p=0.006). Baseline values of coagulation tests, Hb, and lung injury scores did not differ between the two groups, nor did preprocedural use of anticoagulant medica-tion (table 1). Central catheter placement was the most frequent intervention (76% in the FFP transfusion and 78% in the nontransfused group; table 2).

Muller.indd 95 25-7-2014 14:40:24

Chapter 5

96

Figure 1: CONSORT diagram.

Number of screened pa�ents and INR’sPa�ents N=6825

INRs N=13163

Number of INRs ≥ 1.5 en ≤ 3.0N=3499 in 1478 pa�ents

Number of pa�ents withexclusion criterium:

N=615

Number of poten�al eligible pa�ents:N=863

Eligible number of pa�ents undergoinginterven�on:

N=263

No par�cipa�on due to:No informed consent N=65

Missed N=83Physician declines par�cipa�on N=10

Other N=24

Underwent randomiza�onN=81

Assigned to FFPN=40

Assigned to no FFPN=41

Included in analysisN=38

Included in analysisN=38

No interven�on N=3No interven�on N=2

Muller.indd 96 25-7-2014 14:40:24


97

Table 1: Patient characteristics

FFP transfusion(n=38)

No FFP transfusion(n=38)

p value

General characteristicsGender, male 67 (26) 47 (18) 0.09Age (years) 64 (54-70) 66 (62-72) 0.23BMI 24 (22-28) 23 (20-26) 0.15Days admission until intervention 3 (1-8) 2 (1-3) 0.26APACHE IV score 107 (80-129) 101 (83-126) 0.76SOFA score 12 (10-14) 12 (10-15) 0.81Medical historyPulmonary disease 13 (5) 10 (4) 0.72Liver disease 16 (6) 45 (17) 0.006Cardiac failure 16 (6) 16 (6) 1.00Medical condition 24 hours before interventionMechanical ventilation 84 (32) 76 (29) 0.52Sepsis 47 (18) 47(18) 1.00Pneumonia 26 (10) 24 (9) 0.89Disseminated intravascular coagulation 45 (17) 38 (14) 0.54Baseline measurementsINR 1.8 (1.5-2.2) 1.9 (1.6-2.2) 0.29aPTT (sec) 43 (38-52) 41 (36-49) 0.35PLT count (x109/L) 92 (52-180) 110 (52-183) 0.94Haemoglobin (g/dL) 9.3 (8.4-10.2) 9.3 (8.4-10.6) 0.79Lung Injury Score 2 (1-2.7) 1.3 (0.83-2.3) 0.61AnticoagulationAspirin 10 (4) 24 (9) 0.11Clopidogrel 3 (1) 5 (2) 0.54Aspirin and clopidogrel 0 (0) 5 (2) 0.49UFH therapeutic dose 10 (4) 18 (7) 0.31LMWH therapeutic dose 23 (9) 11 (4) 0.14

Data are expressed as percent (number) or median (interquartile range).BMI = Body Mass IndexAPACHE = Acute Physiology and Chronic Health EvaluationSOFA = Sequential Organ Failure AssessmentINR = International Normalized RatioaPTT = Activated Partial Thromboplastin TimePLT = plateletUFH = Unfractionated heparinLMWH = Low-molecular-weight heparin

Muller.indd 97 25-7-2014 14:40:24

Chapter 5

98

Table 2: Intervention and transfusion characteristics in randomized patients.



p value

InterventionCentral venous catheter 76 (29) 76 (29) 0.58Chest tube 11 (4) 8 (3)Tracheotomy 5 (2) 5 (2)Abdominal drain 8 (3) 11 (4)FFP before to interventionUnits 3 (2-4) 0Dose (ml/kg) 12 (10-13) 0Transfusions in first 24 hr after intervention RBCs (units) 1 (0-2) 1 (0-3) 0.91FFP (units) 0 (0-1) 2 (0-2) 0.06PLTs (units) 1 (0-2) 0 (0-1) 0.43

Data expressed as percent (number) or median (interquartile range).

Table 3: Bleeding rates in randomized patients.



Fatal bleeding 0 0Major bleeding 0 1Minor bleeding 8 5No bleeding 30 32

Data are expressed as number.

Occurrence of bleedingNone of the patients experienced a fatal bleeding. Overall, one major and 13 minor bleedings occurred (table 3). The major bleeding complication was a hematothorax after ultrasound-guided insertion of a chest drainage system (Pleural Catheter, RM Temena GmbH, Felsberg, Germany) in a patient randomly assigned not to receive FFP transfusion. The patient was treated with 100 mg of aspirin once daily and unfractionated heparin in a dose of 500 IU/hr. The heparin was discontinued more than 3 hours before the procedure. Coagulation tests 1 hour before the interven-tion showed a PT of 18.9 seconds, an INR of 1.71, activated partial thromboplastin time (aPTT) 47 seconds, fibrinogen 2.3 g/L and a PLT count of 34*109/L. Within the first hour after insertion of the chest drainage system 300 ml of blood ran from the

Muller.indd 98 25-7-2014 14:40:24


99

drain. In response to this complication the patient was transfused with 3 units of RBCs, 3 units of FFP and 1 unit of PLTs, which resulted in cessation of the bleeding and stabilization of the patient.As the primary outcome of major bleeding only occurred once, meaningful statistical analysis was not possible. Therefore, bleeding data (major and minor bleeding com-plications) were aggregated into bleeding and no bleeding categories. Occurrence of bleeding did not differ between groups (eight events in the FFP group compared to six in the nontransfused group, p=0.77; table 3). However, the observed differ-ence in occurrence of bleeding was 4.8% to the detriment of the group transfused with FFP, while a difference of 3% to the detriment of the nontransfused group was allowed. Noninferiority analysis generated a chi-square of 2.05 (df 1, p=0.08 [one sided]) 1 hour after the intervention and a chi-square of less than 0.01 (df 1, p=0.50 [one sided]) 24 hours after the intervention.Preprocedural omission of FFP was not associated with increased occurrence of bleeding (relative risk 1.17; 95% confidence interval [CI] 0.62-2.19; p=0.78). Prepro-cedural use of anticoagulation (PLT inhibitors, therapeutic use of heparin or low-molecular-weight heparin) was not associated with increased occurrence of bleeding (relative risk 1.47; 95%CI 0.55-3.92; p=0.44; figure 2). Patients randomly assigned to FFP transfusion received a median of 3 FFP units [IQR 2-4]. FFP transfusion before the procedure did not affect transfusion requirements in the first 24 hours after the procedure (table 2).

Correction of INRFFP transfusion resulted in a median reduction of INR from 1.8 (IQR 1.5-2.5) to 1.4 (IQR 1.3-1.63; p<0.001). However, only 54% of patients had a corrected INR to less than 1.5 after FFP transfusion (figure 3). The effect of FFP on INR reduction var-ied widely. Patients with higher pretransfusion INR values experienced the greatest reduction after FFP transfusion (r =0.68, p<0.01; figures 3 and 4).

Muller.indd 99 25-7-2014 14:40:24

Chapter 5

100

Figure 2: Occurrence of bleeding complications related to randomization group and use of anticoagulant medication before randomization.

Grey dots indicate patients without FFP transfusion before intervention, black dots indicate patients transfused with 12 ml/kg FFP before intervention. Encircled dots indicate patients who received antiplatelet therapy less than 3 days before the intervention.

Figure 3: Effect of FFP transfusion (12 ml/kg) on correction of INR in 38 patients randomized to the transfusion group.

The grey line indicates an INR value of 1.5.

Muller.indd 100 25-7-2014 14:40:25


101

Figure 4: Correlation of INR value before FFP transfusion and magnitude of effect of FFP transfusion.

Table 4: Clinical outcomes 24 hours after randomization.



p value

Lung injury score 2 (0.8-2.5) 1.25 (0.4-2.4) 0.28P/F ratio 203 (143-308) 248 (180-308) 0.51Ventilator associated pneumonia 8 (3) 5 (1) 0.001Mechanical ventilation days 11 (6-16) 3 (2-12) 0.01SOFA 24 hours after intervention 12 (9-16) 11 (9-14) 0.68ICU length of stay 12 (6-19) 7 (3-17) 0.13Overall mortality 51 (19) 71 (27) 0.08

Data expressed as percent (number) or median (interquartile range)P/F = pO2 (mmHg)/fraction of inspired oxygenICU = intensive care unitSOFA = Sequential Organ Failure Assessment

Muller.indd 101 25-7-2014 14:40:25

Chapter 5

102

Clinical outcomeAt baseline, lung injury scores were increased in both groups, with values indicat-ing moderate to severe lung injury, consistent with underlying morbidity of these patients. After the intervention, lung injury score and oxygenation score tended to be lower in the FFP group, but this did not reach statistical significance. Degree of change in lung injury score after 24 hours did not differ between groups. Duration of mechanical ventilation was significantly longer in patients transfused with FFP com-pared to nontransfused patients (p=0.01). Additional linear regression identified the presence of sepsis and liver disease as important contributors to prolonged dura-tion of mechanical ventilation. Furthermore, the incidence of VAP was significantly higher in the FFP group patients.ICU length of stay did not differ between both groups. Mortality tended to be higher in the nontransfused group (51% mortality in patients transfused with FFP compared to 71% in those not transfused with FFP before the intervention, p=0.08; table 4). Additional linear regression demonstrated that liver disease was the sole predictor for mortality (p=0.056). Hereby, liver disease was a confounder for mortality.

Discussion

In this multicenter randomized controlled trial, occurrence of bleeding complica-tions after an invasive procedure in critically ill patients with a coagulopathy did not differ between patients with or without a prophylactic FFP transfusion. Due to lim-ited inclusion, noninferiority of omitting FFP transfusion could not be demonstrated. Post hoc analysis showed a trend towards noninferiority with a 5% higher bleeding rate in the transfused group compared to nontransfused patients. FFP transfusion was associated with prolonged duration of mechanical ventilation and increased VAP rates.Our study is the first randomized controlled trial on the efficacy of prophylactic FFP transfusion to prevent bleeding in coagulopathic patients undergoing an invasive procedure. Previous studies on prophylactic FFP transfusion reported the effect of FFP on laboratory variables but not on bleeding [12,28]. In this trial, only one major bleeding occurred. Minor bleeding occurred in up to 17% of patients and mostly consisted of prolonged oozing of insertion site or hematoma formation, not

Muller.indd 102 25-7-2014 14:40:25


103

requiring any specific therapy or transfusion. Occurrence of minor bleeding is of limited value for clinical practice and most likely clinicians administer FFP in order to prevent major bleeding complications. From our data is it is not possible to deter-mine whether a lack of difference in occurrence of minor bleeding complications also has predictive value for the occurrence of major bleeding complications. How-ever, retrospective studies also showed that the incidence of major bleeding after an invasive procedure in patients with a prolonged INR is less then 1% [29-31]. Low incidence of bleeding has been reported in observational studies in coagulopathic patients receiving central venous catheter placement [32,33] and percutaneous tra-cheotomy [15] and after thoracocentesis [14,34]. Also, procedure-related bleeding complications were not influenced by the administration of FFP [14,16]. Our data are in line with previous reported observations that invasive procedures can be safely carried out in critically ill patients with a coagulopathy.Bleeding is a subjective endpoint. We could not blind the transfusion bags and man-ufacturing a placebo was not considered feasible. Therefore, we have chosen for an observer blinded to intervention safeguarding objectivity. The endpoint evaluation in a PROBE study is as reliable as a double-blind study, provided that the same crite-ria are applied [35]. In addition, it is shown that results from double-blind, placebo-controlled and PROBE trials are statistically comparable [36]. The HEME tool used to assess procedure-related relevant bleeding in our trial has been validated for ICU patients, with a high inter- rater agreement [25]. Another frequently used tool, the World Health Organization (WHO) bleeding scale, which assesses the presence of petechiae or mucosal bleeding, is validated for patients with cancer but not in the critically ill [37]. A possible disadvantage of the HEME tool is that some items, such as decrease in systolic blood pressure or increase in heart rate, may also occur in the absence of bleeding. However, assessors were asked to consider such physiologic changes only if they occurred in the absence of other causes [25].Although FFP transfusion resulted in a reduction of INR in all patients, only half cor-rected to an INR of less than 1.5, a value which is a reported trigger for FFP transfu-sion in clinical practice [8]. In line, most studies show a reduction and not a complete normalization of INR after FFP transfusion [10,12,19,38]. This is probably due to low FFP doses or attempts to correct minimally elevated INR, which is not possible [28,39]. It can be argued that no effect of FFP transfusion was seen in our study due

Muller.indd 103 25-7-2014 14:40:25

Chapter 5

104

to a low dose. Studies on the dose of FFP to correct INR in ICU patients show conflict-ing results. In a small cohort of 22 patients, a dose of 33 ml/kg was more effective in achieving target levels of coagulation factors compared to 12 ml/kg [12]. However, in a larger trial in ICU patients, 20 ml/kg did not differ from 12 ml/kg in correction of INR or increasing coagulation factors above a minimum hemostatic threshold [38]. Of note, audits on FFP transfusion show that the dose of prophylactic FFP transfu-sion used in clinical practice is often less than 10 ml/kg [9,40]. The strength of this study is the use of a fixed dose, which is in line with doses used in common clinical practice [7,8,10,40,41]. Although we cannot exclude that higher doses of FFP may prevent bleeding, we think that underdosing of FFP is unlikely to have contributed to the absence of a difference in bleeding complications between both groups, because there was no trend towards efficacy of FFP.In line with previous reports, the effect of FFP transfusion on INR decrease corre-lated with pretransfusion INR [12,28]. In the current study, the lower INR boundary was chosen in line with clinical practice, in which an INR of at least 1.5 is frequently reported to be a transfusion trigger [6,8], in particular when an intervention is scheduled [40]. We choose an upper limit of an INR of not more than 3.0 to prevent noncompliance by participating physicians [11].In recent years, many reports have been published on the adverse effects of FFP. FFP transfusion is an independent risk factor for new onset of acute lung injury [42-45], in particular in patients who are mechanically ventilated [46]. Furthermore, studies consistently report an association between FFP and increased length of ICU stay [6,45,46]. We did not confirm an association between FFP and lung injury, probably due to low numbers of patients. Of note, oxygenation tended to be worse and dura-tion of mechanical ventilation was prolonged in patients transfused with FFP, asso-ciated with an increased incidence of VAP in the transfused group. This observation is consistent with reports in trauma and nontrauma patients, where transfusion of FFP has been shown to be independently associated with infection and pneumonia [21,47]. However, in our cohort duration of mechanical ventilation was affected by the presence of sepsis or liver disease, hereby impeding conclusions about the asso-ciation between FFP and duration of mechanical ventilation. Whether FFP transfu-sion indeed can induce transfusion-related immunomodulation is, however under debate [48,49].

Muller.indd 104 25-7-2014 14:40:25


105

The major limitation of our trial is that we stopped early due to slow inclusion. Despite the addition of extra recruitment sites, we were only able to randomize 20% of the targeted patient number. Several factors contributed to this disappoint-ing inclusion rate. First, there is a small window of opportunity to include patients. Placement of central venous catheter may be an urgent procedure in critically ill patients, which does not allow postponing the intervention until informed consent was obtained. This accounted for a 20% loss of potentially eligible patients. Second, despite that certified ICU staff physicians asked informed consent, denial rate was high. Third, treating physicians were at times reluctant to randomize patients for the trial. The reasons were bidirectional: physicians who wanted to correct coagulopa-thy before the invasive procedure did not want to take the risk of randomization to no FFP, and on the other hand physicians who felt no need for FFP transfusion did not want to risk having to administer FFP. Therefore, despite the expressed need for trials on the use of FFP in critically ill patients with a coagulopathy, preferences of treating physicians about the use of FFP were major constraints to the conduct a successful clinical trial. We feel that future trials on the efficacy of FFP may not be feasible to conduct, at least in the Netherlands, and that further knowledge on the efficacy of FFP should be gathered from prospective cohort studies or retrospective studies.Another limitation of our study is the randomization imbalance resulting in a higher proportion of patients with liver disease in the no FFP group. As liver disease influ-enced mortality, it also affected duration of mechanical ventilation. Of note, the combined outcome of major and minor bleeding was assessed within 24 hours; therefore, these outcomes are less likely to be affected by the presence or absence of liver disease. Despite disappointing inclusion rates, our study is the largest ran-domized trial to date on the effectiveness of FFP to prevent bleeding in critically ill patients with a coagulopathy and we believe results are of interest to any physician involved in performing invasive procedures.

Muller.indd 105 25-7-2014 14:40:25

Chapter 5

106

Conclusion

The effect of FFP on major bleeding complications could not be assessed due to low inclusion rate and low incidence of major bleeding complications in critically ill patients with a coagulopathy undergoing an invasive procedure. The combined outcome of major and minor bleeding complications did not differ between patients with or without FFP transfusion. A dose of 12 ml/kg of FFP only partially corrected INR.

Muller.indd 106 25-7-2014 14:40:25


107

References


2. De Backer D, Vandekerckhove B, Stanworth S, et al.: Guidelines for the use of fresh frozen plasma. Acta Clin Belg 2008;63: 381-90

3. Wallis JP, Dzik S: Is fresh frozen plasma overtransfused in the United States? Transfusion 2004;44: 1674-5

4. Serious Hazards of Transfusion Steering Committee: Serious Hazards of Transfusion: annual report 2012. [cited 2013 Nov 5] Available from: http://www.shotuk.org


6. Lauzier F, Cook D, Griffith L, et al.: Fresh frozen plasma transfusion in critically ill patients. Crit Care Med 2007;35: 1655-9




10. Abdel-Wahab OI, Healy B, Dzik WH: Effect of fresh-frozen plasma transfusion on pro-thrombin time and bleeding in patients with mild coagulation abnormalities. Transfu-sion 2006;46: 1279-85




14. Hibbert RM, Atwell TD, Lekah A, et al.: Safety of ultrasound-guided thoracentesis in patients with abnormal preprocedural coagulation parameters. Chest 2013;144: 456-63

15. Rosseland LA, Laake JH, Stubhaug A: Percutaneous dilatational tracheotomy in inten-sive care unit patients with increased bleeding risk or obesity. A prospective analysis of 1000 procedures. Acta Anaesthesiol Scand 2011;55: 835-41

16. Carino GP, Tsapenko AV, Sweeney JD: Central line placement in patients with and with-out prophylactic plasma. J Crit Care 2012;27: 529 e9-13

Muller.indd 107 25-7-2014 14:40:25

Chapter 5

108


18. Yang L, Stanworth S, Hopewell S, et al.: Is fresh-frozen plasma clinically effective? An update of a systematic review of randomized controlled trials. Transfusion 2012;52: 1673-86




22. Westbrook A, Pettila V, Nichol A, et al.: Transfusion practice and guidelines in Australian and New Zealand intensive care units. Intensive Care Med 2010;36: 1138-46


24. Muller MC, de Jonge E, Arbous MS, et al.: Transfusion of fresh frozen plasma in non-bleeding ICU patients--TOPIC trial: study protocol for a randomized controlled trial. Tri-als 2011;12: 266

25. Arnold DM, Donahoe L, Clarke FJ, et al.: Bleeding during critical illness: a prospective cohort study using a new measurement tool. Clin Invest Med 2007;30: E93-102

26. Murray JF, Matthay MA, Luce JM, et al.: An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988;138: 720-3

27. Dunnett CW, Gent M: Significance testing to establish equivalence between treatments, with special reference to data in the form of 2X2 tables. Biometrics 1977;33: 593-602


29. Mumtaz H, Williams V, Hauer-Jensen M, et al.: Central venous catheter placement in patients with disorders of hemostasis. Am J Surg 2000;180: 503-5

30. Fisher NC, Mutimer DJ: Central venous cannulation in patients with liver disease and coagulopathy--a prospective audit. Intensive Care Med 1999;25: 481-5

31. Goldfarb G, Lebrec D: Percutaneous cannulation of the internal jugular vein in patients with coagulopathies: an experience based on 1,000 attempts. Anesthesiology 1982;56: 321-3

32. Doerfler ME, Kaufman B, Goldenberg AS: Central venous catheter placement in patients with disorders of hemostasis. Chest 1996;110: 185-8

33. Foster PF, Moore LR, Sankary HN, et al.: Central venous catheterization in patients with coagulopathy. Arch Surg 1992;127: 273-5

34. McVay PA, Toy PT: Lack of increased bleeding after paracentesis and thoracentesis in patients with mild coagulation abnormalities. Transfusion 1991;31: 164-71

Muller.indd 108 25-7-2014 14:40:25


109

35. Hansson L, Hedner T, Dahlof B: Prospective randomized open blinded end-point (PROBE) study. A novel design for intervention trials. Prospective Randomized Open Blinded End-Point. Blood Press 1992;1: 113-9

36. Smith DH, Neutel JM, Lacourciere Y, et al.: Prospective, randomized, open-label, blinded-endpoint (PROBE) designed trials yield the same results as double-blind, placebo-controlled trials with respect to ABPM measurements. J Hypertens 2003;21: 1291-8

37. Miller AB, Hoogstraten B, Staquet M, et al.: Reporting results of cancer treatment. Can-cer 1981;47: 207-14


39. Holland LL, Foster TM, Marlar RA, et al.: Fresh frozen plasma is ineffective for correcting minimally elevated international normalized ratios. Transfusion 2005;45: 1234-5

40. Hall DP, Lone NI, Watson DM, et al.: Factors associated with prophylactic plasma trans-fusion before vascular catheterization in non-bleeding critically ill adults with pro-longed prothrombin time: a case-control study. Br J Anaesth 2012;109: 919-27




44. Wright SE, Snowden CP, Athey SC, et al.: Acute lung injury after ruptured abdominal aortic aneurysm repair: the effect of excluding donations from females from the pro-duction of fresh frozen plasma. Crit Care Med 2008;36: 1796-802



47. Bochicchio GV, Napolitano L, Joshi M, et al.: Blood product transfusion and ventilator-associated pneumonia in trauma patients. Surg Infect 2008;9: 415-22

48. Vamvakas EC: Pneumonia as a complication of blood product transfusion in the criti-cally ill: transfusion-related immunomodulation (TRIM). Crit Care Med 2006;34: S151-9

49. Vamvakas EC, Carven JH: Exposure to allogeneic plasma and risk of postoperative pneu-monia and/or wound infection in coronary artery bypass graft surgery. Transfusion 2002;42: 107-13

Muller.indd 109 25-7-2014 14:40:25

Muller.indd 110 25-7-2014 14:40:25

Chapter 6

Fresh frozen plasma transfusion fails to alter the hemostatic

balance in critically ill patients with a coagulopathy

Marcella C.A. Müller, Marleen Straat, Joost C.M. Meijers, Evert de Jonge, M. Sesmu Arbous, Marcus J. Schultz, Margreeth B. Vroom, Nicole P. Juffermans


Muller.indd 111 25-7-2014 14:40:26

Chapter 6

112

Abstract

Coagulopathy has a high prevalence in critically ill patients. A prolonged INR is a com-mon trigger to transfuse fresh frozen plasma (FFP), even in the absence of bleeding. Thereby, FFP is frequently administered in these patients. However, efficacy of FFP to correct hemostatic disorders in non-bleeding recipients is questioned. We aimed to assess the effect of a fixed dose of FFP transfusion on the hemostatic balance in non-bleeding critically ill patients with a coagulopathy. Markers of coagulation, indi-vidual factor levels and levels of natural anticoagulants were measured and throm-bin generation assays were performed before and after FFP transfusion (12 ml/kg) to 38 non-bleeding critically ill patients with a prolonged INR (1.5-3.0). At baseline, levels of procoagulant factor II, V and VII as well as levels of anticoagulants protein C, S and antithrombin were below normal values, associated with impaired thrombin generation. FFP transfusion increased both levels of individual coagulation factors as well as levels of anticoagulant proteins. Thrombin generation was not affected by FFP transfusion. In conclusion, in critically ill patients with a coagulopathy, FFP transfusion does not alter hemostatic balance or thrombin generation.

Muller.indd 112 25-7-2014 14:40:26

Effects of FFP on coagulation in the critically ill

113

Background

Coagulopathy occurs in up to 30% of critically ill patients [1]. Seventy percent of Fresh Frozen Plasma (FFP) used by critical care facilities is transfused to patients with prothrombin time (PT) prolongation, mostly in the absence of bleeding [2]. To assess bleeding risk and effectiveness of FFP transfusion, International Normalized Ratio (INR) is often used. An increased INR indicates that at least one of the vitamin K dependent coagulation factors is below the hemostatic threshold [3] and is a com-mon trigger to administer FFP [4]. However, in acquired coagulopathy in critically ill patients, where multiple clotting factors are deficient, the relationship between clotting factor levels and INR prolongation is complex and not well understood. Indeed, INR only represents a part of the coagulation cascade and is insensitive to the activity or concentration of natural anticoagulants and fibrinolytic proteins. In the critically ill, and particularly in patients with disseminated intravascular coagula-tion (DIC), levels of natural anticoagulants are reduced and fibrinolysis is attenuated [5,6].The hemostatic balance is the net result of the presence of pro-coagulant, anti-coagulant and fibrinolytic factors. Disturbance between these components can result in variable in vivo coagulation profiles, ranging from hypocoagulable with increased bleeding tendency to a pro-coagulant state with (micro-) vascular throm-bus formation. Thereby, despite an elevated INR, patients may not necessarily have an increased bleeding tendency. In line with this, INR poorly reflects risk of bleeding in the critically ill [7].The efficacy of FFP to correct coagulopathy has only been evaluated in small and heterogeneous studies [8-10]. Notably, the efficacy of FFP to prevent bleeding in coagulopathic critically ill patients has not been demonstrated [11]. Besides coagula-tion factors, FFP also contains natural occurring anticoagulants antithrombin, pro-tein C and S. The net effect of FFP transfusion on anticoagulant proteins and the hemostatic balance in the critically ill is not well known [12].In a predefined sub study of a multicenter randomized controlled trial on the effi-cacy of prophylactic FFP transfusion to prevent bleeding in critically ill patients with an increased INR who were scheduled to undergo an invasive procedure [13], we investigated whether INR prolongation parallels changes in other tests investigating

Muller.indd 113 25-7-2014 14:40:26

Chapter 6

114

hemostasis and evaluated the effect of a fixed dose of FFP on the hemostatic bal-ance of these patients.


The original study was approved by the medical ethics committee of the Academic Medical Center, Amsterdam, the Netherlands. Before entry in the study, written informed consent was obtained from the patient or legal representative in accord-ance with the Declaration of Helsinki. The study protocol was registered with trial identification numbers NTR 2262 and NCT01143909.

Setting and patientsThe study was performed between May 2010 and June 2013 in four mixed medical-surgical ICUs in the Netherlands. Patients were eligible when INR was 1.5-3.0 and needed to undergo an intervention. Patients younger than 18 years or with a known bleeding diathesis, treated with vitamin K antagonists, activated protein C, abcixi-mab, tirofiban, ticlopidine or prothrombin complex concentrates were excluded. Patients treated with therapeutic doses of heparin or low molecular weight hepa-rin were allowed to participate if medication was discontinued for an appropriate period, which was >2 hours for heparin and >12 hours for low molecular weight heparin. Use of low molecular weight heparin in a prophylactic dose was part of standard care in all patients.

DesignA predefined post hoc study of a randomized controlled clinical trial on risk and ben-efit of FFP transfusion in critically ill patients with a coagulopathy [13]. After inclu-sion, patients were randomized to a single dose of 12 ml/kg FFP or no FFP transfusion before a scheduled intervention. The FFP was quarantine plasma manufactured by Sanquin, the Dutch National Bloodbank. The dose of FFP was based on clinical prac-tice and a previous study with a target of INR reduction to <1.5 [9,14]. Patients were observed until 24 hours after the intervention for bleeding complications.

Muller.indd 114 25-7-2014 14:40:26


115

Patient data and sample collectionPatient data were collected from the electronic patient data management system (PDMS) and consisted of medical history, admission diagnosis, use of anticoagulant medication and occurrence of bleeding. Disseminated intravascular coagulation (DIC) was assessed using the ISTH DIC score, which defines overt DIC as a score of ≥5 points [15].Blood samples were drawn from an indwelling arterial catheter at baseline and the second sample was taken directly after FFP transfusion (but prior to the invasive procedure). Samples were collected in sodium citrate (0.109M 3.2%) tubes. Sam-ples were centrifuged within 30 minutes at 2000 x g for 15 mins at 18°C and subse-quently 5 mins at 15000 x g also at 18°C. Supernatant was collected and stored at -80°C until measurements were performed.

AssaysPT, INR, aPTT and levels of D-dimer and fibrinogen were all determined immedi-ately after the sample was drawn with standard assays on an automated coagula-tion analyzer (Sysmex CA 7000 and all reagents, Siemens Healthcare Diagnostics, Marburg, Germany) according to manufacturers protocols. After termination of the study the remaining assays were performed collectively in all patients. Levels of fac-tors II, V and VII were determined by a PT based one stage clotting assay (ACL TOP 700, Instrumentation Laboratory, USA), using recombiplastin and factor II, V and VII deficient plasma (Instrumentation Laboratory, USA). Antithrombin was assessed by chromogenic substrate method (Symex CA7000) with reagents and protocols of the manufacturer.Protein C activity was measured by a kinetic assay (Coamatic Protein C, Chromog-enix, Mölndal, Sweden). Total protein S levels were determined by enzyme-linked immunosorbent assay (ELISA) as described previously [16]. Free protein S was meas-ured by precipitating the C4b-binding protein-bound fraction with polyethylene gly-col 8000 and measuring the concentration of free protein S in the supernatant [16]. Thrombin-antithrombin (TATc), prothrombin fragment 1 + 2 (F1+2) and plasmin-α2-antiplasmin complex (PAP) levels were measured using specific commercially avail-able ELISAs according to the instruction of the manufacturer (Siemens Healthcare Diagnostics and DRG, Marburg, Germany).

Muller.indd 115 25-7-2014 14:40:26

Chapter 6

116

Thrombin generation assayThe Calibrated Automated Thrombogram® assays the generation of thrombin in clotting plasma using a microtiter plate reading fluorometer (Fluoroskan Ascent, ThermoLab systems, Helsinki, Finland) and Thrombinoscope® software (Thrombino-scope BV, Maastricht, The Netherlands). The assay was carried out as described by Hemker et al. [17] and the Thrombinoscope® manual. Coagulation was triggered by recalcification in the presence of 5 pM recombinant human tissue factor (Innovin®, Siemens Healthcare Diagnostics, Marburg, Germany), 4 μM phospholipids, and 417 μM fluorogenic substrate Z-Gly-Gly-Arg-AMC (Bachem, Bubendorf, Switzerland). Fluorescence was monitored, and the parameters (lag time, peak thrombin, area under the curve or endogenous thrombin potential) were calculated using the Thrombinoscope software.

StatisticsAll variables are expressed as mean (standard deviation) or median (interquartile ranges) depending on whether distribution was normal or not. To compare groups two group t-test or Mann-Whitney test was used. Paired data were compared using the Wilcoxon signed rank test. A p value of less than 0.05 was considered significant. Statistical analyses were performed with SPSS 20.0 (SPSS, Inc, Chicago, IL, USA) and Prism Version 5.0 (Graphpad Software, San Diego, USA).

Results

PatientsIn total 38 patients were randomized to FFP transfusion and all samples before and after transfusion were available for analysis. Patients were critically ill, as reflected by a high disease severity score (e.g. APACHE IV and SOFA score) (table 1). Half of the patients had sepsis and more than a third had DIC. A considerable number of patients received heparin or low molecular weight heparin treatment in a therapeu-tic dose, which was discontinued for at least an appropriate period prior to baseline measurements. In 8 of 38 patients, minor bleeding complications occurred in the first 24 hours after the intervention. None of these events required transfusion of extra FFP or an intervention to cease bleeding.

Muller.indd 116 25-7-2014 14:40:26


117


FFP transfusionn=38

General characteristicsGender, male, % (n) 67 (26)Age (years) 64 (54-70)APACHE IV score 107 (80-129)SOFA score 12 (10-14)Medical conditionLiver disease, % (n) 16 (6)Sepsis, % (n) 47 (18)DIC, % (n) 45 (17)AnticoagulationAspirin, % (n) 10 (4)Clopidogrel, % (n) 3 (1)Heparin, % (n)* 10 (4)Low Molecular Weight Heparin, % (n)* 23 (9)Transfusion FFP (units) 3 (2-4)OutcomeICU length of stay (days) 12 (6-19)28 day mortality, % (n) 51 (19)

Data expressed as median and interquartile ranges.* therapeutic doseAPACHE = Acute Physiology and Chronic Health EvaluationSOFA = Sequential Organ Failure AssessmentDIC = Disseminated Intravascular CoagulationFFP = Fresh Frozen PlasmaICU = Intensive Care Unit

Baseline coagulation tests in critically ill with a prolonged INRMedian INR levels at baseline were 1.8 [1.5-2.2] in all patients. As expected, median baseline levels of coagulation factors were all reduced, with a factor II of 34% [26-46], factor V of 48% [28-76] and VII level of 25% [16-38] (figure 1). Also, patients had decreased levels of endogenous anticoagulant factors, including antithrombin (47% [35-78]), protein C activity (33% [21-50]), as well as of levels of total protein S (51% [36-70]) and free protein S (53% [32-75]) (figure 2). Markers of activation of coagu-lation TATc and F1+2 were above upper reference value at baseline (10 ug/L [5-22]

Muller.indd 117 25-7-2014 14:40:26

Chapter 6

118

Figure 1: Levels of individual coagulation factors at baseline and after Fresh Frozen Plasma transfusion (12 ml/kg).

Figure 2: Levels of anticoagulant proteins before and after Fresh Frozen Plasma transfusion (12 ml/kg).

and 370 pMol/L [113-608] respectively). Also, the fibrinolytic marker PAP was above upper reference value at baseline (842 ug/L [322-1267]).

Effect of FFP transfusion on coagulation tests, factor levels and anticoagulantsFFP transfusion reduced median INR to 1.4 [1.3-1.6] (table 2). As expected, FFP trans-fusion increased median levels of coagulation factors II (44% [38-52]), V (58% [44-

Muller.indd 118 25-7-2014 14:40:26


119

Figure 3: Thrombin generation test results before and after Fresh Frozen Plasma transfusion (12 ml/kg).

Dotted lines indicate reference ranges.FFP = Fresh Frozen PlasmaETP = endogenous thrombin potential

90]) and VII (37% [28-55]). However, all factor levels remained under the lower limit of reference values after transfusion (figure 1). Protein C, S and antithrombin levels also increased in response to FFP transfusion (figure 2). FFP did not increase mark-ers of coagulation, but rather reduced levels of F1+2. TATc levels were unaffected by FFP transfusion. We also measured parameters of fibrinolysis. FFP did not affect PAP levels, although D-dimer concentration was reduced after FFP transfusion (table 2).The response of patients with and without bleeding complications following FFP transfusion did not differ, with no differences in levels of factors II, V and VII and anticoagulants antithrombin, protein C and S (data not shown).

Effect of FFP transfusion on thrombin generationIn 27 patients, paired samples were available to perform thrombin generation tests. At baseline, patients had prolonged lag time, which is the time from start of

Muller.indd 119 25-7-2014 14:40:27

Chapter 6

120

Table 2: Markers of coagulation at baseline and after Fresh Frozen Plasma transfusion (12 ml/kg).

Before FFPn=38

After FFPn=38

p value

INR 1.8 (1.5-2.2) 1.4 (1.3-1.6) <0.001aPTT (sec) 43 (38-52) 39 (32-46) <0.001Platelet count (*109/L) 89 (51-183) 96 (45-158) 0.001D dimer (mg/L) 7.5 (2.1-11.3) 6.4 (3.3-11.0) 0.009Fibrinogen (g/L) 3.1 (2.1-5.2) 2.8 (2.3-5.4) 0.97F1+2 (pMol/L) 370 (113-608) 323 (162-480) 0.02TATc (ug/L) 10 (5-22) 11 (6-22) 0.56PAP (ug/L) 842 (322-1267) 833 (411-1151) 0.11

Data expressed as median (IQR)Wilcoxon Signed Rank test.FFP = Fresh Frozen PlasmaINR = International Normalized RatioaPTT = Activated Partial Thromboplastin TimeF1+2 = Prothrombin Fragment 1+2TATc = Thrombin-antithrombin complexPAP = Plasmin-α2-antiplasmin complex

the assay until detection of the first thrombin (figure 3). Also, peak values and ETP (endogenous thrombin potential) were both reduced, indicating reduced thrombin generation, which is in line with the low individual factor levels in these patients. Transfusion of FFP resulted in slightly increased peak values, however values were still lower compared to the normal range, indicating a persistent hypocoagulable profile. The other thrombin generation parameters were unaffected by FFP transfu-sion, with persistent prolonged lag time and time to peak and reduced ETP (figure 3).

Discussion

In the current study, we demonstrated that critically ill patients with a prolonged INR have reduced levels of individual coagulation factors with a concurrent decrease in levels of natural occurring anticoagulant factors. Transfusion of a fixed dose of FFP improved individual factor levels, but also increased levels of natural anticoagulants. Thrombin generation was impaired in these patients and only improved marginally

Muller.indd 120 25-7-2014 14:40:27


121

in response to FFP transfusion. Thereby, FFP transfusion appears not to influence hemostatic balance in non-bleeding patients with an increased INR.INR is frequently used to assess bleeding risk and elevation is often a trigger to pro-phylactically administer FFP [18]. Indeed, we found that critically ill patients with an increased INR have reduced levels of individual coagulation factors, as found previ-ously [8]. These reduced levels may render them susceptible for enhanced bleeding. Concurrently, also in line with previous reports [6], we found that levels of natural anticoagulants antithrombin, protein C and S were reduced at the same time, hereby suggesting an unchanged hemostatic balance. As INR does not reflect altered levels of anticoagulants, this test poorly reflects in vivo coagulation and actual bleeding risk. In patients with liver disease, it has been shown that INR elevation indeed fails to predict bleeding complications [19]. In these patients, reduced levels of clotting factors, natural anticoagulants [20] and pro- and antifibrinolytic factors [21] led to the concept of “rebalanced hemostasis” [22]. Our results confirm this concept of rebalanced hemostasis in non-bleeding critically ill patients with a coagulopathy.The present study aimed to establish the effect of a standardized dose of FFP on individual components of the coagulation system, as well as on global tests of coag-ulation. We demonstrated that FFP transfusion indeed reduced INR value, in line with an increase in individual factor levels. However, all individual levels remained under the lower limit of the reference values. Of note, the increase in factor levels was equal in those patients experiencing bleeding after the intervention compared to those who did not. Equally important to the concept of the ability of FFP to miti-gate the risk of bleeding, is the effect of FFP on levels of anticoagulant factors. FFP resulted in a concomitant rise of levels of natural anticoagulants antithrombin, pro-tein C and S. Data on the effect of FFP on the hemostatic balance in non-bleeding critically ill patients are scarce. A frequently referenced study investigating factor levels after FFP transfusion in the critically ill showed decreased levels of all factors and small increments after FFP transfusion [8]. However, in this study, different doses of FFP were used in small patients groups and anticoagulant factor levels were not assessed, rendering the net effect of FFP on hemostatic balance unknown. In a study in critically ill neonates, prophylactic FFP transfusion resulted in similar results to ours, with increased coagulation factor levels but also levels of anticoagulants [23]. In our study, levels of F1+2 were slightly reduced and TATc levels were unaffected by

Muller.indd 121 25-7-2014 14:40:27

Chapter 6

122

FFP administration, contradicting enhanced thrombin generation as a result of FFP transfusion and suggesting that hemostatic balance is unaltered by FFP. Of note, in line with the concept of “rebalanced hemostasis”, the paradigm on the use of FFP in liver failure patients has also shifted and liver transplantations are increasingly per-formed without any transfusion despite increased INR values [24]. Taken together, administration of FFP supplies both pro-coagulant and anti-coagulant coagulation factors, resulting in an unchanged hemostatic balance. Despite INR correction, any change to a more pro-coagulant state following FFP transfusion is highly limited.Results of thrombin generation tests in our patients varied widely, but predomi-nantly showed a hypocoagulable profile. Reduced levels of individual coagulation factors, in particular factor VII, probably contributed to a delay in initial thrombin generation, resulting in increasing lag time and time to thrombin peak. In addition, low thrombin peak and ETP values indicate reduced overall thrombin generation. These results are in line with those previously reported in sepsis patients [6,25]. FFP transfusion resulted in only marginally increased peak values and did not affect ETP values, indicating a very limited effect of FFP transfusion on thrombin generation in our patients. In line with this, in a study similar to this one performed in critically ill neonates, prophylactic FFP transfusion was demonstrated to even attenuate throm-bin formation, which was thought to be due to increased levels of anticoagulants following FFP [23]. We did not perform thrombin generation assays with the addi-tion of thrombomodulin or activated protein C, which is a limitation of our study. In patients undergoing liver transplantation, thrombin generation was preserved after addition of thrombomodulin hereby demonstrating a defective endogenous anti-coagulant system. Indeed these patients had reduced levels of antithrombin and protein C and S [20]. Of note, levels of antithrombin, protein C and S were reduced to a similar extent in our patients. Therefore, the observed hypocoagulable profiles in thrombin generation assays do not preclude a preserved hemostatic balance, as stated previously.Our study has several other limitations. First, our group size was relatively small. However, patient characteristics correspond well with those of patients with a coagulopathy in a large prospective cohort study in mixed medical/surgical ICUs [1], supporting the generalizability of our results. Second, the dose of FFP was cho-sen aiming to reduce INR to <1.5 and to not fully correct INR [13]. The observed

Muller.indd 122 25-7-2014 14:40:27


123

limited increment in coagulation factor levels could have been calculated before-hand and higher dose of FFP than 12 ml/kg would have been required to normalize factor levels. However, transfusion with higher doses of FFP is not current practice [2,18,26,27], moreover audits have revealed that a substantial number of patients is transfused with a dose less than 10 ml/kg [14,28]. As transfusion with doses of FFP used in this study reflect current practice, we think that results are relevant to the practicing physician.

Conclusion

Critically ill patients with increased INR display a hypocoagulable state with impaired thrombin generation. Prophylactic FFP transfusion resulted in an increase, but not in correction, of individual coagulation factors with a concomitant increase in levels of anticoagulants. In addition effect on thrombin generation was highly limited, hereby the net hemostatic balance remained unchanged after FFP transfusion. These results underline the lack of rationale to administer FFP to non-bleeding ICU patients with a coagulopathy.

Muller.indd 123 25-7-2014 14:40:27

Chapter 6

124

References



3. Gulati G, Hevelow M, George M, et al.: International normalized ratio versus plasma levels of coagulation factors in patients on vitamin K antagonist therapy. Arch Pathol Lab Med 2011;135: 490-4


5. Levi M, Meijers JC: DIC: which laboratory tests are most useful. Blood Rev 2011;25: 33-76. Collins PW, Macchiavello LI, Lewis SJ, et al.: Global tests of haemostasis in critically ill

patients with severe sepsis syndrome compared to controls. Br J Haematol 2006;135: 220-7




10. Williamson LM, Llewelyn CA, Fisher NC, et al.: A randomized trial of solvent/detergent-treated and standard fresh-frozen plasma in the coagulopathy of liver disease and liver transplantation. Transfusion 1999;39: 1227-34


12. Hildebrandt B, Machotta A, Riess H, et al.: Intraoperative fresh-frozen plasma versus human albumin in craniofacial surgery--a pilot study comparing coagulation profiles in infants younger than 12 months. Thromb Haemost 2007;98: 172-7




Muller.indd 124 25-7-2014 14:40:27


125

16. Kager LM, de Boer JD, Bresser P, et al.: Intrabronchial activated protein C enhances lipopolysaccharide-induced pulmonary responses. Eur Respir J 2013;42: 188-97

17. Hemker HC, Giesen P, Al Dieri R, et al.: Calibrated automated thrombin generation measurement in clotting plasma. Pathophysiol Haemost Thromb 2003;33: 4-15


19. Tripodi A, Caldwell SH, Hoffman M, et al.: Review article: the prothrombin time test as a measure of bleeding risk and prognosis in liver disease. Aliment Pharmacol Ther 2007;26: 141-8

20. Lisman T, Bakhtiari K, Pereboom IT, et al.: Normal to increased thrombin generation in patients undergoing liver transplantation despite prolonged conventional coagulation tests. J Hepatol 2010;52: 355-61

21. Lisman T, Leebeek FW, Mosnier LO, et al.: Thrombin-activatable fibrinolysis inhibitor deficiency in cirrhosis is not associated with increased plasma fibrinolysis. Gastroenter-ology 2001;121: 131-9

22. Lisman T, Porte RJ: Rebalanced hemostasis in patients with liver disease: evidence and clinical consequences. Blood 2010;116: 878-85

23. Hyytiainen S, Syrjala M, Fellman V, et al.: Fresh frozen plasma reduces thrombin forma-tion in newborn infants. J Thromb Haemost 2003;1: 1189-94

24. Jabbour N, Gagandeep S, Shah H, et al.: Impact of a transfusion-free program on non-Jehovah’s Witness patients undergoing liver transplantation. Arch Surg 2006;141: 913-7

25. Petros S, Kliem P, Siegemund T, et al.: Thrombin generation in severe sepsis. Thromb Res 2012;129: 797-800




Muller.indd 125 25-7-2014 14:40:27

Muller.indd 126 25-7-2014 14:40:27

Chapter 7

Correlation of thromboelastometry with conventional

hemostatic tests in critically ill patients with a

coagulopathy treated with fresh frozen plasma

Marcella C.A. Müller, Marleen Straat, J.Harriët Klinkspoor, Margreeth B. Vroom, Nicole P. Juffermans


Muller.indd 127 25-7-2014 14:40:27

Chapter 7

128

Abstract

Introduction: Increased INR values have a high prevalence among critically ill patients and are often a trigger to administer Fresh Frozen Plasma (FFP). However, ability of INR to accurately reflect hemostasis is limited. In contrast to INR, rotational throm-boelastometry (ROTEM) assesses whole clot formation and degradation.Aim: Assessing correlation of ROTEM with conventional hemostatic tests in critically ill patients with a coagulopathy and evaluating the response of thromboelastometry variables to a prophylactic transfusion of a fixed dose of FFP in these patients.Methods: In a sub-study of a prospective trial evaluating the efficacy of FFP to pre-vent bleeding, ROTEM assays were performed in 16 critically ill patients with a pro-longed INR (1.5-3.0) before and after transfusion of a fixed dose of FFP (12 ml/kg).Results: At baseline, only INTEM CFT was prolonged to 133 (51-162), while all other variables were within reference ranges. INR showed weak correlation with INTEM CFT, and was not correlated with EXTEM CFT. In patients with disseminated intra-vascular coagulation (DIC), EXTEM CFT was increased (218 sec (154-409) vs. 60 sec (38-165), p=0.007), whereas alpha (55° (44-65) vs. 79° (70-83), p=0.003) and MCF (47 mm (36-49) vs. 65 mm (55-77), p=0.01) were decreased compared to those with-out DIC, indicative of a hypocoagulable state. Also, CFT, MCF and alpha correlated well with platelet count, fibrinogen and factor II levels, as well as with DIC score. EXTEM CFT, alpha and MCF had high accuracy in discriminating patients with and without DIC. Transfusion of FFP decreased EXTEM CT (70 (46-94) before vs. (58 (46-82) after FFP, p=0.04) and MCF (51 mm (43-65) before FFP vs. 55 (47-75) after FFP, p=0.04) and decreased FIBTEM CT (66 sec (50-96) before vs. 49 (40-66) after FFP, p=0.01).Conclusion: ROTEM parameters were mostly within reference ranges and correlated only modestly with INR, but did correlate with platelet count, fibrinogen and factor II levels. Patients with DIC had more hypocoagulable profiles and ROTEM could be useful to diagnose DIC. Transfusion of FFP improved EXTEM CT and MCF.

Muller.indd 128 25-7-2014 14:40:27

Thromboelastometry and FFP transfusion in critically ill patients with a coagulopathy

129

Background

Coagulopathy, defined as prolonged prothrombin time (PT) or increased Interna-tional Normalized Ratio (INR), has a high prevalence in critically ill patients [1]. Audits of transfusion practice show that clinicians consider increased INR and PT values as indicators of increased bleeding risk and thereby they are an important trigger to transfuse Fresh Frozen Plasma (FFP), in particular in patients undergoing inva-sive procedures [2,3]. However, despite the widespread use of INR and PT to detect coagulopathy, these tests do not accurately reflect in vivo hemostatic potential [4], which may be due to the fact that only a part of the coagulation cascade is assessed. Essential contributors to hemostasis, such as platelet count and levels of endog-enous anticoagulants, both of which are often decreased in the critically ill, are not accounted for [5,6].An important cause of abnormal coagulation test results in critically ill patients is the presence of disseminated intravascular coagulation (DIC). The clinical picture of DIC ranges from severe consumption coagulopathy with increased bleeding tendency to a prothrombotic state with (micro)vascular thrombosis. Conventional coagulation tests lack the ability to discriminate between these profiles.Ideally, a test to assess bleeding risk in the critically ill should take into account all factors contributing to hemostasis and be readily available for clinicians caring for these patients. In recent years there has been a growing interest in rotational thromboelastometry (ROTEM), a point of care test evaluating whole clot formation and degradation. Use of ROTEM was shown to reduce the amount of transfusions in cardiac surgery [7] and ROTEM may also be of use to detect DIC in the critically ill [8,9]. Therefore we aimed to study the correlation of thromboelastometry with other hemostatic tests and ISTH DIC score in critically ill patients with a coagulopa-thy. Also, the effect of transfusion of prophylactic FFP on ROTEM parameters in non-bleeding patients has not been evaluated before. As a substudy of a trial evaluating efficacy of prophylactic FFP in coagulopathic patients prior to undergoing an inter-vention, the response of thromboelastometry variables to a fixed dose of FFP was investigated.

Muller.indd 129 25-7-2014 14:40:27

Chapter 7

130

Methods

Setting and patientsThe study was performed as a predefined post-hoc substudy of a randomized clini-cal trial on the efficacy of prophylactic FFP in critically ill patients with a coagulopa-thy [10]. Patients were either randomized to receive or not to receive 12 ml/kg of FFP transfusion before an intervention. Patients were eligible if INR was ≥1.5 and ≤3.0 and an intervention was scheduled. Defined invasive procedures were inser-tion of a central venous catheter, thoracocentesis, percutaneous tracheotomy or drainage of abscess or fluid collection. Patients randomized to FFP transfusion were included in the current study. Patients with clinically overt bleeding or with a throm-bocytopenia <30 x 109/L were excluded from participation. Patients using platelet aggregation inhibitors, low molecular weight heparin in a therapeutic dose, vitamin K antagonists, activated protein C or prothrombin complex concentrates were also excluded. Patients treated with heparin were eligible if medication was discontin-ued for an appropriate period.The study was conducted in a mixed-medical surgical ICU in a university hospital, the Academic Medical Center in Amsterdam, the Netherlands in accordance with the Declaration of Helsinki. The protocol was approved by our Institutional Review Board. Written informed consent was obtained from patients or legal representative before entry in the study.

Data collectionBaseline data included demographics, admission reason, APACHE IV and SOFA score and previous medical history. DIC was assessed using the ISTH DIC score, which defines DIC as a score of ≥5 points [11]. Blood samples were drawn just prior to FFP transfusion and immediately after FFP transfusion.

Coagulation assaysRoutine coagulation tests included INR, aPTT, fibrinogen, d-dimer levels (Sysmex CA 7000 and all reagents, Siemens Healthcare Diagnostics, Germany) and platelet count. In addition, levels of coagulation factors II, V and VII were assessed using

Muller.indd 130 25-7-2014 14:40:27


131

a one stage clotting assay according to manufacturers instruction (ACL TOP 700, Instrumentation Laboratory, USA).Antithrombin was assessed by chromogenic substrate method (Sysmex CA 7000, Siemens Healthcare Diagnostics, Germany) with reagents and protocols of the man-ufacturer. Plasmin-α2-antiplasmin complex (PAP) levels were measured using spe-cific commercially available ELISAs according to the instruction of the manufacturer (Siemens Healthcare Diagnostics, Germany). Protein C activity was assessed by a kinetic assay (Coamatic Protein C, Chromogenix, Mölndal, Sweden) and protein S levels were determined by ELISA, as described previously [12]. All tests were carried out before and directly after FFP transfusion.

ROTEM measurementsUsing ROTEM (Tem International, Munich, Germany), three separate assays were carried out, including EXTEM to assess tissue factor-initiated coagulation, INTEM to assess the intrinsic pathway and FIBTEM to qualitatively assess fibrinogen status. For EXTEM, 20 μL of 0.2 mol/L CaCl2 (star-tem®) and 20 μL of recombinant tissue factor (r EXTEM®) were added to a test vial, subsequently 300 μL of the citrated blood sam-ple was added. For the INTEM test 20 μL of 0.2 mol/L CaCl2 (star-tem®), 20 μL of par-tial thromboplastin made of rabbit brain (in-tem®) and 300 μL of blood were added to the test cuvette. FIBTEM test was carried out by adding 20 μL of recombinant human tissue factor (r EXTEM®), 20 μL of platelet inactivating cytochalasin D solu-tion 0.2 mol/L CaCl2 and 300 μL of the blood sample to the test vial. The electronic pipette program guided all test steps. For INTEM and EXTEM clotting time (CT), clot formation time (CFT), clot firmness (MCF), alpha angle and maximum lysis (ML) were recorded. For the FIBTEM assay CT and MCF were recorded. All treating physicians were blinded for ROTEM results.

Statistical analysisAll variables are expressed as median (interquartile ranges). To compare groups Mann Whitney was used for independent variables and Wilcoxon signed rank test was used for paired data. Fisher’s exact test was used for comparisons in cross-tabs. Correlations were assessed using Spearman’s rho. To assess value of ROTEM variables to diagnose DIC, receiver operating characteristic (ROC) curves were used.

Muller.indd 131 25-7-2014 14:40:27

Chapter 7

132

Results are given as area under the curve (AUC) and 95% confidence interval (CI). In addition optimal cut-off values ad odds ratio’s were calculated for each parameter. A p value less than 0.05 was considered significant. Statistical analyses were per-formed with SPSS 20.0 (SPSS Inc., Chicago, USA) and Prism Version 5.0 (Graphpad Software, San Diego, USA).

Results

PatientsOf the 16 included patients (table 1), the majority was admitted to the ICU due to sepsis. Accordingly, patients were ill, as reflected by high disease severity scores. Eight patients fulfilled criteria for overt DIC, with a median platelet count of 69 x 109/L. Median levels of natural anticoagulants antithrombin, protein C and S were reduced. Most patients underwent central venous catheterization and minor bleed-ing occurred after six interventions, including hematoma requiring prolonged pres-sure, suture or application of spongostan. No major bleedings were observed.

Thromboelastometry and assessment of coagulation statusBaseline thromboelastometry results are depicted in figure 1. With the exception of INTEM CFT, which was prolonged, all median variables were within manufacturers reference values. However, EXTEM CFT was at the upper reference limit. In addition, both median INTEM and EXTEM MCF were on the lower reference range, indicating a tendency to hypocoagulability.Baseline CT for both INTEM and EXTEM did not correlate with INR, nor did EXTEM CFT, while INTEM CFT showed moderate positive correlation with INR. Both INTEM and EXTEM CFT correlated positively with aPTT (tables 2a and 2b). CFT, alpha and MCF (INTEM and EXTEM) correlated well with platelet count, fibrinogen and factor II levels. Other factor levels did not correlate with ROTEM variables (tables 2a and 2b). Lysis indexes did not correlate with markers of fibrinolysis such as D-dimer and plasmin-antiplasmin levels (data not shown).ROTEM parameters did not differ between patients with and without bleeding (data not shown).

Muller.indd 132 25-7-2014 14:40:27


133

Table 1: Characteristics of critically ill patients with a coagulopathy prior to receiving pro-phylactic FFP for an intervention.

n = 16General characteristicsGender, male, % (n) 63 (10)Age (years) 58 (46-69)APACHE IV score 115 (81-136)SOFA score 14 (8-16)Medical conditionLiver disease, % (n) 25 (4)Sepsis, % (n) 56 (9)DIC, % (n) 50 (8)Coagulation INR 1.60 (1.54-2.12)aPTT (sec) 40 (35-47)Platelet count x109/L 69 (49-177)Fibrinogen (g/L) 2.60 (1.95-4.80)Antithrombin (%) 54 (38-68)Protein C (%) 34 (22-53)Total protein S (%) 47 (39-72)InterventionCentral venous catheter, % (n) 69 (11)Chest tube, % (n) 13 (2)Tracheotomy, % (n) 13 (2)Abdominal drain, % (n) 6 (1)OutcomeOccurrence of bleeding, % (n)* 38 (6)

Data expressed as median and interquartile ranges.*Assessed using a tool validated in critically ill (HEME)[24]APACHE = Acute Physiology and Chronic Health EvaluationSOFA = Sequential Organ Failure AssessmentDIC = Disseminated Intravascular CoagulationaPTT = Activated Partial Thromboplastin TimeINR = International Normalized Ratio

Thromboelastometry in patients with disseminated intravascular coagulationIn patients with DIC, thromboelastometry profiles were more hypocoagulable com-pared to those without DIC (figure 2). Also, prolonged CFT and reduced alpha and MCF were significantly associated with DIC score (tables 2a and 2b). Although the proportion of patients with DIC that had EXTEM test results outside manufacturers

Muller.indd 133 25-7-2014 14:40:27

Chapter 7

134

Figure 1: ROTEM EXTEM variables before and after Fresh Frozen Plasma transfusion (12 ml/kg). Dotted lines indicate reference ranges.

FFP = Fresh Frozen PlasmaCT = clotting timeCFT = clot formation timeMCF = maximum clot firmness

reference range did not differ from those without DIC (p=0.12), a ROC curve analysis was performed to investigate the value of these variables for the diagnosis of DIC. ROC curves for EXTEM CFT, alpha and MCF are shown in figure 3. Comparison of EXTEM CFT, alpha and MCF in patients with and without DIC showed all three vari-ables were capable of detecting differences between these groups with high accu-racy (table 3).

Response of thromboelastometry variables to transfusion with FFPTransfusion with 12 ml/kg FFP reduced median EXTEM CT and improved median EXTEM MCF, indicating enhanced coagulation. However, in most patients, variables

Muller.indd 134 25-7-2014 14:40:28


135

Figure 2: ROTEM EXTEM variables before and after Fresh Frozen Plasma transfusion (12 ml/kg) in patients with and without disseminated intravascular coagulation.

FFP = Fresh Frozen PlasmaDIC = Disseminated intravascular coagulationCT = clotting timeCFT = clot formation timeMCF = maximum clot firmness

only changed marginally after FFP transfusion (figure 1). FIBTEM MCF was unaffected by FFP transfusion and the same applied for the INTEM variables (CT: 183 (175-218) before vs. 175 (159-197) after FFP (p=0.40), CFT: 133 (51-162) before vs. 115 (60-140) after FFP (p=0.33), alpha: 74 (63-80) before vs. 72 (68-78) after FFP (p=0.63) and MCF: 52 (43-70) before vs. 55 (49-69) after FFP (p=0.19)). The effect of FFP transfu-sion did not differ among patients with or without DIC (figure 2).

Muller.indd 135 25-7-2014 14:40:28

Chapter 7

136

Table 2a: Correlation of INTEM ROTEM variables with coagulation tests at baseline.

CT CFT Alpha MCFINR 0.174 0.538* NA NAaPTT (sec) 0.430 0.646* NA NAPlatelet count (x109/L) -0.270 -0.908** 0.725** 0.811**Factor II (%) -0.609* -0.785** 0.839** 0.794**Factor V (%) -0.319 -0.187 0.187 0.279Factor VII (%) -0.187 -0.451 0.487 0.389Fibrinogen (g/L) -0.498 -0.769** 0.865** 0.665**DIC score 0.452 0.802** -0.918** -0.636*

Table 2b: Correlation of EXTEM ROTEM variables with coagulation tests at baseline.

CT CFT Alpha MCFINR -0.022 0.369 NA NAaPTT (sec) 0.318 0.548* NA NAPlatelet count (x109/L) -0.487 -0.833** 0.818** 0.613*Factor II (%) -0.412 -0.689** 0.732** 0.513*Factor V (%) -0.146 -0.190 0.259 0.018Factor VII (%) -0.062 -0.306 0.373 0.484Fibrinogen (g/L) -0.233 -0.743** 0.765** 0.724**DIC score 0.271 0.796** -0.860** -0.790**

*p<0.05**p<0.01INR = International normalized ratioaPTT = activated partial thromboplastin timeDIC = disseminated intravascular coagulationCT = clotting timeCFT = clot formation timeMCF = maximum clot firmnessNA = not applicable

Table 3: Results of Receiver Operating Characteristics curve analysis for EXTEM variables in patients with and without disseminated intravascular coagulation.

AUC 95% CI p value Cut off Sensitivity Specificity Likelihood ratio

CFT 0.906 0.757-1.056 <0.01 >105 sec 100% 75% 3.0alpha 0.945 0.841-1.050 <0.01 <67° 88% 88% 7.0MCF 0.883 0.688-1.078 0.01 <51 mm 88% 88% 7.0

CFT = clot formation timeMCF = maximum clot firmness

Muller.indd 136 25-7-2014 14:40:28


137

Discussion

The current study demonstrates that most ROTEM variables are within reference values in critically ill patients with an increased INR. Correlation with the widely used INR is moderate, while ROTEM correlates well with factor II, platelet count and fibrinogen. ROTEM profiles are more hypocoagulable in patients with DIC compared to those without DIC and CFT, alpha and MCF correlate strongly with DIC score. Prophylactic transfusion of FFP slightly improved ROTEM EXTEM results and FIBTEM CFT.In the present study, ROTEM measurements were done in severely ill patients with increased INR values. Thromboelastometry results were heterogeneous ranging from hyper- to hypocoagulable profiles, as shown before in ICU patients [13]. How-ever for the whole cohort, most results were within reference ranges, which is in line with reports from patients with severe sepsis [6,14]. Thereby, it can be questioned whether reference values as validated in healthy volunteers for a single outcome should be used to diagnose hypo- or hypercoagulability in the critically ill, or rather

Figure 3: Receiver operating characteristics curves for EXTEM clot formation time, alpha and maximum clot firmness for the diagnosis of disseminated intravascular coagulation in criti-cally ill patients with coagulopathy.

CFT = clot formation timeMCF = maximum clot firmness

Muller.indd 137 25-7-2014 14:40:28

Chapter 7

138

that new reference values should be formulated for specific profiles by combining several ROTEM values.We assessed the correlation of ROTEM with conventional coagulation tests and indi-vidual factor levels. INR values increase when levels of factor II, VII, IX and X are reduced and a reduction of these factors would result in concomitant prolonged coagulation times in the ROTEM assay, as was observed in patients with sepsis [15]. Indeed, INTEM CFT was prolonged and EXTEM CFT was at the upper reference limit in our patients. However, correlation between INR values and ROTEM was only seen in the INTEM assay and not in the EXTEM, as reported previously [16]. A pos-sible explanation for this observation could be that compared to activation of the intrinsic pathway, EXTEM is less standardized due to the use of tissue factor [17]. This has led to the recommendation to use INTEM for patients with known bleeding diathesis [18]. Of note, correlation with INR was only modest. A stronger correlation was found between CFT, alpha and MCF with fibrinogen, factor II levels and platelet count in our patients, which is in line with recent reports in surgical patients [16] and sepsis patients [14]. Hereby, ROTEM could provide valuable information on different components of the coagulation system in a bedside manner, abating the need to perform different assays.Half of the patients in the current study had overt DIC. Compared to patients with-out DIC, ROTEM CFT was prolonged and alpha and MCF were reduced, indicative of a more hypocoagulable profile. Also, ROTEM correlated highly with the DIC score. Comparable observations have been reported in sepsis patients [6,8]. Similar to our findings, the combination of CFT, alpha and MCF values was able to discriminate between patients with and without DIC [8]. This suggests that ROTEM could be a useful diagnostic test for DIC [19]. Further research is warranted enabling develop-ment of a thromboelastographic DIC score for critically ill with appropriate refer-ence values.Transfusion with a fixed dose of FFP reduced CT in EXTEM and FIBTEM and increased MCF in EXTEM, whereas INTEM variables remained unaffected. Although FFP is widely used to correct coagulopathy in critically ill patients [1], effect of this prac-tice on coagulation parameters is limited. An increase in individual factor levels and decrease of INR following FFP transfusion has been demonstrated in different patient groups [20-22], however to our best knowledge this is the first report of the

Muller.indd 138 25-7-2014 14:40:28


139

effect of FFP on thromboelastometric variables. Of note, observed increments in this study were only small and the clinical significance of the observed improvement is doubtful.Our study has several limitations. First, group size is relatively small, although patient characteristics correspond well with those of larger cohort studies on thromboelas-tometry in the ICU [13], supporting the generalizability of our results on correlation between ROTEM and other hemostatic tests. However, although we found no differ-ences in ROTEM parameters between patients with and without bleeding following an intervention, assessment of the ability of ROTEM to predict bleeding complica-tions requires a larger sample size. Another limitation is test precision of ROTEM, which is a subject of debate with reported high coefficients of variance [23]. How-ever, in order to limit variation as much as possible, all tests were carried out on the same device by only two experienced researchers.

Conclusion

In critically ill patients with a coagulopathy, ROTEM parameters were mostly within manufacturers reference ranges. Correlation with INR was moderate, but CFT, alpha and MCF correlated well with platelet count and levels of factor II and fibrinogen, as well as with DIC score. Also, EXTEM variables showed high accuracy in discriminating patients with and without DIC. Thereby, ROTEM may be a useful tool in diagnosing DIC in the critically ill. Transfusion of FFP resulted in improved EXTEM CT and MCF and reduced CT in FIBTEM.

Muller.indd 139 25-7-2014 14:40:28

Chapter 7

140

References





5. Strauss R, Wehler M, Mehler K, et al.: Thrombocytopenia in patients in the medical intensive care unit: bleeding prevalence, transfusion requirements, and outcome. Crit Care Med 2002;30: 1765-71

6. Brenner T, Schmidt K, Delang M, et al.: Viscoelastic and aggregometric point-of-care testing in patients with septic shock - cross-links between inflammation and haemosta-sis. Acta Anaesthesiol Scand 56: 1277-90

7. Bolliger D, Tanaka KA: Roles of thrombelastography and thromboelastometry for patient blood management in cardiac surgery. Transfus Med Rev 2013;27: 213-20

8. Sivula M, Pettila V, Niemi TT, et al.: Thromboelastometry in patients with severe sepsis and disseminated intravascular coagulation. Blood Coagul Fibrinolysis 2009;20: 419-26

9. Müller MC, Meijers JC, Vroom MB, et al.: Utility of thromboelastography and/or throm-boelastometry in adults with sepsis: a systematic review. Crit Care 2014;18: R30

10. Müller MC, de Jonge E, Arbous MS, et al.: Transfusion of fresh frozen plasma in non-bleeding ICU patients--TOPIC trial: study protocol for a randomized controlled trial. Tri-als 2011;12: 266


12. Kager LM, de Boer JD, Bresser P, et al.: Intrabronchial activated protein C enhances lipopolysaccharide-induced pulmonary responses. Eur Respir J 2013;42: 188-97

13. Johansson PI, Stensballe J, Vindelov N, et al.: Hypocoagulability, as evaluated by thrombelastography, at admission to the ICU is associated with increased 30-day mor-tality. Blood Coagul Fibrinolysis 21: 168-74



Muller.indd 140 25-7-2014 14:40:28


141

16. Theusinger OM, Schroder CM, Eismon J, et al.: The influence of laboratory coagulation tests and clotting factor levels on Rotation Thromboelastometry (ROTEM(R)) during major surgery with hemorrhage. Anesth Analg 2013;117: 314-21

17. Young G, Sorensen B, Dargaud Y, et al.: Thrombin generation and whole blood viscoe-lastic assays in the management of hemophilia: current state of art and future perspec-tives. Blood 2013;121: 1944-50

18. Chitlur M, Rivard GE, Lillicrap D, et al.: Recommendations for performing thromboe-lastography/thromboelastometry in hemophilia: communication from the SSC of the ISTH. J Thromb Haemost 2014;12: 103-6

19. Sharma P, Saxena R: A novel thromboelastographic score to identify overt disseminated intravascular coagulation resulting in a hypocoagulable state. Am J Clin Pathol 134: 97-102




23. Kitchen DP, Kitchen S, Jennings I, et al.: Quality assurance and quality control of thromb-elastography and rotational Thromboelastometry: the UK NEQAS for blood coagulation experience. Semin Thromb Hemost 2010;36: 757-63

24. Arnold DM, Donahoe L, Clarke FJ, et al.: Bleeding during critical illness: a prospective cohort study using a new measurement tool. Clin Invest Med 2007;30: E93-102

Muller.indd 141 25-7-2014 14:40:28

Muller.indd 142 25-7-2014 14:40:28

Chapter 8

Effect of transfusion of fresh frozen plasma on

parameters of endothelial condition and inflammatory

status in non-bleeding critically ill patients

Marleen Straat, Marcella C.A. Müller, Joost C.M. Meijers, M. Sesmu Arbous, Angelique M.E. Spoelstra – de Man, Charlotte J.P. Beurskens,

Anita M Tuip – De Boer, Margreeth B. Vroom, Nicole P. Juffermans


Muller.indd 143 25-7-2014 14:40:28

Chapter 8

144

Abstract

Much controversy exists on the effect of a fresh frozen plasma (FFP) transfusion on systemic inflammation and endothelial damage. Adverse effects of FFP have been well described, including acute lung injury. However, it is also suggested that a higher amount of FFP decreases mortality in trauma patients requiring a massive transfu-sion. Furthermore, FFP has an endothelial stabilizing effect in experimental models. In 38 coagulopathic non-bleeding critically ill patients receiving a prophylactic trans-fusion of FFP (12 ml/kg) prior to an invasive procedure, we measured inflamma-tory cytokines and markers of endothelial condition before and after transfusion. At baseline, systemic cytokine levels were mildly elevated in critically ill patients. Surprisingly, FFP transfusion resulted in a decrease of TNF-α (from 11.3 to 2.3 pg/ml, p=0.01). Other cytokines were not affected. FFP also resulted in a decrease in sys-temic syndecan-1 levels from 675 to 565 pg/ml (p=0.01) and a decrease in median factor VIII levels (from 246 to 246%, p<0.01), suggestive of an improved endothelial condition. This was associated with an increase in ADAMTS13 levels (from 24 to 32%, p<0.01), with a concomitant decrease in von Willebrandfactor (vWF)(from 474 to 423%, p<0.01), In conclusion, a fixed dose of FFP transfusion seems to stabilize endothelial condition, possibly by increasing ADAMTS13 which cleaves vWF.

Muller.indd 144 25-7-2014 14:40:28

FFP transfusion and endothelial function in the critically ill

145

Introduction

Substantial amounts of Fresh Frozen Plasma (FFP) are utilized in the intensive care unit (ICU)[1,2]. FFP is effective in correcting clotting factor deficiencies [3] and is therefore transfused in patients with active bleeding, but also frequently in patients with abnormal coagulation tests to prevent bleeding [2,4]. In sepsis patients, FFP transfusion rates of up to 57% have been reported [5]. However, there is an asso-ciation between FFP transfusion and adverse outcome in the critically ill, including transfusion-related acute lung injury (TRALI) [4,6-8], transfusion-related circulatory overload [9,10], multi-organ failure [8,11] and an increased risk of infections [12]. Although not entirely understood, the pathological mechanisms underlying the association between FFP transfusion and lung injury is thought to result from an inflammatory response including a neutrophil influx into the lungs and elevated pulmonary levels of IL-8 and IL-1, as demonstrated in TRALI patients [13,14]. In line with this, FFP increased expression of endothelial adhesion molecules in human pulmonary endothelial cells [15]. Together, these data suggest that endothelial cell activation and disruption may be an early event following lung injury due to transfu-sion [16].On the other hand, FFP also seems to have protective effects. In trauma patients requiring a massive transfusion, it is suggested that resuscitation with a higher ratio of FFP to red blood cell units decreases mortality [17,18]. Of interest, this decreased mortality is irrespective of correction of coagulopathy by restoring coagulation fac-tor levels [18,19]. Instead, a beneficial effect of FFP may be related to the restoration of injured endothelium. Patients in hemorrhagic shock have a disrupted endothelial integrity and glycocalix layer, with decreased syndecan-1 expression [20]. Synde-can-1 is a proteoglycan on the luminal surface of endothelial cells which inhibits leu-kocyte adhesion and is shed during endothelial damage, resulting in increased levels of syndecan-1 in the systemic compartment [21]. In a hemorrhagic shock model, FFP was found to improve endothelial integrity, associated with increased expression of syndecan-1 on endothelial cells [22].Vascular integrity is also disrupted in various populations of critically ill patients, as demonstrated by increased levels of syndecan-1 [23,24]. The effect of a transfu-sion of a fixed dose of FFP on endothelial and cytokine host response in patients is

Muller.indd 145 25-7-2014 14:40:28

Chapter 8

146

not known. In a study investigating the risk-benefit ratio of FFP transfusion in non-bleeding critically ill patients with a coagulopathy, we investigated host response to a fixed dose of FFP transfusion.

Methods

Study designThis was a predefined post-hoc sub-study of a multicenter trial in which non-bleeding critically ill patients with an increased International Normalized Ratio (INR 1.5-3.0) were randomized between May 2010 and June 2013 to omitting or administering a prophylactic transfusion of FFP (12 ml/kg) prior to an invasive procedure. Only patients randomized to receive FFP were included in this sub-study. Patients were enrolled at 3 sites in The Netherlands: two university hospitals (Academic Medical Center, Amsterdam and Leiden University Medical Center, Leiden) and one large teaching hospital (Tergooi Ziekenhuizen, Hilversum). The Institutional Review Board of the Academic Medical Center - University of Amsterdam, Amsterdam, The Neth-erlands, approved the study protocol. Before entry in the study, written informed consent was obtained from the patient or legal representative in accordance with the Declaration of Helsinki. The study protocol was registered with trial identifica-tion numbers NTR2262 and NCT01143909 [25].Exclusion criteria were clinically overt bleeding, thrombocytopenia of <30 x 109/L, treatment with vitamin K antagonists, activated protein C, abciximab, tirofiban, ticlopidine or prothrombin complex concentrates and a history of congenital or acquired coagulation factor deficiency or bleeding diathesis. Patients treated with low molecular weight heparin (LMWH) or heparin in therapeutic dose were eligible if medication was discontinued for an appropriate period. Sepsis was defined by the Bone criteria [26]. Disseminated intravascular coagulation (DIC) was defined by an ISTH score of ≥5 [27]. The FFP was quarantaine plasma manufactured by Sanquin, the Dutch National Bloodbank. As of 2007, women are deferred from donation for preparation of FFP in the Netherlands.

Muller.indd 146 25-7-2014 14:40:28


147

Sample collectionCitrated blood samples were drawn from an indwelling arterial catheter before and directly after FFP transfusion. Samples were collected in sodium citrate (0.109M 3.2%) tubes and were centrifuged twice within 30 minutes: first 15 minutes at 2000 x g and then 5 mins at 15000 x g, both at 18°C. Supernatant was collected and stored at -80°C.

AssaysTumor necrosis factor-α (TNF- α) levels were measured by enzyme-linked immuno-sorbent assay (ELISA), according to instructions of the manufacturer (R&D Systems Inc., Minneapolis, MN, USA). Serum levels of IL–1β, IL–1RA, IL–8, IL–10, Macrophage Inflammatory Proteins (MIP)-1A, Monocyte Chemotactic Protein (MCP)–1 and sol-uble CD40 ligand were determined by Luminex, according to instructions of the manufacturer (Merck Millipore Chemicals BV; Amsterdam; the Netherlands). When less than 50 beads were measured by the Luminex assay, samples were excluded from further analysis. Von Willebrand Factor (VWF) antigen (VWF:Ag) levels were determined with an in-house enzyme-linked immunosorbent assay (ELISA) assay using commercially available polyclonal antibodies against VWF (DAKO, Glostrup, Denmark). ADAMTS13 antigen levels were measured using a commercially available ELISA according to the manufacturer’s instructions (Sekisui Diagnostics, Stamford, CT) and ADAMTS13 activity was assessed using the FRETS-VWF73 assay (Peptanova, Sandhausen, Germany) based on the method described by Kokame et al. [28].

Statistical analysisVariables are expressed as medians and interquartile ranges or means and standard deviations. For comparisons, a paired t-test was used, or in case of not normally dis-tributed data the Wilcoxon signed rank test. Statistical uncertainties are expressed in 95% confidence limits, with a significance level of 0.05. For the analyses, we used SPSS 20 and Graphpad Prism 5.

Muller.indd 147 25-7-2014 14:40:28

Chapter 8

148

Results

PatientsFrom 38 patients receiving FFP, paired samples from 33 patients were available for analysis before and after FFP transfusion. Patients were ill, as reflected by a high disease severity score and half of the patients had sepsis (table 1). Patients received a mean dosage of 11.2 (2.8) ml/kg FFP.

Inflammatory cytokine and chemokine levels before and after transfusion of 12 ml/kg FFPAt baseline, levels of cytokines were mildly elevated in this cohort. After FFP transfu-sion, median TNF-α decreased (p=0.01, table 2). Levels of all other cytokines were not affected by FFP transfusion. Chemokine levels IL-8 and MCP1 were elevated at baseline but also not influenced by FFP transfusion. Levels of soluble CD40L, which has been implicated as a mediator in TRALI [29], were also not significantly altered by FFP transfusion.

Parameters of endothelial condition before and after transfusion of 12 ml/kg FFPAfter FFP transfusion, levels of ADAMTS13 increased (p<0.01, figure 1). This increase was accompanied by a decrease in VWF (p<0.01) and in levels of syndecan-1. Factor VIII levels were slightly decreased following FFP transfusion (p=0.01).

Discussion

Our patients had mildly elevated levels of inflammatory cytokines at baseline, which corresponds to levels measured before in critically ill patients [30]. We observed no aggravation of this inflammatory response after FFP transfusion. Rather, there was a decrease in TNF-α level. This is not in line with a study in which FFP elicited an inflammatory response in endothelial cells [31], nor with an in vitro model of transfusion, in which whole blood incubated with FFP induced TNF production [32].Of the cytokines we measured, only TNF-α changed after FFP transfusion. As TNF-α is known to be the quickest responder among all cytokines, we may have timed our measurement too early after FFP transfusion to note an effect of FFP on other

Muller.indd 148 25-7-2014 14:40:28


149


FFP transfusionn=33

General characteristicsGender, male, % (n) 64 (21)Age (years) 61 (50-70)APACHE IV score 96 (79-128)SOFA score 11 (10-14)Medical historyPulmonary disease, % (n) 8 (3)Liver disease, % (n) 16 (6)Cardiac failure, % (n) 16 (6)Medical condition 24 hours before transfusionMechanical ventilation, % (n) 82 (27)Sepsis, % (n) 45 (15)Disseminated intravascular coagulation, % (n) 49 (16)

Data expressed as median and interquartile ranges

Table 2: Inflammatory cytokines in critically ill patients with a coagulopathy before and after a transfusion of fresh frozen plasma (12 ml/kg)

Before FFP After FFP p valuepro-inflammatory parameters (pg/ml)TNF-α 11.3 (2.3 – 52.3) 2.3 (2.3 – 41.0) 0.01IL1-β 15.0 (11.7 – 18.8) 14.4 (13.1 – 23.3) 0.97IL-8 178 (124 – 418) 187 (113 – 412) 0.23MCP1 1255 (503 – 3376) 1101 (434 – 5802) 0.89MIP1A 19.6 (15.7 – 33.6) 19.1 (13.3 – 34.4) 0.12sCD40L 409 (257 – 614) 324(216 – 537) 0.08anti-inflammatory parameters (pg/ml)IL-1RA 69.3 (52.1 – 110.6) 73.5 (47.8 – 104.9) 0.11IL-10 36.1 (15.5 – 100.1) 31.5 (14.8 – 279.6) 0.62

Data expressed as median (IQR)FFP = Fresh Frozen Plasma

cytokine levels. However, lung injury following transfusion is thought to be an early event. Also, we choose this early time point to minimize confounding by other fac-tors. Taken together, FFP does not appear to elicit an early inflammatory response.

Muller.indd 149 25-7-2014 14:40:28

Chapter 8

150

Figure 1: Effect of fresh frozen plasma transfusion (12 ml/kg) on markers of endothelial con-dition: ADAMTS13, Von Willebrandfactor, Factor VIII and Syndecan-1.

vWF = von WillebrandfactorFVIII = Factor VIII

Muller.indd 150 25-7-2014 14:40:29


151

Of interest, recent in vitro studies support an endothelial stabilizing role of FFP, as FFP reduced vascular endothelial cell permeability [22,33] and decreased expres-sion of endothelial adhesion markers [34] and endothelial white blood cell binding [22,34,35]. Effects of FFP were investigated in a rat hemorrhagic shock model, char-acterized by systemic shedding of syndecan-1, decreased syndecan-1 expression on pulmonary cells and increased pulmonary vascular permeability. Syndecan-1 is a proteoglycan on the luminal surface of endothelial cells which inhibits leukocyte adhesion during inflammation and is shed in case of endothelial damage [21]. Resus-citation with FFP was found to abrogate these effects, whereas resuscitation with crystalloids did not [22], associated with preservation of the endothelial glycocalyx [36] and improvement of lung injury [37].In trauma patients with hemorrhagic shock, syndecan-1 levels are also increased [20]. Studies of the effect of FFP on endothelial condition in patients are however lacking. Of note, recent evidence in trauma patients requiring a massive transfu-sion suggests that more and earlier administration of FFP decreases mortality [17,18,38,39]. This effect was not associated with improved coagulation ability, as the reduction in mortality in their study was irrespective of the admission INR [18] and coagulopathy does not seem to improve with higher amounts of FFP [19]. Given that FFP restores coagulation factors but also anti-coagulant proteins and that the net effect on hemostasis is unclear, FFP may exert protection via other mechanisms. This study found that FFP decreased levels of syndecan-1, associated with decreased levels of factor VIII, the latter which may also reflect improved endothelial condi-tion. These results support earlier experimental work indicating that FFP preserves endothelial integrity.The mechanism underlying this beneficial effect of FFP has not yet been described. We found that FFP transfusion was associated with an increase in ADAMTS13 and a decrease in vWF. Thereby, ADAMTS13 may have increased the ability to cleave large vWF multimers present on the activated endothelium. As large vWF multimers dam-age the endothelium, this effect may have preserved endothelial condition. This is also the rationale behind the treatment of thrombotic thrombocytopenic purpura by therapeutic plasma exchange.A protective effect of FFP on the endothelium is in apparent contrast with studies that have linked FFP to the occurrence of TRALI [4,6-8]. In an effort to reconcile

Muller.indd 151 25-7-2014 14:40:29

Chapter 8

152

these findings, we suggest that FFP associated with TRALI occurs as a result of an antibody-mediated pathogenesis. Indeed, efforts to reduce antibody positive blood products by male only policies, are associated with a significant reduction in TRALI [40]. In patients in whom transfusion is associated with lung injury in the absence of antibodies, other products such as red blood cells and platelets may be more impor-tant in inducing lung injury. Although dissecting these effects in multiple transfused patients is a challenge, future research should focus on differential effects of differ-ent blood products.This study is limited by a small and heterogeneous patient population. Thereby, some of the effects may be caused by chance or by regression to the mean. Findings need to be confirmed in a larger sample. Also, as mentioned, the timing of measure-ment may have been too early following FFP transfusion. We cannot exclude that FFP induces a pro-inflammatory host response later in time.In conclusion, this study is the first to describe the effect of a fixed dose of FFP trans-fusion in detail in critically ill patients. Results suggest that FFP stabilizes endothelial condition.

Muller.indd 152 25-7-2014 14:40:29


153

References



3. O’Shaughnessy DF, Atterbury C, Bolton Maggs P, et al.: Guidelines for the use of fresh-frozen plasma, cryoprecipitate and cryosupernatant. Br J Haematol 2004;126: 11-28





8. Watson GA, Sperry JL, Rosengart MR, et al.: Fresh frozen plasma is independently asso-ciated with a higher risk of multiple organ failure and acute respiratory distress syn-drome. J Trauma 2009;67: 221-7; discussion 8-30

9. Li G, Rachmale S, Kojicic M, et al.: Incidence and transfusion risk factors for transfusion-associated circulatory overload among medical intensive care unit patients. Transfusion 2011;51: 338-43

10. Narick C, Triulzi DJ, Yazer MH: Transfusion-associated circulatory overload after plasma transfusion. Transfusion 2012;52: 160-5

11. Johnson JL, Moore EE, Kashuk JL, et al.: Effect of blood products transfusion on the development of postinjury multiple organ failure. Arch Surg 2010;145: 973-7


13. Toy P, Gajic O, Bacchetti P, et al.: Transfusion-related acute lung injury: incidence and risk factors. Blood 2012;119: 1757-67

14. Vlaar AP, Hofstra JJ, Determann RM, et al.: Transfusion-related acute lung injury in car-diac surgery patients is characterized by pulmonary inflammation and coagulopathy: a prospective nested case-control study. Crit Care Med 2012;40: 2813-20

15. Letourneau PA, Pati S, Gerber MH, et al.: Fresh frozen plasma increases adhesion mol-ecule expression on human pulmonary endothelial cells. J Surg Res 2010;163: 317-22

16. Vlaar AP, Juffermans NP: Transfusion-related acute lung injury: a clinical review. Lancet 2013;382: 984-94

Muller.indd 153 25-7-2014 14:40:29

Chapter 8

154

17. Holcomb JB, Wade CE, Michalek JE, et al.: Increased plasma and platelet to red blood cell ratios improves outcome in 466 massively transfused civilian trauma patients. Ann Surg 2008;248: 447-58

18. Brown LM, Aro SO, Cohen MJ, et al.: A high fresh frozen plasma: packed red blood cell transfusion ratio decreases mortality in all massively transfused trauma patients regardless of admission international normalized ratio. J Trauma 2011;71: S358-63

19. Chambers LA, Chow SJ, Shaffer LE: Frequency and characteristics of coagulopathy in trauma patients treated with a low- or high-plasma-content massive transfusion proto-col. Am J Clin Pathol 2011;136: 364-70

20. Haywood-Watson RJ, Holcomb JB, Gonzalez EA, et al.: Modulation of syndecan-1 shed-ding after hemorrhagic shock and resuscitation. PLoS One 2011;6: e23530

21. Hayashida K, Parks WC, Park PW: Syndecan-1 shedding facilitates the resolution of neutrophilic inflammation by removing sequestered CXC chemokines. Blood 2009;114: 3033-43

22. Peng Z, Pati S, Potter D, et al.: Fresh frozen plasma lessens pulmonary endothelial inflammation and hyperpermeability after hemorrhagic shock and is associated with loss of syndecan 1. Shock 2013;40: 195-202

23. Steppan J, Hofer S, Funke B, et al.: Sepsis and major abdominal surgery lead to flaking of the endothelial glycocalix. J Surg Res 2011;165: 136-41

24. Ostrowski SR, Berg RM, Windelov NA, et al.: Coagulopathy, catecholamines, and bio-markers of endothelial damage in experimental human endotoxemia and in patients with severe sepsis: a prospective study. J Crit Care 2013;28: 586-96


26. Bone RC, Sibbald WJ, Sprung CL: The ACCP-SCCM consensus conference on sepsis and organ failure. Chest 1992;101: 1481-3


28. Kokame K, Nobe Y, Kokubo Y, et al.: FRETS-VWF73, a first fluorogenic substrate for ADAMTS13 assay. Br J Haematol 2005;129: 93-100


30. Shorr AF, Thomas SJ, Alkins SA, et al.: D-dimer correlates with proinflammatory cytokine levels and outcomes in critically ill patients. Chest 2002;121: 1262-8

31. Urner M, Herrmann IK, Buddeberg F, et al.: Effects of blood products on inflammatory response in endothelial cells in vitro. PLoS One 2012;7: e33403

32. Schneider SO, Rensing H, Graber S, et al.: Impact of platelets and fresh frozen plasma in contrast to red cell concentrate on unstimulated and stimulated cytokine release in an in vitro model of transfusion. Scand J Immunol 2009;70: 101-5

Muller.indd 154 25-7-2014 14:40:29


155

33. Pati S, Matijevic N, Doursout MF, et al.: Protective effects of fresh frozen plasma on vas-cular endothelial permeability, coagulation, and resuscitation after hemorrhagic shock are time dependent and diminish between days 0 and 5 after thaw. J Trauma 2011;69 Suppl 1: S55-63

34. Nohe B, Kiefer RT, Ploppa A, et al.: The effects of fresh frozen plasma on neutrophil-endothelial interactions. Anesth Analg 2003;97: 216-21

35. Wataha K, Menge T, Deng X, et al.: Spray-dried plasma and fresh frozen plasma modu-late permeability and inflammation in vitro in vascular endothelial cells. Transfusion 2013;53 Suppl 1: 80S-90S


37. Kozar RA, Peng Z, Zhang R, et al.: Plasma restoration of endothelial glycocalyx in a rodent model of hemorrhagic shock. Anesth Analg 2011;112: 1289-95

38. Borgman MA, Spinella PC, Perkins JG, et al.: The ratio of blood products transfused affects mortality in patients receiving massive transfusions at a combat support hospi-tal. J Trauma 2007;63: 805-13

39. Sperry JL, Ochoa JB, Gunn SR, et al.: An FFP:PRBC transfusion ratio >/=1:1.5 is associated with a lower risk of mortality after massive transfusion. J Trauma 2008;65: 986-93

40. Arinsburg SA, Skerrett DL, Karp JK, et al.: Conversion to low transfusion-related acute lung injury (TRALI)-risk plasma significantly reduces TRALI. Transfusion 2012;52: 946-52

Muller.indd 155 25-7-2014 14:40:29

Muller.indd 156 25-7-2014 14:40:29

Chapter 9

Clinicians’ attitude towards prophylactic

transfusion of fresh frozen plasma – an evaluation

of a multicenter randomized clinical trial

Marcella C.A. Müller, Rob J. de Haan, Margreeth B. Vroom, Nicole P. Juffermans


Muller.indd 157 25-7-2014 14:40:29

Chapter 9

158

Abstract

Background: Prophylactic use of Fresh Frozen Plasma (FFP) to prevent bleeding in critically ill patients with a coagulopathy is widespread practice, while evidence of efficacy is lacking and detrimental effects of FFP in this population have clearly been demonstrated. This has led to a call for randomized trials on the risks and benefit of FFP in the non-bleeding critically ill patients with a coagulopathy. However, to date, conducting a trial on this subject has not been successful. Recently a multicenter randomized trial was stopped prematurely due to slow inclusion.Aim: To examine clinicians’ attitude towards correction of coagulopathy with FFP in non-bleeding critically ill patients undergoing an intervention and to assess clini-cians’ opinions regarding a trial on prophylactic administration of FFP in critically ill patients with a coagulopathy prior to undergoing an intervention.Methods: A survey in 55 critical care physicians working at 4 intensive care depart-ments which participated in a randomized trial on the risks and benefits of FFP in critically ill patients.Results: Response rate was 84%. Of all respondents, 61% stated a need to assess International Normalized Ratio (INR) before placement of a central venous catheter. Before insertion of a chest tube (89%) or tracheostomy (91%), nearly all respondents indicated that INR should be assessed. Reasons to withhold transfusion of FFP to non-bleeding critically ill patients are risk of TRALI (46%), fluid overload (39%) and allergic reaction (24%). Although the majority of respondents expressed the opinion that the trial was clinically relevant and had a clear protocol, up to 56% of them indicated that one or more patient subgroups should have been excluded from par-ticipating in a trial on prophylactic administration of FFP.Conclusion: Critical care physicians express the need for more evidence on the pro-phylactic use of FFP in critically ill patients with a coagulopathy. However, the major-ity expressed reluctance to include patients in a trial on the effectiveness of FFP due to their personal beliefs about the preferable transfusion strategy in certain patient categories. Results suggest that trials on FFP, at least in the Netherlands, are not feasible.

Muller.indd 158 25-7-2014 14:40:29

Clinicians' attitude towards FFP transfusion in the critically ill

159

Background

The prevalence of coagulopathy as detected by prolonged coagulation tests in criti-cally ill patients is high [1]. In order to prevent bleeding complications, fresh frozen plasma (FFP) is frequently administered prophylactically to critically ill patients with prolonged prothrombin times (PT) or elevated International Normalized Ratio (INR) values [2-4]. However, evidence that prophylactic transfusion of FFP reduces bleed-ing risk in these patients is limited [5,6]. On the other hand, it has well been demon-strated that FFP transfusion in the critically ill contributes to adverse outcome [3,7]. This has resulted in a call for randomized trials evaluating the benefit of prophylactic FFP in patients with a coagulopathy [8-12]. However, to date only small trials [13,14], cohort studies [15] or retrospective studies have been reported [16-18].Conducting a successful trial on the efficacy of FFP transfusion in critically ill patients is likely to encounter several difficulties. First, critically ill patients are often not able to give consent to participate in a trial [19] and permission from a legal repre-sentative is warranted. However, most substitute-decision makers are unaware of patients wishes regarding research and are anxious to make a decision to participate or not [20]. Furthermore, time is limited in decision making about participation in a trial conducted on the Intensive Care Unit (ICU). Besides issues on informed consent there may be factors related to clinical practice which hamper performing transfu-sion trials. It has been reported that clinicians have strong beliefs about the ability of FFP to influence bleeding risk in coagulopathic patients [21-23]. Altogether, this may have contributed to a lack of successfully conducted large randomized trials on the effectiveness of prophylactic FFP transfusion in critically ill patients. In line with this, results of two randomized trials have not been completed (NCT00953901) or published [24].We carried out a multicenter randomized clinical trial to assess risks and benefits of FFP to prevent bleeding in coagulopathic critically ill patients undergoing an invasive procedure (TOPIC trial) [25]. However, despite a proactive screening approach at all trial sites, the trial was stopped early due to slow inclusion. In order to give direc-tions that help to improve conduction and subject enrolment in future transfusion trials, we conducted a survey among the ICU physicians who participated in this trial.

Muller.indd 159 25-7-2014 14:40:29

Chapter 9

160

Methods

SettingIn June 2013, a questionnaire was sent to staff physicians and fellows of the ICU departments of two university hospitals and two large teaching hospitals in the Netherlands just after ending enrolment of patients in the TOPIC trial. The TOPIC trial randomized critically ill patients with a coagulopathy prior to undergoing an intervention to administration of a fixed dose of FFP (12 ml/kg) or to no transfusion [25]. At each research site, a local investigator was responsible for active screening of eligible patients, and at two sites, research nurses supported the trial. Medical staff of participating sites was informed and updated on relevant aspects of the trial by regular presentations by the main or local investigator and a digital newslet-ter. All involved medical staff members were provided with pocket cards containing essential information of the trial and for inclusion and randomization 24/7 phone assistance was available.

SurveyThe questionnaire comprised multiple-choice questions regarding the need to assess coagulation status in non-bleeding patients, factors influencing the decision to administer FFP, factors affecting patient inclusion, and overall evaluation of our TOPIC trial. In addition, data on hospital setting, previous training and age of partici-pants were collected.

Statistical analysisDescriptive statistics were used to summarize respondents’ characteristics. Cate-gorical data were presented in percentages and analyses were carried out with SPSS version 20.0 (SPSS, Inc, Chicago IL, USA) and Prism version 5.0 (Graphpad Software, San Diego, USA).

Results

A fully completed survey was obtained in 46 of 55 participants (84%). The majority of respondents worked at an ICU in a university hospital, however other character-

Muller.indd 160 25-7-2014 14:40:29


161

istics such as medical specialty and age did not differ from previously reported sur-veys among Dutch critical care physicians [26]. Characteristics of the respondents are summarized in table 1.

Need to assess coagulation statusThe most commonly performed invasive procedures in critically ill are placement of a central venous catheter (CVC), a chest tube, a tracheotomy and drainage of fluid/abscess collections. 61% of respondents indicated that INR should be assessed before placement of a CVC. For placement of a chest tube or tracheostomy, nearly all respondents indicated that INR should be assessed (89% and 91% respectively to CVC). According to 75% of the respondents, there is a need to assess INR before fluid or abscess drainage (figure 1).

Decision to administer FFP and inclusion of patientsRespondents were asked to indicate which factors determine the decision to pro-phylactically administer FFP prior to an invasive procedure. Reducing risk of bleeding is the main reason to administer FFP prophylactically. However, risk of bleeding was a concern in only a minority of respondents (15%). In addition, 60% indicated to have no concerns about the occurrence of bleeding complications after an intervention.

Table 1: Characteristics of respondents

N=46Medical specialtyMedicineAnesthesiologyCardiology

62%28%10%

QualificationCritical care specialistFellow training in intensive care

72%28%

Type of ICUUniversity hospitalTeaching hospital

83%17%

Age20-35 year36-50 year51-65 year

24%63%13%

SexMale 65%

Muller.indd 161 25-7-2014 14:40:29

Chapter 9

162

Figure 1: Percentage of respondents indicating that INR should be assessed before place-ment of a central venous catheter, chest tube, tracheostomy or abscess or fluid drainage.

CVC = central venous catheter

Figure 2: Factors reported by respondents to omit transfusion of fresh frozen plasma in criti-cally ill patients with a coagulopathy needing to undergo an intervention.

TRALI = transfusion related acute lung injury

Muller.indd 162 25-7-2014 14:40:30


163

Factors determining the decision to withhold FFP were more frequently reported. The risk of inducing a transfusion related acute lung injury (TRALI) was the most frequently reported factor to omit administering FFP. Participants also reported risk of fluid overload, an allergic reaction or loss of coagulation status as a marker of disease severity as reasons not to administer FFP before an intervention in non-bleeding critically ill patients with a coagulopathy (figure 2).

Overall evaluation of the TOPIC trialThe majority of the respondents had no objection towards the conduct of research in incapacitated patients (76%) or intervention studies in critically ill patients (91%). Nearly all physicians (98%) indicated that it was ethically justified to study the effect of FFP on prevention of bleeding and correction of INR in critically ill patients. Clini-cians indicated to have a need for more evidence with respect to the treatment of coagulopathy (80%) and administration of FFP in critically ill (76%). In addition, the majority judged the trial to be well designed with a clear protocol and 87% stated that the chosen endpoint of the trial, major bleeding complication after an interven-tion, was clinically relevant.Of respondents, 89% had no objection towards requesting informed consent of criti-cally ill patients or relatives. However, 15% considered asking patients consent to be complex for the TOPIC trial, moreover 24% indicated that asking consent of patients’ relatives was complex. Indeed, declined consent rate for the TOPIC trial was 25%.Three quarters of participants indicated that the chosen range of INR (1.5-3.0) for inclusion was appropriate. However, 30% classified a FFP dose of 12 ml/kg as too high, whereas 11% indicated that the trial had interference with prevailing treat-ment protocols. In contrast to these answers, only 41% stated that the trial was widely supported by the medical staff in their department. Furthermore, the major-ity of respondents had the opinion that certain patient groups should have been excluded from the trial (table 2). Altogether, 56% of respondents indicated one or more clinical conditions to exclude a patient from participating in the TOPIC trial.

Muller.indd 163 25-7-2014 14:40:30

Chapter 9

164

Discussion

The current survey demonstrates that the majority of critical care physicians requires assessment of INR before commonly performed invasive procedures in critically ill patients. In addition, they indicated a need for more evidence regarding the treatment of coagulopathy and administration of prophylactic FFP. Neverthe-less, a randomized trial on the efficacy of prophylactic FFP transfusion, which was judged to be ethical and well designed, was not widely supported by the medical staff of participating centers. Moreover, the majority of physicians were reluctant to include all patients in this trial, indicating a lack of knowledge and the presence of strong personal beliefs regarding FFP administration in the critically ill. These per-sonal beliefs were major contributors to a lack of sufficient inclusion in a trial on the efficacy of prophylactic FFP transfusion.Although the majority of respondents deemed it necessary to obtain an INR before performing an invasive procedure in a critically ill patient, evidence supporting that increased INR or PT values predict bleeding risk is scarce [27]. Despite the limited value of INR to predict bleeding risk, increased values are an important trigger for clinicians to administer FFP [21]. In the current survey, respondents indicated that the TOPIC trial was clinically relevant. This is in line with the call in the medical com-munity for trials on this subject [6,8]. Moreover, the design of the TOPIC trial was in line with a proposal done by a committee of experts on clinical trial opportunities in transfusion medicine [28]. Therefore, the TOPIC trial was expected to be widely supported by involved physicians.

Table 2: Reasons to exclude a patient from participating in a trial on the effectiveness of fresh frozen plasma.

Medical condition Percentage indicating that patient should be excluded

Heart failure 24State of fluid overload 33Liver failure 24Disseminated intravascular coagulation 33Reduced P/F ratio 13

P/F = PaO2 divided by the fraction of inhaled oxygen

Muller.indd 164 25-7-2014 14:40:30


165

The survey indicated several barriers for inclusion of patients in the TOPIC trial. A general well-known problem is the difficulty of obtaining informed consent in critically ill patients. A substantial amount of respondents indicated that obtaining informed consent of substitute decision maker was complex. In a quarter of eligi-ble patients, consent was declined which is in line with other intervention studies in critically ill patients [29]. However, a more important barrier for patient inclu-sion in our FFP trial turned out to be physician-driven. Physicians involved in the TOPIC trial indicated to be reluctant to include certain patients. The reasons were bidirectional. Physicians who wanted to correct increased INR with FFP before the planned invasive procedure did not want to take the risk of randomization to no FFP, although this concerned a minority. The majority of physicians felt no need for FFP transfusion and did not want to risk having to administer FFP. Nearly half of the surveyed physicians indicated to be concerned about possible detrimental effects of FFP, with TRALI being the most frequently mentioned. This is consistent with reported increased awareness of TRALI [30]. Reluctance to include patients was also reflected by a relatively low percentage reporting support by the medical staff for the trial. Taken together, despite the recognition that a trial on the efficacy of FFP is highly warranted, reasons for slow inclusion with subsequent early stopping the TOPIC trial, are, at least in part, physician-driven.The strong beliefs among physicians about the administration of FFP are in line with surveys reporting high rates of inappropriate use of FFP [22,23]. Moreover, in line with previous reports [22], involved clinicians demonstrate a lack of knowledge about appropriate dosing of FFP, as 30% indicated that the dose of 12 ml/kg FFP used in the TOPIC trial was too high, while the recommended dose is 12 to 15 ml/kg [31] or even as high as 35 ml/kg to obtain full normalization of factor levels [13]. When designing future trials on the efficacy of FFP in critically ill patients with a coagulopathy, abovementioned obstacles need to be addressed in order to obtain improved cooperation of involved ICU physicians.The current survey has some limitations. First, the questionnaire was not formally tested for reliability and validity. Furthermore, we did not address potential improve-ments of the design and protocol of the TOPIC study. Modifiable factors in enhanc-ing enrolment of critically ill patients in clinical trials are research culture, support of the clinical team and the consent procedure [29]. Also, it has been reported

Muller.indd 165 25-7-2014 14:40:30

Chapter 9

166

that passive dissemination of research information does not affect clinicians’ atti-tudes [32]. Therefore, we actively informed medical staff at all sites throughout the whole study period by repeated presentations, digital newsletters and pocket cards. Although we did not ask respondents for potential improvements on this informa-tion strategy, the majority indicated that enrolment criteria of the trial were clear. Therefore, we do not think that a lack of information or knowledge about the trial by involved physicians has contributed to the disappointing inclusion rates. Of note, in line with previous reports [29], the presence of dedicated experienced research personnel contributed to improved enrolment of subjects and minimized potential additional workload for the bedside staff. In our opinion, exclusion of certain patient categories, although desired by the respondents, should be avoided because this will limit generalizability of results. However, in future trials these barriers should be specifically addressed, so they can be overcome.

Conclusion

Despite that critical care physicians express a need for trials on the use of FFP in patients with a coagulopathy, strong preferences of treating physicians about the use of FFP were major constraints to the conduction a successful clinical trial. We feel that results indicate that future trials on the efficacy of FFP may not be feasible to conduct, at least in the Netherlands. Thereby, further knowledge on the efficacy of FFP should be gathered from prospective cohort studies or retrospective studies.

Muller.indd 166 25-7-2014 14:40:30


167

References






6. Yang L, Stanworth S, Hopewell S, et al.: Is fresh-frozen plasma clinically effective? An update of a systematic review of randomized controlled trials. Transfusion 2012;52: 1673-86



9. Tinmouth AT, McIntyre L: The conundrum of persistent inappropriate use of frozen plasma. Crit Care 2011;15: 160

10. Desborough M, Stanworth S: Plasma transfusion for bedside, radiologically guided, and operating room invasive procedures. Transfusion 2012;52 Suppl 1: 20S-9S

11. Murad MH, Stubbs JR, Gandhi MJ, et al.: The effect of plasma transfusion on morbidity and mortality: a systematic review and meta-analysis. Transfusion 2010;50: 1370-83



14. Veelo DP, Vlaar AP, Dongelmans DA, et al.: Correction of subclinical coagulation disor-ders before percutaneous dilatational tracheotomy. Blood Transfus 2012;10: 213-20


16. Carino GP, Tsapenko AV, Sweeney JD: Central line placement in patients with and with-out prophylactic plasma. J Crit Care 2012;27: 529 e9-13


Muller.indd 167 25-7-2014 14:40:30

Chapter 9

168

18. Davis JW, Davis IC, Bennink LD, et al.: Placement of intracranial pressure monitors: are “normal” coagulation parameters necessary? J Trauma 2004;57: 1173-7

19. Fan E, Shahid S, Kondreddi VP, et al.: Informed consent in the critically ill: a two-step approach incorporating delirium screening. Crit Care Med 2008;36: 94-9

20. Chenaud C, Merlani P, Verdon M, et al.: Who should consent for research in adult intensive care? Preferences of patients and their relatives: a pilot study. J Med Ethics 2009;35: 709-12






26. Vlaar AP, Honselaar WB, Binnekade JM, et al.: Diagnosing acute lung injury in the critically ill: a national survey among critical care physicians. Acta AnaesthesiolScand 2009;53: 1293-9


28. Blajchman MA, Glynn SA, Josephson CD, et al.: Clinical trial opportunities in Transfu-sion Medicine: proceedings of a National Heart, Lung, and Blood Institute State-of-the-Science Symposium. Transfus Med Rev 2010;24: 259-85

29. Smith OM, McDonald E, Zytaruk N, et al.: Enhancing the informed consent process for critical care research: Strategies from a thromboprophylaxis trial. Intensive Crit Care Nurs 2013;29: 300-9



32. Dale C, Fowler RA, Adhikari NK, et al.: Implementation of a research awareness pro-gram in the critical care unit: effects on families and clinicians. Intensive Crit Care Nurs 2010;26: 69-74

Muller.indd 168 25-7-2014 14:40:30

Part IIIRisks of transfusion

in the criti cally ill: TRALI

Muller.indd 169 25-7-2014 14:40:30

Muller.indd 170 25-7-2014 14:40:30

Chapter 10

Risk factors for transfusion-related

acute lung injury in ICU patients

Marcella C.A. Müller, Nicole P. Juffermans

Annual Update in Intensive Care and Emergency Medicine 2013; 527-535

Muller.indd 171 25-7-2014 14:40:30

Chapter 10

172

Introduction

Multiple observational studies have shown an association between transfusion and lung injury in the critically ill [1-6]. The definition of transfusion-related acute lung injury (TRALI) is the onset of acute lung injury (ALI) occurring within 6 hours follow-ing a blood transfusion. When other risk factors for ALI are present, the term ‘pos-sible TRALI’ is used [7,8]. Although a temporal relation with a blood transfusion is an obvious requirement for the definition of TRALI, the time frame of 6 hours is arbi-trarily chosen. Indeed, it has been recognized that transfusion can result in delayed respiratory complications, termed ‘delayed TRALI syndrome’ [9].Recent observational studies suggest that the incidence of TRALI is high in several critically ill risk populations. Cohort studies in critically ill patients report a 100-fold higher incidence of TRALI when compared to the general hospital population [10,11]. Of note, the attributive morbidity and mortality of TRALI is considerable. Up to 70% of patients who develop TRALI needs referral to the intensive care unit (ICU) and invasive mechanical ventilation [12]. When TRALI develops in patients already admitted to the ICU, there is an association with prolonged mechanical ventilation, increased hospital mortality and a decreased long-term survival [10,11,13]. Progno-sis of the ‘delayed TRALI syndrome’ is particularly bad, with an estimated mortality of 40% [9].Obviously, blood transfusion cannot be avoided altogether. The high incidence and considerable morbidity and mortality of TRALI warrant insight in the risks and be nefits of the decision to administer a blood transfusion. This chapter aims to address the susceptibility of critically ill patients to the development of a TRALI reac-tion. Risk factors for critically ill patients to develop TRALI as well product-related risk factors will be discussed. Efforts to reduce TRALI have focused to date on modi-fication of blood products, which has resulted in a substantial reduction in TRALI cases, but has not completely mitigated TRALI. In this chapter, we suggest that com-plete mitigation requires physicians to develop new approaches for patient-specific transfusion practice, aimed at pre-emptively taking action to decrease the risk of TRALI. It may be time to improve our assessment of the critically ill patient in need of a transfusion.

Muller.indd 172 25-7-2014 14:40:30

Risk factors for TRALI in ICU patients

173

Pathogenesis of TRALI

The high incidence of TRALI reported in critically ill patient populations may be a consequence of its pathogenesis. TRALI is mediated by the interaction of neutro-phils with pulmonary endothelial cells and can be caused by any cell-containing or plasma rich blood product. TRALI is thought to occur as a result of a ‘two hit’ insult [14,15]. The first ‘hit’ is an inflammatory condition of the patient at the time of the transfusion (e.g. sepsis, recent surgery), causing sequestration and priming of neu-trophils and the expression of adhesion molecules in the pulmonary compartment. In response to priming, neutrophils undergo changes allowing for close contact with the endothelial surface of the pulmonary vasculature. The second ‘hit’ is mediated by components in the transfusion product. Donor antibodies against human neu-trophil antigens (HNA) or human leukocyte antigens (HLA) present on leukocytes in the lungs of the recipient have been implicated. Also bioactive substances that have accumulated during blood storage may play a role in TRALI. These mediators are thought to stimulate the primed neutrophils to release proteases resulting in endothelial damage, vascular permeability and the full clinical picture of TRALI [16,17].Findings from experimental models support the assumption that in TRALI, neutro-phils are primed by inflammation. Priming with a ‘first hit’ of endotoxin (LPS) was found to be a prerequisite for induction of TRALI using plasma from stored blood products [16-19]. In the presence of a priming ‘hit’, it was demonstrated that lower amounts of antibody were sufficient to elicit the TRALI reaction [20], which supports the notion that the presence of an inflammatory response increases susceptibility to TRALI. Thereby, both transfusion and patient-related factors play a role in TRALI pathogenesis.

Blood product-related risk factors for TRALI

Anti-leukocyte antibodiesFrom case series, the presence of donor anti-leukocyte antibodies in the trans-fused product is thought to be implicated in TRALI. Involved antibodies are mainly directed against HLA Class I, HLA Class II or HNA. Activation of neutrophils occurs

Muller.indd 173 25-7-2014 14:40:30

Chapter 10

174

via HLA Class I antigens. Mononuclear cells expressing HLA Class II are involved by activation through HLA Class II antibodies [21]. Recently, a large, prospective, case-controlled multicenter study systematically identified factors associated with an increased risk of TRALI [22]. In this study, blood product-related risk factors were the transfused volume of high titer cognate HLA class II antibody and the volume of high titer HNA antibody. Antibodies not associated with increased risk of TRALI were non-cognate or weak anti-HLA class II antibody, as well as HLA class I antibodies. The finding that the presence of HLA Class II antibodies in plasma of blood donors was of greater importance then class I, underlines results from previous studies [23]. Antibodies against the HNA-3a or 5b antigen appear to be especially important in causing severe cases of TRALI [24].

Bioactive response modifiersBioactive lipids (lysoPCs and neutral lipids) increase during storage of red blood cells (RBCs) and platelet concentrates [25,26]. These substances were shown to have in vitro neutrophil priming capacity and can induce TRALI in animal models [16-18]. A retrospective clinical study of 10 TRALI patients linked the occurrence of TRALI with transfusion of blood products containing lipids with neutrophil-priming activity [27]. However, this finding was not confirmed in a study in cardiac surgery patients developing TRALI [28]. Also, in the largest prospective case-control study to date, no association was found between bioactive lipids and TRALI [22]. Other bioactive substances, such as pro-inflammatory cytokines, have not been consistently linked to TRALI [22].

The red cell storage lesionDuring storage, RBC products undergo changes that affect their function and in vivo survival, which are collectively termed the “storage lesion”. In vitro experiments have shown increased reactive oxygen species (ROS) formation by endothelial cells upon incubation with aged red blood cells but not with fresh cells [29]. Indeed, in a murine “two hit” transfusion model, aged RBCs were capable of inducing a TRALI reaction [30]. The proposed mechanism is either a decrease in the chemokine scav-enging function of the RBC [31] or increased adhesion of the erythrocyte to the endothelium due to a deficiency of anti-adhesive adenosine-5’-triphosphate release

Muller.indd 174 25-7-2014 14:40:30


175

from the RBC [32]. Hereby, transfused RBCs may promote or exacerbate microvas-cular pathophysiology in the lung.

In conclusion, cognate HLA class II antibodies are blood product-related risk factors for TRALI. Bioactive lipids can function as a second ‘hit’ and are able to induce a TRALI reaction, but clinical data are conflicting. The aged RBC seems to play a role in preclinical models of TRALI. To date, interventions aimed at decreasing TRALI inci-dence have focused on product-related risk factors.

Mitigating TRALI by modifying blood products

Exclusion of female donorsThe prevalence of anti-leukocyte antibodies in the donor population depends on donor allo-exposure. Transfusion can lead to sensitization of recipients, albeit with very different reported numbers, ranging from 1 to 12% [33]. However, pregnancy is the most important cause of sensitization in the donor population. Approximately 10% of previously pregnant women have HLA antibodies and this number increases to 26-39% in women who have had three or more pregnancies [33]. In order to pre-vent antibody mediated TRALI, a predominantly male donor strategy for prepara-tion of plasma rich products (including fresh frozen plasma and buffy coat-derived pool platelets) was first implemented in the United Kingdom in 2003. Other coun-tries followed. This resulted in a ~70% reduction in reported TRALI cases [33-37]. Hence, it seems that exclusion of female plasma donors prevents the majority of TRALI cases caused by plasma-rich products. However, TRALI cases that occur due to RBC transfusion and pooled platelets are not prevented.

Transfusion of fresh red blood cells onlyThe accumulation of compounds, such as bioactive lipids, during storage suggests that the occurrence of TRALI may be associated with the use of older blood. A meta-analysis of studies on the effect of transfusion of stored blood showed an association between transfusion of stored blood and mortality in various patient populations [38]. In cardiac surgery patients, an association of prolonged storage of blood prod-ucts and lung injury was found [39]. However, a large prospective study did not find

Muller.indd 175 25-7-2014 14:40:30

Chapter 10

176

evidence for an association between red cell storage duration and TRALI [22]. In a prospective case-control study in ICU patients, stored red blood cell products did not have any effect on pulmonary function or gas exchange [40]. Prospective rand-omized controlled clinical trials investigating the effect of fresh compared to stored RBCs on outcome in the critically ill are currently underway.

Patient-related risk factors for TRALI

Predisposing factors for TRALI presumably lower the recipient’s threshold for devel-oping TRALI. Specific host factors have recently been identified (table 1). In a mouse model of TRALI, mechanical ventilation synergistically augmented lung injury, which was even further enhanced by the use of injurious ventilator settings [41], support-ing the concept that mechanical ventilation aggravates the course of a TRALI reac-tion. There are also clinical data suggesting that mechanical ventilation may be a risk factor for lung injury following blood transfusion. As much as 33% of mechanically ventilated critically ill patients develop lung injury within 48 hours after transfu-sion in an observational study [3]. In accordance, it was found that the presence of mechanical ventilation predisposed to acquiring TRALI in a retrospective study in ICU patients [11]. Furthermore, a recent study confirmed that high peak airway pres-sures (>30 cm H2O) contributed to an increased TRALI risk (OR 5.6 [2.1-14.9]) [22].In addition to pulmonary hits, systemic inflammatory conditions are often present in critically ill patients. In a retrospective study on risk factors for TRALI in ICU patients, 87% of the patients that confirmed to the diagnostic criteria for TRALI had a risk factor for ALI prior to onset of (possible) TRALI [11]. Hence, it is not surprising that there is an overlap between risk factors for ALI and TRALI. Of note, comparable to possible TRALI, ALI is a syndrome which occurs as a complication of an inflammatory condition.Sepsis has been identified as a risk factor for TRALI in several studies in ICU patients [10,11]. This is in line with animal models showing that endotoxemia reduceds the amount of antibodies or bioactive response modifiers needed to induce TRALI [16,20]. The presence of shock prior to transfusion increased TRALI risk [22]. Of interest, plasma interleukin (IL)-8 levels in TRALI patients on a general hospital ward were elevated prior to transfusion, underlining the contribution of an inflammatory

Muller.indd 176 25-7-2014 14:40:31


177

status to the risk of developing TRALI [22]. In cardiac surgery patients prospectively followed for the onset of TRALI, IL-8 levels prior to transfusion were elevated com-pared to transfused controls [42]. Because IL-8 has neutrophil-priming capacity, we hypothesize that during an insult, endothelial cells produce IL-8, contributing to attraction of neutrophils to the pulmonary compartment [42]. Whether IL-8 can serve as a biomarker for diagnosing TRALI remains to be determined.Coronary artery bypass grafting (CABG) was found to be a risk factor for TRALI [11,43]. In this context, the incidence of TRALI was found to be 2.4% in a prospective study in cardiac surgery patients [28], which is about 8 to 10-fold higher compared to the general patient population. We speculate that this may be due to a longer time on cardiopulmonary bypass (CPB) [28]. Hematologic malignancy is also a TRALI risk factor in a general hospital patient populations as well as in ICU patients [11,43]. Massive transfusion is a risk factor for acute respiratory distress syndrome (ARDS) [1,4] as well as for TRALI [11]. In patients admitted to the ICU because of bleeding, the presence of liver failure was a strong risk factor for developing TRALI when com-pared to transfused patients without liver failure [13]. It is not clear whether these conditions are risk factors solely because these patients commonly receive multi-ple transfusions, or that fluid overload induces priming. Of note, a prospective case control study in TRALI patients in the general hospital population revealed that an increased TRALI risk was associated with a positive fluid balance [22]. This is in line with findings in ALI patients [44] and indeed suggests that fluid overload may play role in TRALI pathogenesis.Taken together, critical illness contributes to increased susceptibility to a TRALI reaction. Of note, in a multivariate analysis of a retrospective study in ICU patients, patient-related risk factors were more important for the onset of TRALI than transfu-sion-related risk factors, suggesting that development of a TRALI reaction depends more on host factors then on factors in the blood product [11]. This finding suggests that taking an active approach in daily ICU practice may be effective in mitigating the risk of TRALI.

Muller.indd 177 25-7-2014 14:40:31

Chapter 10

178

Table 1: Patient-related risk factors for transfusion-related acute lung injury (TRALI)

Author, date [ref] Type of study Relation between risk factor and TRALI

Sepsis Rana, 2006 [46] Retrospective case control

Sepsis in 48% TRALI patients (p<0.01 vs. controls)

Gajic, 2007 [10] Prospective case control

Sepsis in 37% patients developing lung injury after transfusion (p=0.016 vs. controls)

Vlaar, 2010 [11] Retrospective cohort

Sepsis predisposes to TRALI (OR 2.5 [1.2-5.2])

Shock Toy, 2012 [22] Prospectivecase control

Shock prior to transfusion predis-poses to TRALI(OR 4.2 [1.7- 10.6])

Mechanical Ventilation

Gajic, 2004 [3] Retrospective cohort

33% of ventilated patients develop lung injury within 48 hours after transfusion


Mechanical ventilation predisposes to TRALI (OR 3.0 [1.3-7.1])

Toy, 2012 [22] Prospective case control

Peak airway pressure of >30 cm H2O predisposes to TRALI (OR 5.6 [2.1-14.9])

Cardiac surgery


Emergency cardiac surgery predis-poses to TRALI (OR 17.6 [1.8-168.5])


Cardiac surgery predisposes to TRALI (OR 3.3 [1.21-9.2])

Vlaar, 2011 [28] Prospective case control

Time on cardiopulmonary bypass predisposes to TRALI (OR 1.0 [1.0-1.03])

Hematologic malignancy

Rana, 2006 [46] Retrospective case control

Hematologic malignancy in 25% TRALI patients (p<0.01 vs. controls)


Hematologic malignancy predisposes to TRALI (OR 13.1 [2.7-63.8])

Silliman, 2003 [43] Retrospective cohort

Hematologic malignancy in 46% of TRALI patients (p=0.0004 vs. trans-fused controls)

Massive transfusion


Massive transfusion predisposes to TRALI (OR 4.5 [2.1-9.8])

Positive fluid balance


Fluid balance, increment per liter, predisposes to TRALI (OR 1.17 [1.08-1.28], p<0.001)

Liver failure Benson, 2010 [13] Retrospective cohort

Liver failure predisposes to TRALI (OR 13.1 [2.7-63.8])


Liver transplant surgery predisposes to TRALI (OR 6.7 [1.3-35.7])

Data are percentages or odds ratio (OR) with confidence interval

Muller.indd 178 25-7-2014 14:40:31


179

Mitigating TRALI by individualized patient intervention

Appropriate management of critically ill patients has decreased their risk of develop-ing ALI. The same may hold true for TRALI. Given the high number of ICU patients who receive a blood transfusion, the association between blood transfusion and adverse outcome in the critically ill [10,11,13] and the recent identification of TRALI risk factors, it is possible and perhaps even mandatory for ICU physicians to take an individualized approach towards their patients in need of a transfusion. Patient-focused strategies that may decrease the risk of TRALI include: monitoring fluid bal-ance; shock prior to transfusion should be avoided; a restrictive fluid balance should be maintained. Decreasing airway pressures in patients on mechanical ventilation prior to transfusion may also decrease the risk of TRALI. Although not proven in clini-cal trials, low tidal volume ventilation may further reduce TRALI risk. With the use of electronic medical records, protocols can be developed to further decrease the risk of TRALI. An electronic screening algorithm has been developed which accurately identifies TRALI patients [45]. Improvement in identification may contribute to bet-ter patient management of a TRALI case. Electronic health record surveillance may also be a tool for further implementing measures to decrease TRALI.

Conclusion

The presence of an inflammatory condition increases the susceptibility to TRALI. Specific patient-related risk factors have been identified, empowering the ICU physi-cian to take an individualized approach to the patient in need of a transfusion, which may contribute to mitigating the risk of TRALI.

Muller.indd 179 25-7-2014 14:40:31

Chapter 10

180

References

1. Croce MA, Tolley EA, Claridge JA, et al.: Transfusions result in pulmonary morbidity and death after a moderate degree of injury. J Trauma 2005;59: 19-23



4. Gong MN, Thompson BT, Williams P, et al.: Clinical predictors of and mortality in acute respiratory distress syndrome: potential role of red cell transfusion. Crit Care Med 2005;33: 1191-8


6. Kiraly LN, Underwood S, Differding JA, et al.: Transfusion of aged packed red blood cells results in decreased tissue oxygenation in critically injured trauma patients. J Trauma 2009;67: 29-32


8. Toy P, Popovsky MA, Abraham E, et al.: Transfusion-related acute lung injury: definition and review Crit Care Med 2005;33: 721-6

9. Marik PE, Corwin HL: Acute lung injury following blood transfusion: expanding the defi-nition. Crit Care Med 2008;36: 3080-4

10. Gajic O, Rana R, Winters JL, et al.: Transfusion Related Acute Lung Injury in the Critically Ill: Prospective Nested Case-Control Study. Am J Respir Crit Care Med 2007;176: 886-91

11. Vlaar AP, Binnekade JM, Prins D, et al.: Risk factors and outcome of transfusion-related acute lung injury in the critically ill: A nested case-control study. Crit Care Med 2010;38: 771-8


13. Benson AB, Austin GL, Berg M, et al.: Transfusion-related acute lung injury in ICU patients admitted with gastrointestinal bleeding. Intensive Care Med 2010;36: 1710-7



16. Kelher MR, Masuno T, Moore EE, et al.: Plasma from stored packed red blood cells and MHC class I antibodies causes acute lung injury in a 2-event in vivo rat model. Blood 2009;113: 2079-87


Muller.indd 180 25-7-2014 14:40:31


181

18. Silliman CC, Bjornsen AJ, Wyman TH, et al.: Plasma and lipids from stored platelets cause acute lung injury in an animal model. Transfusion 2003;43: 633-40

19. Tung JP, Fung YL, Nataatmadja M, et al.: A novel in vivo ovine model of transfusion-related acute lung injury (TRALI). Vox Sang 2011;100: 219-30

20. Looney MR, Nguyen JX, Hu Y, et al.: Platelet depletion and aspirin treatment pro-tect mice in a two-event model of transfusion-related acute lung injury. J Clin Invest 2009;119: 3450-61

21. Kopko PM, Paglieroni TG, Popovsky MA, et al.: TRALI: correlation of antigen-antibody and monocyte activation in donor-recipient pairs. Transfusion 2003;43: 177-84

22. Toy P, Gajic O, Bacchetti P, et al.: Transfusion-related acute lung injury: incidence and risk factors. Blood 2012;119:1757-67

23. Kopko PM, Popovsky MA, MacKenzie MR, et al.: HLA class II antibodies in transfusion-related acute lung injury. Transfusion 2001;41: 1244-8

24. Curtis BR, McFarland JG: Mechanisms of transfusion-related acute lung injury (TRALI): anti-leukocyte antibodies. Crit Care Med 2006;34: S118-S123

25. Silliman CC, Dickey WO, Paterson AJ, et al.: Analysis of the priming activity of lipids gen-erated during routine storage of platelet concentrates. Transfusion 1996;36: 133-9


27. Silliman CC, Paterson AJ, Dickey WO, et al.: The association of biologically active lipids with the development of transfusion-related acute lung injury: a retrospective study. Transfusion 1997;37: 719-26


29. Mangalmurti NS, Chatterjee S, Cheng G, et al.: Advanced glycation end products on stored red blood cells increase endothelial reactive oxygen species generation through interaction with receptor for advanced glycation end products. Transfusion 2010;50: 2353-61

30. Vlaar AP, Hofstra JJ, Levi M, et al.: Supernatant of aged erythrocytes causes lung inflam-mation and coagulopathy in a “two-hit” in vivo syngeneic transfusion model. Anesthe-siology 2010;113: 92-103

31. Mangalmurti NS, Xiong Z, Hulver M, et al.: Loss of red cell chemokine scavenging pro-motes transfusion-related lung inflammation. Blood 2009;113: 1158-66

32. Zhu H, Zennadi R, Xu BX, et al.: Impaired adenosine-5’-triphosphate release from red blood cells promotes their adhesion to endothelial cells: A mechanism of hypoxemia after transfusion. Crit Care Med 2011;39: 2478-86


34. Arinsburg SA, Skerrett DL, Karp JK, et al.: Conversion to low transfusion-related acute lung injury (TRALI)-risk plasma significantly reduces TRALI. Transfusion 2012;52: 946-52

Muller.indd 181 25-7-2014 14:40:31

Chapter 10

182

35. Eder AF, Herron R, Strupp A, et al.: Transfusion-related acute lung injury surveillance (2003-2005) and the potential impact of the selective use of plasma from male donors in the American Red Cross. Transfusion 2007;47: 599-607

36. Vlaar AP, Binnekade JM, Schultz MJ, et al.: Preventing TRALI: ladies first, what follows? Crit Care Med 2008;36: 3283-4


38. Wang D, Sun J, Solomon SB, et al.: Transfusion of older stored blood and risk of death: a meta-analysis. Transfusion 2012;52: 1184-95

39. Koch CG, Li L, Sessler DI, et al.: Duration of red-cell storage and complications after cardiac surgery. N Engl J Med 2008;358: 1229-39

40. Kor DJ, Kashyap R, Weiskopf RB, et al.: Fresh red blood cell transfusion and short-term pulmonary, immunologic, and coagulation status: a randomized clinical trial. Am J Respir Crit Care Med 2012;185: 842-50


42. Vlaar AP, Hofstra JJ, Determann RM, et al.: Transfusion-related acute lung injury in car-diac surgery patients is characterized by pulmonary inflammation and coagulopathy: A prospective nested case-control study. Crit Care Med 2012;40: 2813-20

43. Silliman CC, Boshkov LK, Mehdizadehkashi Z, et al.: Transfusion-related acute lung injury: epidemiology and a prospective analysis of etiologic factors. Blood 2003;101: 454-62

44. Sakr Y, Vincent JL, Reinhart K, et al.: High tidal volume and positive fluid balance are associated with worse outcome in acute lung injury. Chest 2005;128: 3098-108

45. Clifford L, Singh A, Wilson GA, et al.: Electronic health record surveillance algorithms facilitate the detection of transfusion-related pulmonary complications. Transfusion 2013;53: 1205-16


Muller.indd 182 25-7-2014 14:40:31

Chapter 11

Contribution of damage-associated molecular patterns to

transfusion-related acute lung injury in cardiac surgery

Marcella C.A. Müller, Pieter R. Tuinman, Alexander P. Vlaar, Anita M. Tuip, Kelly Maijoor, Achmed Achouiti, Cornelis van t Veer, Margreeth B. Vroom,

Nicole P. Juffermans

Blood Transfusion 2014;2:1-8

Muller.indd 183 25-7-2014 14:40:31

Chapter 11

184

Abstract

Background: The incidence of transfusion-related acute lung injury (TRALI) in car-diac surgery patients is high and this contributes to an adverse outcome. Damage-associated molecular pattern (DAMP) molecules, HMGB1 and S100A12, are thought to mediate inflammatory changes in acute respiratory distress syndrome. We aimed to determine whether DAMP are involved in the pathogenesis of TRALI in cardiac surgery patients.Materials and methods: This was a secondary analysis of a prospective observa-tional trial in cardiac surgery patients admitted to the Intensive Care of a university hospital in the Netherlands. Fourteen TRALI cases were randomly matched with 32 transfused and non-transfused controls. Pulmonary levels of HMGB1, S100A12 and inflammatory cytokines (IL-1β, Il-6, IL-8 and tumour necrosis factor-α) were deter-mined when TRALI evolved. In addition, systemic and pulmonary levels of soluble receptor for advanced glycation end products (sRAGE) were determined.Results: HMGB1 expression and levels of sRAGE in TRALI patients did not differ from those in controls. There was a trend towards higher S100A12 levels in TRALI patients compared to the controls. Furthermore, S100A12 levels were associated with increased levels of markers of pulmonary inflammation, prolonged cardiopul-monary bypass, hypoxemia and duration of mechanical ventilation.Conclusion: No evidence was found that HMGB1 and sRAGE contribute to the devel-opment of TRALI. S100A12 is associated with duration of cardiopulmonary bypass, pulmonary inflammation, hypoxia and prolonged mechanical ventilation and may contribute to acute lung injury in cardiac surgery patients.

Muller.indd 184 25-7-2014 14:40:31

DAMPs in TRALI in cardiac surgery patients

185

Introduction

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related morbidity and mortality. Incidence of TRALI is much higher in patients undergoing cardiac surgery than in the general hospital population and contributes to an adverse outcome [1,2]. TRALI is thought to result from the interaction of acti-vated neutrophils with endothelial cells, thereby inducing endothelial damage, vas-cular leakage and pulmonary oedema. However, the precise pathways of TRALI are largely unknown.The receptor for advanced glycation end products (RAGE) is pivotal in the pro-inflammatory response in acute lung injury. RAGE is a multi-ligand pattern recog-nition receptor expressed on type I alveolar cells, endothelium and neutrophils [3], ligation with damage associated molecular pattern (DAMP) molecules, includ-ing high-mobility group box 1 (HMGB1) and S100A12 (also known as EN-RAGE) is thought to contribute to a pro-inflammatory response [4].In lung injury inflicted by mechanical ventilation or by trauma, increased levels of HMGB1 are associated with adverse outcome [5-8]. HMGB1 plays a role in neutro-phil trafficking, as demonstrated by a study in which intratracheally instilled HMGB1 resulted in neutrophil accumulation, increased cytokine release and pulmonary oedema [9]. HMGB1 may, therefore play an important role in neutrophil-mediated acute lung injury (ALI). Interestingly, HMGB1 was shown to accumulate during red blood cell storage [10].The DAMP molecule S100A12 is thought to mediate the early phase of ALI. In vitro experiments showed a direct pro-inflammatory effect of S100A12 on lung endothe-lial cells [11-13]. Soluble RAGE (sRAGE) is a marker of alveolar cell injury in ALI/acute respiratory distress syndrome (ARDS) [14] and increased levels have been demon-strated in ALI following trauma [15], injurious mechanical ventilation [16] and lung transplantation [17]. Interestingly, sRAGE levels were associated with both blood transfusion as well as with the use of cardiopulmonary bypass [17,18]. Of note, advanced glycation end products formed in red blood cells were found to ligate to endothelial-bound RAGE, resulting in endothelial damage [19]. Thereby, RAGE ligands may play a role in the neutrophil-endothelial interactions in the pathogen-esis of TRALI. We aimed to elucidate whether RAGE-activating DAMP contribute to

Muller.indd 185 25-7-2014 14:40:31

Chapter 11

186

a TRALI reaction in patients undergoing cardiac surgery. To do this, we assessed HMGB1 and S100A12 levels, as these are important RAGE activating DAMP [20] and were shown to be associated with transfusion, the use of cardiopulmonary bypass and ALI. In addition we determined sRAGE levels.

Materials and Methods

SettingThe study is a secondary analysis of a trial in cardiac surgery patients performed in an Intensive Care Unit (ICU) of a university hospital in the Netherlands [1] and was approved by the medical ethics committee of the Academic Medical Center, Amster-dam, the Netherlands (06/201#08.17.1328as). The study was carried out in accord-ance with the Declaration of Helsinki. Patients of 18 years or older were asked to give written informed consent prior to valvular and/or coronary artery surgery for participation in the study. Exclusion criteria were off-pump surgery and emergency surgery.

DesignPatients were prospectively screened for the onset of TRALI up to 30 hours after sur-gery. Using the Canadian Consensus Conference definition [21], TRALI was defined as new onset hypoxemia or deterioration demonstrated by a PaO2/FiO2 < 300, occurring within 6 hours after transfusion, with bilateral pulmonary changes, in the absence of cardiac pulmonary oedema [21-23]. Cardiogenic pulmonary oedema was identified when pulmonary arterial occlusion pressure was >18 mmHg. Chest radio-graphs taken before surgery and on arrival at the ICU were scored for the presence of new onset bilateral interstitial abnormalities by two independent physicians who were blinded to the predictor variables.Sixteen TRALI cases were identified and non-directed bronchoalveolar lavage (NBL) fluid and plasma was available from 14 patients of these for analysis. Cases were randomly matched with controls. Transfused cardiac surgery patients not develop-ing ALI and cardiac surgery patients not transfused not developing ALI served as controls.

Muller.indd 186 25-7-2014 14:40:31


187

Cardiothoracic surgery, anaesthesia procedures and Intensive Care Unit managementPatients were anesthetized with lorazepam, etomidate, sufentanil and rocuronium for induction of anaesthesia and sevoflurane plus propofol for maintenance of anaesthesia. As part of standard care, a pulmonary artery catheter was inserted for peri-operative monitoring. In all patients, cardiopulmonary bypass was performed under mild to moderate hypothermia (28°C-34°C), using a membrane oxygenator and a non-pulsatile blood flow. During the procedure, lungs were deflated. After surgery, all patients were transferred to the ICU with mechanical ventilation. The postoperative ICU protocol involved fluid infusions with normal saline and starch solutions and transfusion of leucodepleted erythrocytes to maintain haemoglobin level above 8.5 g/dL. If indicated, norepinephrine was used to maintain a mean arte-rial blood pressure of 65 mmHg and dobutamine and/or milrinone were used to achieve a cardiac index of ≥2.5 L/min/m2.

Data collectionThe pre-operative European System for Cardiac Operative Risk Evaluation (Euro-SCORE), the physical status according American Society of Anaesthesiologists (ASA score), predicted vital capacity and forced expiratory volume in 1 second (FEV1) were determined. Left ventricular function was categorized on the basis of ejection fraction (EF), into good (EF>45%), moderate (EF<45% but >30%) or poor (EF ≤30%). Data on operation time, clamp time and time on cardiopulmonary bypass, the dura-tion of mechanical ventilation, and the partial pressure of oxygen in arterial blood were all registered.

Plasma collection and analysisArterial blood samples, collected in test tubes containing EDTA, before cardiopul-monary bypass and after arrival at the ICU, were centrifuged at 1500 x g for 15 minutes at 4°C. The supernatant was stored at -80°C until sRAGE was measurement was performed.Levels of sRAGE were determined by an enzyme-linked immunosorbent assay (ELISA) developed in our laboratory [24]. In short, 96-well plates were coated over-night with mouse anti-human RAGE antibody (R&D systems, Minneapolis, Minne-

Muller.indd 187 25-7-2014 14:40:31

Chapter 11

188

sota, USA). Samples diluted as appropriate were added and incubated for 2 hours. Next, biotinylated goat anti-human RAGE antibody (R&D Systems) was added and incubated for another 2 hours. Streptavidin poly-horseradish peroxidase (HRP) was added for 30 minutes. Finally, sodium-acetate buffer (pH 5.5) containing 100 μ/mL tetramethylbenzidine and 0.003% H2O2 was added and the colour reaction was stopped by 1 N H2SO4. All measurements were made in duplicate.

Non-directed bronchoalveolar lavage techniqueAt onset of TRALI, we performed NBL, with which we have ample experience [25-28]. Controls were lavaged within 30 hours of admission to the ICU. There was no signifi-cant difference in timing of the NBL between the groups [1]. As described previously [25-28], a standard 50 cm, 14-gauge tracheal suction catheter was introduced via the endotracheal tube and advanced until significant resistance was encountered. Immediately after instillation of 20 mL of sterile 0.9% saline over 10-15 seconds, fluid was aspirated before withdrawal of the catheter. Generally, 4 to 5 mL of fluid was recovered. NBL samples were centrifuged at 1500 x g for 10 minutes at 4°C. The supernatant was collected and stored at -80°C until assays were performed. The cell pellet was re-suspended in phosphate-buffered saline and cell counts were deter-mined using a haemocytometer (Beckman Coulter, Fullerton, CA, USA) [26].

Analyses in non-directed bronchoalveolar lavage fluidThe correlation between HMGB1 concentration detected by ELISA and immunoblot is poor [29]. Since HMGB1 and HMGB2 show high homology there may be simul-taneous determination by ELISA [30]. We, therefore, choose to use western blot-ting which is more specific for HMGB1. The NBL fluid samples were mixed with 3-fold concentrated sodium dodecyl sulphate (SDS) sample buffer containing 6% β-mercaptoethanol in a 2:1 ratio and denatured for 5 minutes at 95°C. Twenty microliters of each sample was run on a 15% polyacrylamide gel and subsequently transferred to a polyvinylidene fluoride (PVDF) membrane (Pharmacia, Piscataway, NJ, USA). Cell extracts of 293T cells were included as positive control. After blocking with 5% non-fat dry milk in phosphate-buffered saline + 0.05 % Tween-20 (PBS-T), the membrane was incubated overnight at room temperature with a rabbit polyclonal antibody directed against human HMGB1 (ab18256, Abcam, Cambridge, MA, USA) in

Muller.indd 188 25-7-2014 14:40:31


189

PBS-T with 1% non-fat dry milk followed by secondary labeling with a HRP-labelled goat-anti-rabbit IgG polyclonal antibody (Bioké, Leiden, the Netherlands) in PBS-T with 1% non-fat dry milk. PVDF membranes were developed using Lumilight plus ECL substrate (Roche, Darmstadt, Germany) and a chemoluminescence detector with a cooled CCD camera (Syngene, Cambridge, UK). The density of HMGB1 bands at 35 kD were measured using AIDA image analysis software (Raytest, Straubenhardt, Ger-many). HMGB1 levels were compared to a standard curve prepared by serial dilution of 293T cell lysates which were run on the same gel with a lower detection limit of 500 cells. NBL fluid HMGB1 was then expressed in cell units (band densitometry cor-related to the number of lysed cells).Levels of sRAGE were determined by an ELISA, as described above [24]. S100A12 was detected by an ELISA (CircuLexTM, Nagano, Japan). Diluted samples were added to a microplate pre-coated with a monoclonal antibody specific for S100A12/EN-RAGE and incubated for 1 hour. After washing, HRP-conjugated anti-S100A12/EN-RAGE polyclonal antibody was added and incubated for 1 hour. Next, the substrate rea-gent tetra-methylbenzidine was added for 20 minutes. The reaction was stopped with 1 N H2SO4. All measurements were made in duplicate. Lower detection limit was 61 pg/mL.As described previously, interleukin (IL)-1β, IL-6, IL-8 and tumour necrosis factor α (TNFα) were measured using specific commercially available ELISA (PeliKine-com-pact™ kit), according to the instructions of the manufacturer (Sanquin, Amsterdam, the Netherlands) [26,31].

Statistical analysisContinuous data are expressed as mean and standard deviation (SD) or as medi-ans and interquartile ranges (IQR) according to their distribution. Categorical vari-ables are expressed as numbers (%). TRALI patients were compared with the control groups using one-way ANOVA and Dunnett’s post-test or the Kruskall-Wallis test and Mann-Whitney U test depending on data distribution. The chi-square test was used to compare categorical variables. Correlations were determined by Spearman’s Rho. Because of the small size of the individual groups (TRALI and controls), associations were determined for the total group of patients.

Muller.indd 189 25-7-2014 14:40:31

Chapter 11

190

Statistical significance was defined as p<0.05. Statistical analysis was performed with SPSS 18.0 (SPSS, Inc, Chicago, IL, USA).

Results

The patients’ characteristics are shown in table 1. Patients in the TRALI group were older, had higher ASA scores and lower FEV1 values prior to surgery when compared to controls. There were no differences in cardiac function, nor in other risk factors for ALI (data not shown). TRALI patients more often underwent combined bypass and valve replacement surgery and operation time and duration of cardiopulmo-nary bypass were significantly longer than those in controls. Patients who devel-oped TRALI received more red blood cells and fresh-frozen plasma compared to transfused controls. Cases had a lower PaO2/FiO2 ratio as well as a longer duration of mechanical ventilation.

Levels of cytokines in non-directed bronchoalveolar lavage fluidPulmonary levels of the pro-inflammatory cytokines IL-8 and IL-6 in NBL fluid were higher in TRALI patients than in controls [26]. Pulmonary levels (median (IQR)) of IL-1β did not differ among TRALI patients (29.5 (4.1-204.2) pg/mL) and either trans-fused (14.7 (10.2-44.5) pg/mL) or non-transfused controls (9.7 (1.4-14.7) pg/mL), (p=0.11). Likewise the levels of TNF-α did not differ among the groups, being 17 (4-121) pg/mL in TRALI patients compared to 36 (10-59) pg/mL in transfused and 34 (13-44) pg/mL in non-transfused controls (p=0.19).

Levels of sRAGE in plasma and bronchoalveolar lavage fluid of cardiac surgery patients with TRALI and controlsIn all patients, cardiac surgery resulted in a modest increase in plasma sRAGE levels compared to baseline values (figure 1). Following surgery, levels (median and IQR) of sRAGE in NBL fluid were low, being 34.1 (14.7-135.9) pg/mL in TRALI patients, 97.1 (19.3-150.4) pg/mL in transfused controls and 71.8 (27.3-200) pg/mL in non-trans-fused controls (p=0.57). Pulmonary levels of sRAGE were not elevated compared to plasma levels and none of the levels differed between TRALI patients and controls (figure 1).

Muller.indd 190 25-7-2014 14:40:31


191


Non transfused TransfusedNo ALIn=16

No ALIn=16

TRALIn=14

Patient dataAge, years* 63±13 65±10 74±8a

Sex, male (%) 13 (81) 9 (56) 12 (79)EuroSCORE* 3.8±2.8 5.4±3.2 6.4±2.9ASA score# 3 (1) 3 (0)d 3 (1)b

FEV1, % predicted* 100±17 83±17d 81±19a

Left ventricular function Poor (%) 0 (0) 3 (19) 1 (7)Moderate (%) 3 (21) 1 (6) 6 (43)Good (%) 11 (79) 12 (75) 7 (50)Surgery data CABG (%) 11 (69) 9 (56) 6 (43)Valve replacement (%) 3 (19) 3 (19) 0CABG + valve replacement (%) 0 3 (19) 8 (57)a

Other (%) 2 (12) 1 (6) 0Peri-operative dataRed blood cells, units# NA 2.0 (1) 2.5 (2)a

Fresh frozen plasma, units# NA 0.0 (2) 2.0 (2.8)a

Platelets, units# NA 0.0 (0) 0.5 (1)Clamp time, min* 74±38 92±37 114±45a

Pump time, min* 104±43 137±51 165±58b

Operation time, min* 313±66 381±104d 414±81c

MV, hours* 13.4±4.6 16.6±8,4 127.4±211a

PaO2/FiO2 ratio* 223±108 170±78 118±44b

ASA: American Society of Anesthesiologists; FEV1: forced expiratory volume in 1 second; CABG: coronary artery bypass grafting; MV: mechanical ventilation*normally distributed data, expressed as means ± standard deviation (SD)#not normally distributed data, expressed as median and interquartile range (IQR)ap<0.05, bp<0.01,cp≤0.001 TRALI compared to transfused or non-transfused controlsdp<0.05 transfused controls compared to non-transfused controlsNA: not applicable

Muller.indd 191 25-7-2014 14:40:31

Chapter 11

192

Figure 1: sRAGE levels in plasma before and after cardiac surgery and sRAGE levels in non-directed bronchoalveolar lavage (NBL) fluid after surgery.TRALI: TRALI cases.

Transfusion: transfused patients who did not develop lung injury.No transfusion: patients not transfused and without lung injury.Data are expressed as median and interquartile range.

Levels of HMGB1 in bronchoalveolar lavage fluid of cardiac surgery patients with TRALI and controlsNo differences were found in pulmonary HMGB1 expression (mean ± SD) between patients with TRALI (712 ± 379 [cell units]) and either transfused (605 ± 205 [cell units]) or non-transfused controls (556 ± 123 [cell units]) (p=0.86, figure 2). There was a wide variation in expression of HMGB1. If dichotomized (detected versus not detected), no differences were found among the three groups (data not shown).

Levels of S100A12 in bronchoalveolar lavage fluid of cardiac surgery patients with TRALI and controlsS100A12 could be detected at substantial levels in NBL fluid of all patients, with the highest median levels measured in TRALI patients (702 (77-1661) ng/mL) compared to 388 (217-1199) ng/mL in transfused and 119 (71-349) ng/mL in non-transfused controls (figure 3); this trend was, however, not statistically significant (p=0.15).There was a negative correlation between S100A12 levels and pO2 6 hours post-operatively (r=-0.584, p=0.014) and a positive correlation between levels of S100A12 in NBL fluid and duration of mechanical ventilation (r=0.445, p=0.005).

Muller.indd 192 25-7-2014 14:40:32


193

Furthermore, duration of ventilation was positively associated with the total num-ber of units transfused (r=0.537, p=0.002). We then determined whether there was a correlation between the degree of pulmonary inflammation and pulmonary levels of S100A12. Patients with higher NBL levels of S100A12, had higher NBL cell counts and higher NBL levels of TNF-α, IL-1B, IL-6 and IL-8 (figure 4). Levels of S100A12 were not significantly associated with EUROscore, ASA score and amount of transfusion. Levels of S100A12 did show a positive correlation with time on cardiopulmonary bypass (r=0.337, p<0.05).

Figure 2: Levels of HMGB1 in non-directed bronchoalveolar lavage (NBL) fluid after cardiac surgery.

TRALI: TRALI casesTransfusion: transfused patients who did not develop lung injury.No transfusion: patients not transfused and without lung injury.Data expressed as mean and standard deviation.The western blot is representative for HMGB1 from a patient with TRALI (A) and two control patients (B and C).

Muller.indd 193 25-7-2014 14:40:32

Chapter 11

194

Figure 3: Levels of S100A12 in non-directed bronchoalveolar lavage (NBL) fluid after cardiac surgery.

TRALI: TRALI casesTransfusion: transfused patients who did not develop lung injury.No transfusion: patients not transfused and without lung injury.Data expressed as median and interquartile range.

Discussion

In the present study, cardiac surgery patients had strongly elevated pulmonary lev-els of S100A12, which was associated with longer time on cardiopulmonary bypass, hypoxemia, prolonged mechanical ventilation and pulmonary inflammation. The levels of S100A12 were highest in those patients developing TRALI and this molecule may, therefore, play a role in pathogenesis of TRALI.Levels of S100A12 following cardiac surgery were very high compared to those in previous studies in other postoperative patients [12]. Furthermore, the level of pul-monary S100A12 was about 10-fold higher than that reported in ARDS [11]. Increased S100A12 correlated with decreased oxygenation, prolonged time on mechanical ventilation and elevated pulmonary levels of pro-inflammatory cytokines. This is in line with the recent finding that S100A12 actively contributes to the early inflamma-tory response by the amplification of inflammatory processes [32].As time on cardiopulmonary bypass was significantly associated with elevated pulmonary levels of S100A12 in this study, use of the bypass machine may have

Muller.indd 194 25-7-2014 14:40:32


195

Figure 4: Pulmonary levels of inflammatory markers grouped by S100A12 in all patients. Cell count in non-directed bronchoalveolar lavage (NBL) fluid was significantly higher in patients with higher S100A12 levels

(a). Percentage of neutrophils in NBL fluid (b). Levels of IL-1β (c), IL-6 (d), IL-8 (e) and TNFα (f) were significantly higher in patients with high levels of S100A12.Data presented in quartiles.

Muller.indd 195 25-7-2014 14:40:33

Chapter 11

196

contributed to this increase. Increased levels of S100A12 after cardiopulmonary bypass in children were found to correlate with severity of ALI [33]. Of note, levels in TRALI patients were clearly higher compared to patients with ALI after cardiopulmo-nary bypass [33]. Given that S100A12 is a pro-inflammatory mediator whose levels have been shown to rise early in the course of lung injury [11,12] and the TRALI syndrome is classically a rapid pulmonary reaction to a blood transfusion, S100A12 might contribute to pulmonary inflammation in cardiac surgery patients with a TRALI reaction. TRALI is thought to be a ‘two hit syndrome’, in which an inflamma-tory condition primes lung neutrophils, rendering them susceptible to lung injury inflicted by mediators in the blood product [34]. As S100A12 mediates inflammation via activated granulocytes [32], we speculate that S100A12 may amplify the TRALI reaction or may contribute to the “first hit” in TRALI, thereby lowering the threshold for developing TRALI [35]. This may explain the high incidence of TRALI found previ-ously [1,36].As shown before, plasma levels of sRAGE increased modestly in all patients follow-ing cardiac surgery. However, levels were low compared to those reported previ-ously [18, 33]. In addition, sRAGE in NBL fluid after surgery was not elevated and did not differ among TRALI patients and their controls. In line with this finding, sRAGE levels in NBL were not elevated in an experimental model of TRALI [37]. RAGE is expressed abundantly in pulmonary tissue and it has been proposed that S100A12 exerts its effects via RAGE. Interestingly, it was recently demonstrated that Toll-like receptor 4 (TLR4), not RAGE, is pivotal in S100A12 mediated inflammatory response [32]. In the lung, TLR4 expression has been demonstrated on vascular endothelial cells and activation has shown to contribute to pulmonary neutrophil sequestration [38]. Therefore, the relatively low levels of sRAGE in our patients do not preclude S100A12-mediated pulmonary inflammation.In this study, expression of HMGB1 was not increased in TRALI patients compared to that in controls. HMGB1 has been demonstrated to be an important mediator of inflammation in ARDS due to various causes [5,6,9]. However, its relation to out-come has not unequivocally been demonstrated. In trauma-induced ARDS, HMGB1 was associated with a poor outcome [8]. In another group of ARDS patients, HMGB1 levels did not differ between survivors and non-survivors, nor did they correlate with lung injury score and hypoxemia [6]. These conflicting results may be related

Muller.indd 196 25-7-2014 14:40:33


197

to the time course of lung injury. A late increase in HMGB1 was seen in patients with ARDS, with peak levels occurring between day 1 and 7, but not at onset of dis-ease [6]. In addition, in mechanically ventilated patients, pulmonary HMGB1 release could only be demonstrated after days and not hours of ventilation [7]. Altogether, HMGB1 has been suggested to be a late mediator of inflammation during ALI and it does not seem to play a role in TRALI.Taken together, no clear association between DAMP molecules and TRALI was found. However, S100A12 was associated with pulmonary inflammation, suggesting that this early DAMP molecule may be a strong driver of the inflammatory response in ALI following cardiac surgery. Although the increase in pulmonary S100A12 levels in TRALI patients was not statistically different from controls, the observed trend suggests that the study was underpowered to establish a clear association. Whether S100A12 is a mediator in TRALI does, therefore, remain to be established. Of note, other DAMP molecules have been shown to accumulate during the storage of blood products. These include haem and lipid peroxidation products, both ligands for TLR4, and extracellular ATP and microparticles, which are capable of activating the NLRP3 inflammasome [35]. The latter has been shown to contribute to pulmonary inflammation [39]. Whether these DAMP molecules contribute to the inflammatory changes in TRALI should be addressed in future research.There are limitations to our study. First, the number of patients is small, which may have underpowered the study. Second, NBL was collected only once, precluding conclusion about the time course of the role of DAMP molecules in the maintenance and amplification of the inflammatory response in TRALI. However, this cohort of TRALI patients in whom lavage samples were obtained is the largest to date.In conclusion, no evidence was found that HMGB1 and sRAGE contribute to the development of TRALI. The early DAMP molecule S100A12 is associated with pro-longed cardiopulmonary bypass, pulmonary inflammation, ventilation duration and hypoxemia after cardiac surgery and may mediate the priming phase of ALI in car-diac surgery patients who develop this condition. Further research is warranted to establish the role of DAMP molecules in the inflammatory response in TRALI.

Muller.indd 197 25-7-2014 14:40:33

Chapter 11

198

References



3. Morbini P, Villa C, Campo I, et al.: The receptor for advanced glycation end products and its ligands: a new inflammatory pathway in lung disease? Mod Pathol 2006;19: 1437-1445

4. van Zoelen MA, Achouiti A, van der Poll T: The role of receptor for advanced glycation endproducts (RAGE) in infection. Crit Care 2011;15: 208

5. Ogawa EN, Ishizaka A, Tasaka S, et al.: Contribution of high-mobility group box-1 to the development of ventilator-induced lung injury. Am J Respir Crit Care Med 2006;174: 400-407

6. Ueno H, Matsuda T, Hashimoto S, et al.: Contributions of high mobility group box pro-tein in experimental and clinical acute lung injury. Am J Respir Crit Care Med 2004;170: 1310-1316

7. van Zoelen MA, Ishizaka A, Wolthuls EK, et al.: Pulmonary levels of high-mobility group box 1 during mechanical ventilation and ventilator-associated pneumonia. Shock 2008;29: 441-445

8. Bitto A, Barone M, David A, et al.: High mobility group box-1 expression correlates with poor outcome in lung injury patients. Pharmacol Res 2010;61: 116-120

9. Abraham E, Arcaroli J, Carmody A, et al.: HMG-1 as a mediator of acute lung inflamma-tion. J Immunol 2000;165: 2950-2954

10. Peltz ED, Moore EE, Eckels PC, et al.: HMGB1 is markedly elevated within 6 hours of mechanical trauma in humans. Shock 2009;32: 17-22

11. Wittkowski H, Sturrock A, van Zoelen MA, et al.: Neutrophil-derived S100A12 in acute lung injury and respiratory distress syndrome. Crit Care Med 2007;35: 1369-1375

12. Kikkawa T, Sato N, Kojika M, et al.: Significance of measuring S100A12 and sRAGE in the serum of sepsis patients with postoperative acute lung injury. Dig Surg 2010;27: 307-312

13. Lorenz E, Muhlebach MS, Tessier PA, et al.: Different expression ratio of S100A8/A9 and S100A12 in acute and chronic lung diseases. Respir Med 2008;102: 567-573

14. Uchida T, Shirasawa M, Ware LB, et al.: Receptor for advanced glycation end-products is a marker of type I cell injury in acute lung injury. Am J Respir Crit Care Med 2006;173: 1008-1015

15. Cohen MJ, Brohi K, Calfee CS, et al.: Early release of high mobility group box nuclear protein 1 after severe trauma in humans: role of injury severity and tissue hypoperfu-sion. Crit Care 2009;13: R174

Muller.indd 198 25-7-2014 14:40:33


199

16. Calfee CS, Ware LB, Eisner MD, et al.: Plasma receptor for advanced glycation end prod-ucts and clinical outcomes in acute lung injury. Thorax 2008;63: 1083-1089

17. Christie JD, Shah CV, Kawut SM, et al.: Plasma levels of receptor for advanced glycation end products, blood transfusion, and risk of primary graft dysfunction. Am J Respir Crit Care Med 2009;180: 1010-1015

18. Agostoni P, Banfi C, Brioschi M, et al.: Surfactant protein B and RAGE increases in the plasma during cardiopulmonary bypass: a pilot study. Eur Respir J 2011;37: 841-847

19. Mangalmurti NS, Chatterjee S, Cheng G, et al.: Advanced glycation end products on stored red blood cells increase endothelial reactive oxygen species generation through interaction with receptor for advanced glycation end products. Transfusion 2010;50: 2353-2361

20. Kuipers MT, van der Poll T, Schultz MJ, et al.: Bench-to-bedside review: Damage-associ-ated molecular patterns in the onset of ventilator-induced lung injury. Crit Care 2011;15: 235

21. Kleinman S, Caulfield T, Chan P, et al.: Toward an understanding of transfusion-related acute lung injury: statement of a consensus panel. Transfusion 2004;44: 1774-1789

22. Toy P, Popovsky MA, Abraham E, et al.: Transfusion-related acute lung injury: definition and review. Crit Care Med 2005;33: 721-726


24. Determann RM, Wolthuis EK, Choi G, et al.: Lung epithelial injury markers are not influ-enced by use of lower tidal volumes during elective surgery in patients without preex-isting lung injury. Am J Physiol Lung Cell Mol Physiol 2008;294: L344-L350

25. Cornet AD, Hofstra JJ, Vlaar AP, et al.: Activated protein C attenuates pulmonary coagu-lopathy in patients with acute respiratory distress syndrome. J Thromb Haemost 2013;11: 894-901

26. Vlaar AP, Hofstra JJ, Determann RM, et al.: Transfusion-related acute lung injury in car-diac surgery patients is characterized by pulmonary inflammation and coagulopathy: a prospective nested case-control study. Crit Care Med 2012;40: 2813-2820

27. Schultz MJ, Determann RM, Royakkers AA, et al.: Bronchoalveolar Activation of Coagu-lation and Inhibition of Fibrinolysis during Ventilator-Associated Lung Injury. Crit Care Res Pract 2012;2012: 961784

28. Hofstra JJ, Vlaar AP, Knape P, et al.: Pulmonary activation of coagulation and inhibition of fibrinolysis after burn injuries and inhalation trauma. J Trauma 2011;70: 1389-1397

29. Urbonaviciute V, Furnrohr BG, Weber C, et al.: Factors masking HMGB1 in human serum and plasma. J Leukoc Biol 2007;81: 67-74

30. Yamada S, Inoue K, Yakabe K, et al.: High mobility group protein 1 (HMGB1) quantified by ELISA with a monoclonal antibody that does not cross-react with HMGB2. Clin Chem 2003;49: 1535-1537

31. Tuinman PR, Vlaar AP, Cornet AD, et al.: Blood transfusion during cardiac surgery is associated with inflammation and coagulation in the lung: a case control study. Crit Care 2011;15: R59

Muller.indd 199 25-7-2014 14:40:33

Chapter 11

200

32. Foell DMD, Wittkowski H, Kessel C, et al.: Pro-inflammatory S100A12 can Activate Human Monocytes via Toll-like Receptor 4. Am J Respir Crit Care Med 2013;187: 1324-34

33. Liu X, Chen Q, Shi S, et al.: Plasma sRAGE enables prediction of acute lung injury after cardiac surgery in children. Crit Care 2012;16: R91


35. Land WG: Transfusion-Related Acute Lung Injury: The Work of DAMPs. Transfus Med Hemother 2013;40: 3-13


37. Su X, Looney MR, Gupta N, et al.: Receptor for advanced glycation end-products (RAGE) is an indicator of direct lung injury in models of experimental lung injury. Am J Physiol Lung Cell Mol Physiol 2009;297: L1-L5

38. Andonegui G, Bonder CS, Green F, et al.: Endothelium-derived Toll-like receptor-4 is the key molecule in LPS-induced neutrophil sequestration into lungs. J Clin Invest 2003;111: 1011-1020

39. Kuipers MT, Aslami H, Janczy JR, et al.: Ventilator-induced lung injury is mediated by the NLRP3 inflammasome. Anesthesiology 2012;116: 1104-1115

Muller.indd 200 25-7-2014 14:40:33

Chapter 12

Transfusion-related acute lung injury: a preventable syndrome?

Marcella C.A. Müller, Nicole P. Juffermans

Expert Review of Hematology 2012;5(1):97-106

Muller.indd 201 25-7-2014 14:40:33

Chapter 12

202

Abstract

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related morbidity and mortality. Recent insights into the pathophysiology of TRALI have led to various preventive strategies. Strategies in donor management range from antibody testing of sensitized donors to the deferral of female plasma donors altogether. However, knowledge on the efficacy of measures to reduce TRALI is limited. In addition, the various measures may lead to a substantial loss of donors, hampering steady blood supply. Thereby, consensus among countries and blood collecting facilities regarding the optimal strategy to prevent TRALI is lacking. In this review, the advantages and disadvantages of various preventive measures to prevent TRALI are discussed, related to both patient factors as well as blood com-ponent-processing strategies, including transfusion policy, donor management and practices of preparation and storage conditions of blood components.

Muller.indd 202 25-7-2014 14:40:33


203

Introduction

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related morbidity and mortality. Any plasma-containing blood product, non-cellu-lar and cellular, can elicit a TRALI reaction and even small volumes can trigger the reaction. TRALI is defined as the onset of Acute Lung Injury (ALI) [1] in temporal relationship to a blood transfusion, characterized by pulmonary edema resulting in decreased oxygenation and bilateral infiltrates on a chest radiograph [2,3].TRALI is generally considered to be under-reported, underlined by the finding that the incidence in TRALI cases reported to the blood bank is increasing [4,5]. Given the persistent finding of an association between blood transfusion and the occur-rence of ALI in observational studies [6], it is probable that TRALI cases reported to the blood bank represent the “tip of the iceberg”. Although it is not known whether adverse outcome in patients that develop TRALI is causal or merely reflective of underlying disease severity, the clear association between transfusion and adverse outcome [6-8] highlights the need for preventive measures.Although the exact pathophysiology of TRALI is not known, there is near-universal agreement that anti-leukocyte antibodies play a role. Human Leukocyte Antibodies (HLA) and Human Neutrophil Antibodies (HNA) have been found in up to 85% of sera of donors implicated in a TRALI reaction. HLA and HNA are thought to activate pul-monary neutrophils in the recipient, resulting in endothelial damage and pulmonary edema. However, antibodies are not detected in all TRALI cases [9]. Also, many anti-body-containing blood products fail to produce TRALI [10]. These supposed differ-ences in susceptibility between individuals have led to the alternative hypothesis of a ‘two-event’ model, in which the first hit is determined by recipient factors at the moment of transfusion (e.g., mechanical ventilation or infection) causing priming of the neutrophils. The second hit consists of the transfusion of antibodies present in the blood product. Alternatively, biologically active substances which accumulate during storage of blood products may provide additional signals in the activation of neutrophils [11].In recent years, there has been evolving interest in the prevention of TRALI. One of the most important preventive measures has been the exclusion of female plasma donors, which has led to a relevant reduction of the incidence of (reported)

Muller.indd 203 25-7-2014 14:40:33

Chapter 12

204

TRALI [4,12], as well as a reduction of respiratory complications following multiple blood transfusions [13,14]. In addition to the antibody hypothesis, the ‘two-hit’ concept may have several implications for efforts to prevent TRALI. Recipient fac-tors can be modified aiming to reduce susceptibility for TRALI as much as possible. Furthermore, optimization of preparation and storage conditions of blood products may attenuate the second hit.This review will focus on strategies to prevent TRALI. Advantages and disadvantages of various preventive measures will be discussed, related to both patient factors and blood component-processing strategies, including donor management and practices of preparation and storage conditions of blood components and transfusion policy.

Incidence of TRALI

Hemovigilance programs report variable TRALI incidences, ranging from 1:29.000 to 1:260.000 for all transfused blood products [5,15-17], suggesting that TRALI is a rare complication of transfusion. Other data suggest much higher incidences with TRALI occurring in 1 per 1120-5000 transfused products [18,19].For several reasons, records from hemovigilance programs may underestimate TRALI incidence. Hemovigilance programs rely on spontaneous reporting. Active follow-up of transfused patients with computer-assisted screening showed a TRALI incidence of 0.85%, in contrast to the spontaneously reported incidence that was only 0.24% [20]. Look-back investigations and retrospective studies confirm under-reporting of TRALI [10,21]. A lack of knowledge among clinicians further contributes to under-recognition [22]. Also, despite the consensus definition, imputability criteria exist in various countries. These differences in local diagnostic criteria (i.e., France and UK take the presence of antibodies into account) contribute to variance in reported incidences and probably to underdiagnosing, in particular of nonimmune TRALI. Of note, some of the criteria of the consensus definition are subjective, leading to inter-observer variability in classifying cases of respiratory distress after transfusion.TRALI incidence may be higher in critically ill patients and cardiac surgery patients. Not only are these patients frequently exposed to allogenic blood products, they often have an underlying condition (e.g., sepsis or mechanical ventilation) that pro-motes sequestration and priming of neutrophils in pulmonary capillaries. In line with

Muller.indd 204 25-7-2014 14:40:34


205

the ‘two-event’ hypothesis, primed neutrophils may be prone to signals in the blood product that mediate activation and additional lung injury after transfusion. Pro-spective cohort studies showed an incidence of 5-8% in transfused intensive care unit (ICU) patients, with an incidence of TRALI per transfused unit of approximately 1% [7,8].

Outcome of TRALI

Most patients who develop TRALI need some kind of respiratory support; indeed up to 70% require invasive mechanical ventilation [18] and subsequent ICU admis-sion. When TRALI develops in critically ill patients, duration of mechanical ventila-tion is prolonged and ICU and hospital length of stay increased compared to controls [8,23]. The mortality of TRALI is considerable; up to 21% of TRALI patients die [4]. In the critically ill, TRALI is an important contributor to mortality, with 42-47% mor-tality in TRALI patients compared to 23-25% in transfused controls who did not develop ALI [7,8]. In addition to increased hospital mortality, long-term survival is also significantly decreased for critically ill medical patients with TRALI compared to critically ill transfused controls without TRALI [23]. Taken together, TRALI may not be a rare disorder and the syndrome contributes to adverse outcomes. Therefore, efforts to decrease TRALI are mandatory.

Pathogenesis of TRALI

Immune-mediated TRALIIn 65-85 % of TRALI cases, antibodies to HLA or granulocytes can be demonstrated in one of the administered blood products [18,24,25]. Involved antibodies are mainly directed against HLA class I, HLA class II or HNA and are thought to react with the cognate antigen on the recipients’ neutrophils [26,27]. This interaction triggers a cascade of inflammatory responses, culminating in pulmonary endothelial damage leading to an increase in pulmonary vascular permeability and subsequent leakage of a protein-rich pulmonary exudate [28]. In approximately 60% of the recipients, presence of the cognate antigen can be demonstrated [18]. Most involved donors in immune-mediated TRALI are multiparous women, since pregnancy is strongly

Muller.indd 205 25-7-2014 14:40:34

Chapter 12

206

associated with allo-immunization [29]. Both HLA class I and HLA class II antibodies are implicated in TRALI. Antibodies against HNA-3a (5b) are less frequently involved but associated with more severe cases of TRALI [30-32].

Non-immune-mediated TRALIAs antibodies do not seem to play a role in a substantial number of TRALI cases [8], an alternative pathophysiological explanation implicates a ‘two event’ model [11], in which the underlying condition of the patient predisposes to a TRALI reaction [33]. In primed recipients, biological response modifiers that accumulate during storage of cellular blood components have been implicated, including lysophosphatidylcho-lines (LysoPCs), nonpolar lipids (arachidonic acid and 5-, 12-, and 15-hydroxyeicso-tetranoic acid) [34-36], inflammatory cytokines and soluble CD40 ligand (sCD40L) [37]. These substances induce activation of primed neutrophils in the recipient, resulting in release of inflammatory cytokines and endothelial damage [11,35,38].Results from experimental studies show that transfused antibodies also function as the second hit [27], implicating that the two hit concept and the immune-mediated concept are not mutually exclusive. A threshold model has been suggested [39], in which a threshold must be overcome to induce a TRALI reaction. Whether TRALI evolves, depends both on the presence of primed neutrophils in the recipients as well as the potential of the transfused product to activate these neutrophils.

Prevention of TRALI using donor-management strategies

Ideally, donor-management strategies should be consistent within individual organi-zations and across transfusion medicine [40]. However, there is a lack of consensus among institutions and different countries regarding the management of donors implicated in TRALI [41,42]. In 2006 and 2007, the American Association of Blood Banks (AABB) recommended US blood-collecting facilities to take actions to miti-gate the risk of TRALI [43,44]. By 2009, 90% of blood centers instituted a platelet and/or plasma risk-reduction policy, although a wide variation regarding the type of implemented measures exists [42].

Muller.indd 206 25-7-2014 14:40:34


207

Screening all donors for HLA and/or HNA antibodiesVarious cohort studies have shown that the prevalence of HLA antibodies ranges from 1 to 7% in individuals who have never been alloexposed [45-47]. An impor-tant limitation of antibody screening is the lack of a gold standard and uniformity in antibody testing. To detect HLA antibodies, many different tests are available, with various cutoff values. Currently there are no published data to assess the sensitivity of a particular assay cutoff value for preventing TRALI [42,46]. Thereby, consensus regarding deferral of donors who test positive for antibodies is lacking [41,48]. Test-ing for granulocyte antibodies is time consuming and requires large amounts of test cells. Novel, more efficient assays for HLA and HNA antibody detection are under-way [49,50], but not yet widely available.In conclusion, testing of all donors for antibodies is a costly and time consuming strategy to prevent TRALI, due to both the low prevalence of antibodies in unex-posed donors as well as the limitations of the available assays [51].

Screening previously alloexposed donors for HLA and/or HNA antibodiesTransfusion can lead to sensitization of recipients, albeit with very different reported numbers, ranging from 1 to 12% [45,51]. Pregnancy is the most important reason for sensitization of the donor population [29,45,51,52]. About 10% of previously preg-nant women have HLA-antibodies [32,53] and this number increases to 26-39% in women who have had three or more pregnancies [29,45,51,52].In 2009, a US survey on TRALI risk-reduction strategies showed that some, but not all centers, tested for leukocyte allo-antibodies in alloexposed donors [42]. Although indications for HLA testing varied widely (e.g., number of pregnancies, history of transfusion or transplantation), the most common scenario was to screen women with one or more pregnancies [42]. None of the centers screened for HNA antibod-ies. Powers et al. demonstrated that donor history indeed is a reliable predictor of HLA alloimmunization [51]. In addition, the predictive value of a positive HLA anti-body screen increases when the test is restricted to individuals with a higher pre-test probability of HLA sensitization [47,51]. However, it is still unclear if alloexposed donors who harbor antileukocyte antibodies should be deferred.

Muller.indd 207 25-7-2014 14:40:34

Chapter 12

208

Donor deferral based on antibody screeningProspective follow-up of transfusion outcomes suggests that HLA antibodies infre-quently cause TRALI [50] and deferral of all donors that test positive for HLA anti-bodies could result in an unnecessary loss of donors. In contrast to HLA antibodies, HNA-3a (5b) antibodies are associated with more severe cases of TRALI [30,32]. In 2008, the International Society of Blood Transfusion recommended to consider deferral of donors with HNA-3a antibodies from blood donation [54].To establish if a TRALI case is antibody mediated, implicated donors and recipients should be tested to detect antibody and cognate antigen interaction. However, investigations are often incomplete as recipient samples for HLA and HNA typing can be obtained only in a minority of TRALI cases [40]. Therefore, it was recommended not to test donors for leukocyte antibodies in clinically unsubstantiated cases of TRALI, or in cases in which critical information is lacking [40].In conclusion, it is obvious that donors implicated in a proven case of TRALI should be permanently deferred from further donations [10]. Donors with HLA or nonspe-cific HNA antibodies who are implicated in a case of TRALI, could be directed to donating low plasma-volume products if causation is excluded by cross-match or recipient typing studies [40]. Finally, it seems reasonable to defer all donors who test positive for HNA-3a antibodies.

Donor deferral based on alloexposureIn Europe, most previously transfused donors are already deferred, in order to decrease the risk of prion transmission [47,55]. In the USA, donors are deferred for 12 months after receiving a transfusion. As less than 5% blood donations are made by previously transfused donors [56] and the rate of alloimmunization is low after previous transfusion [45], the deferral of transfused blood donors is not considered to be an effective TRALI mitigation strategy.As noted, alloexposure also occurs after pregnancy. Transfusion of plasma from mul-tiparous female donors led to a significant worsening in oxygenation and increased inflammatory response compared to control plasma of non-alloexposed donors [14]. Furthermore, a case referent study of 83 TRALI cases showed that the relative risk of developing TRALI was 19 (CI 95% 1.9-191) after transfusion of plasma-rich products from female donors [12]. In addition, in 71% of TRALI fatalities reported

Muller.indd 208 25-7-2014 14:40:34


209

to the American Red Cross, a female antibody-positive donor was identified [57]. Thereby, a strategy could be to defer women reported to have been pregnant.Half of blood donations in the USA are made by female donors, of which two thirds reported one or more pregnancies [45,51]. In 2006, in the USA it was recommended to exclude parous women as fresh frozen plasma (FFP) and platelet apheresis donors, as one of the several possible interventions to prevent TRALI [44]. This led to a reduction in reported TRALI cases from plasma transfusion [58]. Also, in 2007, Switzerland adopted a policy to exclude parous women and donors testing positive for HLA antibodies from the production of quarantine FFP [55].A limitation to deferring parous women only will be the failure to defer female donors with antibodies, who could not or did not accurately inform the blood bank on abortion or miscarriages. About 8% of nulliparous women have HLA antibod-ies, probably owing to missed ectopic pregnancies and premature abortions [29]. Thereby, an alternative strategy is to defer all women from donation of plasma.

Deferral of all female donorsIn 2003, the National Blood Service, which supplies 83% of blood components in the UK, was the first to introduce a predominantly male donor strategy for FFP produc-tion and for suspension of buffy coat-derived pool platelets. This resulted in a sig-nificant drop in reported TRALI cases associated with transfused FFP and platelets. Since 2005, no concordant antibody-positive cases of TRALI have been reported [4]. Also, in surgical patients receiving multiple transfusions, use of predominantly male plasma significantly reduced the onset of ALI [13] or respiratory distress [59].In 2007, The American Red Cross implemented a strategy based on preferential use of plasma collected from male donors. As a result, 95% of distributed plasma for transfusion was collected from male donors [40]. This strategy resulted in a signifi-cant reduction in TRALI associated with plasma transfusion (odds ratio: 0.21 95% CI: 0.08-0.45) [58]. Although not all blood collecting facilities in the USA implemented a gender-based donor strategy, male only was the most frequently implemented policy to reduce TRALI risk from apheresis platelets. Furthermore, 91% of the blood centers had a plasma risk-reduction policy, which in some centers included male-only plasma [42].

Muller.indd 209 25-7-2014 14:40:34

Chapter 12

210

In conclusion, exclusion of female donors prevents the majority of TRALI cases caused by plasma rich products. However, TRALI due to red blood cell transfusion and pooled platelets are not prevented.

Consequences of donor deferral for blood-banking practice

Preventing TRALI by screening or excluding (sub) groups of donors will have implica-tions for blood-banking practice. The key issue is finding a balance between safety and maintaining adequate supply of blood products.

Deferral based on antibody screeningLeukocyte antibody screening of only previously alloexposed and consequential deferral of those with HLA antibodies, will lead to a loss of up to 10% of apheresis donors and 7% of whole blood donations [32,51,56]. Deferral of all donors with any leukocyte antibody from donation of any blood product would result in a loss of 9% of donors [47]. An alternative option is exclusion based on the type of antibody, such as, deferral of those with HNA-3a antibodies and HLA antibodies that react with HLA antigens that are highly prevalent. Obviously, donors with circulating antibodies that have been implicated in TRALI in which the cross match with the recipient is con-firmed, should be deferred from further donations [40].

Deferral based on alloexposureAs most previously transfused donors are already deferred to decrease the risk of prion transmission [47,55], implementation of this strategy will have limited effect on blood-banking practice.However, deferral of all previous pregnant women from donating plasma-rich blood components however, would lead to a substantial loss of donors. As discussed pre-viously, approximately 50% of donors are female and up to 65% of them have been pregnant [29]. If all these women would be excluded, it is estimated that this would result in a loss of 30% of whole blood donations and nearly a quarter of all apheresis platelet donations [56]. In Spain, apheresis donations from females with a history of pregnancy and positive for HLA antibodies are used for plasma derivates. This led to 16% loss of transfusable apheresis plasma and 12% loss of apheresis platelets from

Muller.indd 210 25-7-2014 14:40:34


211

female donors. The latter could be compensated with platelets obtained from whole blood donors. However, the loss of transfusable apheresis plasma was temporar-ily troublesome and it was occasionally necessary to import FFP from other blood banks [60].

Deferral of all womenThis strategy carries the risk of unnecessary loss of donors and blood that would be safe for most recipients. However, in the UK, the implementation of predominantly male FFP was not accompanied by loss of any donors [4]. Also in The Netherlands, the male-only plasma measure was implemented without significant costs or seri-ous threat to the plasma supply [61]. By contrast, calculations from the American Red Cross Blood Service estimated that deferring all female donors would result in a nearly 50% reduction in the units of whole blood available for plasma manufac-turing. Deferral of female apheresis platelet donors would result in a loss of 37% of apheresis platelet donations [56]. It is expected that the loss of female-derived transfusable plasma units can be compensated, but the losses of apheresis platelet would impact clinical care and are considered unacceptable.A possible explanation for this discrepancy in loss of platelet donations is the differ-ence in how platelets are collected. In the USA the predominant method is aphere-sis, while in Europe buffycoat-derived platelet pools are used. The latter also contain some residual plasma, but can be resuspended in male plasma or additive solution (AS), which has only recently became available in the USA [4]. In general, apher-esis platelet risk-reduction policies led to an increased production of whole-blood derived platelets in 25% of the blood centers [42].In conclusion, deferral of all female plasma donations is proven to be feasible and efficient in reducing TRALI. The deferral of all female platelet apheresis donations without substantial losses is currently not feasible. Furthermore, to prevent unnec-essary loss of donors, it should be emphasized that female donors remain eligible to donate cellular components, but that their plasma would be preferentially diverted to fractionation rather than infusion [57].

Muller.indd 211 25-7-2014 14:40:34

Chapter 12

212

Prevention of TRALI by modifying product preparation and storage condition

LeukoreductionMany countries and numerous blood-collecting facilities in the USA had imple-mented universal prestorage leukoreduction by the end of the 1990s. Its imple-mentation is associated with decreased in-hospital mortality, which can possibly be attributed to a decrease in the amount of inflammatory transfusion reactions [62].According to the two-event model, TRALI can result from transfusion of bioactive lipids that accumulate during storage of cellular blood components. Prestorage leukoreduction of red blood cells reduces accumulation of cytokines [63], with a concomitant attenuation of neutrophil-priming activity compared with non-leu-koreduced blood [64]. However, accumulation of non-polar lipids is not attenuated by prestorage leukoreduction of red blood cells and these lipids have been shown to induce ALI in a two-event in vivo model of TRALI [36]. In addition, observational studies on the effect of leukoreduction on ALI and TRALI incidence have shown con-flicting results [65-69]. Of note, only a small minority (<10%) of TRALI cases result from the interaction of transfused white blood cells with antibodies in the recipient [60].Blood product leukoreduction leads to reduced allo-exposure of transfused indi-viduals, which consequently reduces the sensitization rates of these individuals [68]. However, as rates of alloimmunization are modest and transfused individuals are excluded from donation, it is doubtful whether leukoreduction attributes to a reduced incidence of TRALI.

Decreasing storage timeDuring storage, the erythrocyte undergoes numerous changes and acquires pro-inflammatory properties, collectively referred to as the “red cell storage lesion” [70]. In addition, bioactive lipids (LysoPCs) that accumulate in cellular blood com-ponents during storage are associated with increased neutrophil-priming activity of the blood product [71,72]. Indeed, the accumulated lipids have been shown to induce TRALI in a two-event model of TRALI in rats [35,38] and the occurrence of TRALI has been linked with transfusion of blood containing lipids with significant

Muller.indd 212 25-7-2014 14:40:34


213

neutrophil-priming activity [34]. Of note, association of aged red blood cells and transfusion-related morbidity and mortality has mainly been found in studies per-formed in the USA [27,35,73], but not outside the USA [8,74], which may suggest a difference in product preparation or storing processes of red blood cells. In a study performed in The Netherlands, stored red blood cells did not show lysoPC accumu-lation, nor neutrophil priming [72].Besides lipids, sCD40L has been implicated in TRALI [36-38]. Levels of sCD40L were increased in platelet products associated with TRALI compared to control plate-let products and sCD40L had neutrophil-priming capacity in vitro [37]. Of note, although observational studies report associations between prolonged storage of blood products and respiratory failure in patients after cardiac surgery and trauma [73,75], storage time could not be linked to TRALI in observational trials [7,8].

Altering storage conditionsOf interest, the accumulation of LysoPCs depends on the percentage of plasma present in the storage medium [72]. The use of ASs for the resuspension of plate-lets (platelet additive solution) or erythrocytes (SAG-M) can reduce the amount of plasma content in the final product. However, since even small amounts of plasma can cause TRALI, these substances will not totally eliminate this risk [76]. Whether the use of ASs will reduce the incidence of TRALI remains to be determined.Washing of red blood cells can reduce the concentration of HLA and HNA antibodies, which could reduce the incidence of TRALI [41]. Also, Silliman et al. demonstrated that washing of stored packed red blood cells reduces neutrophil-priming activity of the product [77]. Indeed, in a two-hit rat transfusion model, transfusion of washed red blood cells prevented onset of TRALI [78]. Whether washing can be performed without diminishing the oxygen-delivering capacity remains to be determined. At present, there are no clinical data available about washing of blood components and TRALI.In conclusion, preparation and storage conditions of blood components are involved in the pathophysiology of TRALI. However, it is not yet known whether specific measures regarding the preparation or storage of blood products influences TRALI incidence [41].

Muller.indd 213 25-7-2014 14:40:34

Chapter 12

214

Preparation of solvent detergent plasmaSolvent/detergent (S/D) plasma (Octaplas®) is prepared from pools containing plasma from 500-1600 donations, which leads to at least a 500-fold dilution of donations containing leukocyte antibodies. Indeed, no HLA Class I or II or granulocyt agglutinating antibodies could be detected in S/D plasma [79]. After introduction of S/D plasma in Norway [79], no cases of TRALI have been reported in recipients so far [17,80].A concern of the use of pooled plasma has been the occurrence of prion diseases. However, to date, this has not been reported. Another limitation to implementation of S/D plasma is financial. The replacement of all units of FFP in the UK by S/D plasma is estimated to cost £9.2 million/year to prevent 107 cases of TRALI [81]. The intro-duction of the use of male-only FFP may further negate supposed beneficial effects of S/D plasma on the incidence of TRALI.

Reducing recipient exposure to blood products

The most important measure to prevent TRALI is to enforce appropriate use of blood products. A restrictive transfusion strategy reduces the incidence of ALI in the critically ill [82] and is associated with decreased mortality [83]. Multiple-unit trans-fusions to correct for anaemia are still common practice. Also, inappropriate use of blood products (i.e., outside of guidelines) is still frequent, including the use of FFP for the reversal of warfarin and for prevention of hemorrhage in case of prolonged coagulation parameters. A measure to reduce inappropriate transfusion may be the use of computerized transfusion algorithms [84]. However, blood transfusion cannot be completely avoided. Indeed, data from combat trauma patients have led to more aggressive use of FFP and platelets for the resuscitation of severe bleeding patients [85]. Thereby, FFP and platelet use in the UK has remained virtually unchanged in the past decade [86].Of note, there is an emerging interest in the use of alternative products to correct coagulation disturbances. A recent trial showed that the use of tranexamic acid in trauma patients significantly reduced mortality [87]. Furthermore, the use of point of care tests to assess coagulation (e.g., ROTEM®) in combination with transfusion

Muller.indd 214 25-7-2014 14:40:34


215

algorithms has shown to reduce the amount of transfused blood products in trauma and cardiac surgery [88,89].

Modifying host factors to reduce susceptibility for TRALI

Patient-related risk factors for TRALI have recently been identified. These include mechanical ventilation, cardiothoracic surgery, sepsis, trauma and hematologic malignancy [7,8]. Most of these factors cannot be modified. However, measures that attenuate ALI are likely to raise the threshold for TRALI. Protective mechanical venti-lation strategies using low tidal volumes in mechanical ventilation reduce mortality with 22% in ALI/acute respiratory distress syndrome [90] and injurious mechanical ventilation with high tidal volumes is shown to aggravate TRALI in a mouse model [91]. Therefore, it seems reasonable to apply lung protective ventilation strategies in patients subjected to a blood transfusion. In addition, fluid-restrictive management has been shown to reduce ventilation days in patients with ALI [92]. It is conceivable that these results also apply to TRALI.

Expert commentary

TRALI is a relevant clinical problem with substantial morbidity and mortality in cer-tain patient populations. In the past decade, important preventive measures have been implemented. The deferral of female donors for FFP production has led to a reduction of TRALI cases. Depending on blood bank practices, such a policy is feasi-ble without inferring with adequate blood supply. Preventing TRALI due to platelet transfusion is more complicated. Deferral of all women from donating apheresis platelets does not seem feasible, since this will lead to a substantial loss of donors and, consequently, clinical supply problems. Antibody testing of platelet donors can be considered, but has considerable limitations. However, at present, antibody test-ing of alloexposed platelet donors seems a reasonable TRALI risk-reduction strat-egy. Alternative measures to prevent TRALI are the resuspension of pooled platelets in male plasma, which was relatively easily implemented in the UK and The Nether-lands, or the addition of platelet AS instead of plasma.

Muller.indd 215 25-7-2014 14:40:34

Chapter 12

216

However, the above measures will only prevent immune-mediated cases of TRALI due to FFP or platelet transfusion. Also, patients with a decreased threshold to develop TRALI (e.g., those who are critically ill) are at risk to develop TRALI, even if the transfused product only contains a small amount of plasma. Further research in optimizing preparation and storage conditions with the aim to reduce accumulation of bioactive substances is warranted.It needs to be elucidated whether it is beneficial to implement tailor-made transfu-sion therapy for specific patient populations at risk for TRALI. This could potentially include the preferential use of ‘fresh’ blood products, S/D plasma or washed blood products, platelets from male-only or antibody-negative donors and possibly even male-only red blood cell products.Finally, despite extensive evidence about the harmful effects of blood products, inappropriate use is still common. Awareness needs to be heightened among clini-cians and studies on the use of alternative products need to be encouraged.

Five-year view

Consensus needs to be reached concerning donor deferral and antibody screen-ing of donors. To improve donor screening, standardized HLA and HNA tests are needed. In addition, stronger evidence is needed regarding the effect of deferral of donors that harbor antibodies, in order to maximize TRALI prevention with the least possible loss of donors. Furthermore, to prevent unnecessary deferral of donors with leukocyte antibodies, both donor testing and crossmatching with the recipi-ent’s cells or typing of the recipient’s cells should be performed when TRALI is sus-pected. A prerequisite to prevent use of products from donors implicated in TRALI is universal use of the consensus definition for reporting of TRALI cases to the blood bank.The deferral of female plasma donors has been shown to be effective and feasible and is expected to be applied in more countries. For other plasma-rich products (e.g., platelets), deferral of all females is not feasible. Guidelines on donor screening and deferral are warranted.It remains to be elucidated whether the use of fresh blood products is beneficial compared with prolonged storage, which is associated with the accumulation of bio-

Muller.indd 216 25-7-2014 14:40:34


217

active substances. Results of the ABLE trial, which compared fresh blood with stand-ard transfusion care in critically ill patients [101], may help to answer this question.Awareness among clinicians about risk factors for TRALI needs to be heightened, especially among those who are involved in the care of high-risk patients, such as ICU physicians, anesthesiologists and hematologists.Results should be awaited of trials investigating the (preventive) use of plasma-rich products in critically ill patients with a coagulopathy [102-104]. In addition, the ongo-ing RELIEVE trial might clarify whether a restrictive transfusion policy in critically ill is safe and advantageous [105]. Hopefully, the results of these trials will further con-tribute to the appropriate use of blood products.

Muller.indd 217 25-7-2014 14:40:34

Chapter 12

218

References

1. Bernard GR, Artigas A, Brigham KL, et al.: The American-European Consensus Confer-ence on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordina-tion. Am J Respir Crit Care Med 1994;149: 818-824

2. Toy P, Popovsky MA, Abraham E, et al.: Transfusion-related acute lung injury: definition and review. Crit Care Med 2005;33: 721-726



5. Wiersum-Osselton JC, Porcelijn L, van Stein D, et al.: Transfusion-related acute lung injury (TRALI) in the Netherlands in 2002-2005. Ned Tijdschr Geneeskd 2008;152: 1784-1788




9. Kopko PM, Popovsky MA, MacKenzie MR, et al.: HLA class II antibodies in transfusion-related acute lung injury. Transfusion 2001;41: 1244-1248

10. Kopko PM, Marshall CS, MacKenzie MR, et al.: Transfusion-related acute lung injury: report of a clinical look-back investigation. JAMA 2002;287: 1968-1971


12. Middelburg RA, van Stein D, Zupanska B, et al.: Female donors and transfusion-related acute lung injury: A case-referent study from the International TRALI Unisex Research Group. Transfusion 2010;50: 2447-2454


14. Palfi M, Berg S, Ernerudh J, et al.: A randomized controlled trial oftransfusion-related acute lung injury: is plasma from multiparous blood donors dangerous? Transfusion 2001;41: 317-322

15. Ozier Y, Muller JY, Mertes PM, et al.: Transfusion-related acute lung injury: reports to the French Hemovigilance Network 2007 through 2008. Transfusion 2011;51: 2102-10

16. Keller-Stanislawski B, Reil A, Gunay S, et al.: Frequency and severity of transfusion-related acute lung injury--German haemovigilance data (2006-2007). Vox Sang 2010;98: 70-77

Muller.indd 218 25-7-2014 14:40:34


219

17. Flesland O: A comparison of complication rates based on published haemovigilance data. Intensive Care Med 2007;33: S17-S21



20. Finlay HE, Cassorla L, Feiner J, et al.: Designing and testing a computer-based screening system for transfusion-related acute lung injury. Am J Clin Pathol 2005;124: 601-609


22. Kram R, Loer SA: Transfusion-related acute lung injury: lack of recognition because of unawareness of this complication? Eur J Anaesthesiol 2005;22: 369-372


24. Win N, Massey E, Lucas G, et al.: Ninety-six suspected transfusion related acute lung injury cases: investigation findings and clinical outcome. Hematology 2007;12: 461-469

25. Wallis JP, Lubenko A, Wells AW, et al.: Single hospital experience of TRALI. Transfusion 2003;43: 1053-1059

26. Looney MR, Su X, Van Ziffle JA, et al.: Neutrophils and their Fc gamma receptors are essential in a mouse model of transfusion-related acute lung injury. J Clin Invest 2006;116: 1615-1623

27. Kelher MR, Masuno T, Moore EE, et al.: Plasma from stored packed red blood cells and MHC class I antibodies causes acute lung injury in a 2-event in vivo rat model. Blood 2009;113: 2079-2087

28. Silliman CC, McLaughlin NJ: Transfusion-related acute lung injury. Blood Rev 2006;20: 139-159

29. Densmore TL, Goodnough LT, Ali S, et al.: Prevalence of HLA sensitization in female apheresis donors. Transfusion 1999;39: 103-106

30. Davoren A, Curtis BR, Shulman IA, et al.: TRALI due to granulocyte-agglutinating human neutrophil antigen-3a (5b) alloantibodies in donor plasma: a report of 2 fatalities. Trans-fusion 2003;43: 641-645

31. Gottschall JL, Triulzi DJ, Curtis B, et al.: The frequency and specificity of human neutro-phil antigen antibodies in a blood donor population. Transfusion 2011;51: 820-827

32. Reil A, Keller-Stanislawski B, Gunay S, et al.: Specificities of leucocyte alloantibodies in transfusion-related acute lung injury and results of leucocyte antibody screening of blood donors. Vox Sang 2008;95: 313-317

33. Van Buren NL, Stroncek DF, Clay ME, et al.: Transfusion-related acute lung injury caused by an NB2 granulocyte-specific antibody in a patient with thrombotic thrombocyto-penic purpura. Transfusion 1990;30: 42-45

Muller.indd 219 25-7-2014 14:40:34

Chapter 12

220

34. Silliman CC, Paterson AJ, Dickey WO, et al.: The association of biologically active lipids with the development of transfusion-related acute lung injury: a retrospective study Transfusion 1997;37: 719-726




38. Silliman CC, Bjornsen AJ, Wyman TH, et al.: Plasma and lipids from stored platelets cause acute lung injury in an animal model. Transfusion 2003;43: 633-640


40. Eder AF, Benjamin RJ: TRALI risk reduction: donor and component management strate-gies. J Clin Apher 2009;24: 122-129

41. Wendel S, Biagini S, Trigo F, et al.: Measures to prevent TRALI. Vox Sang 2007;92: 258-277


43. Connor JD, Lipton KS: Clarifications to recommendations to reduce the risk of TRALI. Association Bulletin of the AABB 2007;#03-07

44. Strong DM, SLK: Transfusion Related Lung Injury. Association Bulletin of the AABB 2006;#06-07


46. Kakaiya RM, Triulzi DJ, Wright DJ, et al.: Prevalence of HLA antibodies in remotely trans-fused or alloexposed volunteer blood donors. Transfusion 2010;50: 1328-1334

47. Middelburg RA, Porcelijn L, Lardy N, et al.: Prevalence of leucocyte antibodies in the Dutch donor population. Vox Sang 2011;100: 327-335

48. Kopko P, Silva M, Shulman I, et al.: AABB survey of transfusion-related acute lung injury policies and practices in the United States. Transfusion 2007;47: 1679-1685

49. Nguyen XD, Dengler T, Schulz-Linkholt M, et al.: A novel tool for high-throughput screen-ing of granulocyte-specific antibodies using the automated flow cytometric granulo-cyte immunofluorescence test (Flow-GIFT). Scientif World J 2011;11: 302-309

50. Quillen K, Medrano C, Adams S, et al.: Screening plateletpheresis donors for HLA anti-bodies on two high-throughput platforms and correlation with recipient outcome. Transfusion 2011;51: 504-510

Muller.indd 220 25-7-2014 14:40:34


221

51. Powers A, Stowell CP, Dzik WH, et al.: Testing only donors with a prior history of preg-nancy or transfusion is a logical and cost-effective transfusion-related acute lung injury prevention strategy. Transfusion 2008;48: 2549-2558

52. Sachs UJ, Link E, Hofmann C, et al.: Screening of multiparous women to avoid transfu-sion-related acute lung injury: a single centre experience. Transfus Med 2008;18: 348-354

53. Maslanka K, Michur H, Zupanska B, et al.: Leucocyte antibodies in blood donors and a look back on recipients of their blood components. Vox Sang 2007;92: 247-249

54. Bierling P, Bux J, Curtis B, et al.: Recommendations of the ISBT Working Party on Granu-locyte Immunobiology for leucocyte antibody screening in the investigation and pre-vention of antibody-mediated transfusion-related acute lung injury. Vox Sang 2009;96: 266-269

55. Jutzi M, Levy G, Taleghani BM: Swiss Haemovigilance Data and Implementation of Meas-ures for the Prevention of Transfusion Associated Acute Lung Injury (TRALI). Transfus Med Hemother 2008;35: 98-101

56. Rios JA, Schlumpf KS, Kakaiya RM, et al.: Blood donations from previously transfused or pregnant donors: a multicenter study to determine the frequency of alloexposure. Transfusion 2011;51: 1197-1206


58. Eder AF, Herron RM, Strupp A, et al.: Effective reduction of transfusion-related acute lung injury risk with male-predominant plasma strategy in the American Red Cross (2006-2008). Transfusion 2010;50: 1732-1742

59. Nakazawa H, Ohnishi H, Okazaki H, et al.: Impact of fresh-frozen plasma from male-only donors versus mixed-sex donors on postoperative respiratory function in surgical patients: a prospective case-controlled study. Transfusion 2009;49: 2434-2441

60. Insunza A, Romon I, Gonzalez-Ponte ML, et al.: Implementation of a strategy to prevent TRALI in a regional blood centre. Transfus Med 2004;14: 157-164


62. Hebert PC, Fergusson D, Blajchman MA, et al.: Clinical outcomes following institution of the Canadian universal leukoreduction program for red blood cell transfusions. JAMA 2003;289: 1941-1949

63. Shanwell A, Kristiansson M, Remberger M, et al.: Generation of cytokines in red cell concentrates during storage is prevented by prestorage white cell reduction. Transfu-sion 1997;37: 678-684

64. Sparrow RL, Patton KA: Supernatant from stored red blood cell primes inflammatory cells: influence of prestorage white cell reduction. Transfusion 2004;44: 722-730

65. Watkins TR, Rubenfeld GD, Martin TR, et al.: Effects of leukoreduced blood on acute lung injury after trauma: a randomized controlled trial. Crit Care Med 2008;36: 1493-1499

Muller.indd 221 25-7-2014 14:40:34

Chapter 12

222

66. Yazer MH, Podlosky L, Clarke G, et al.: The effect of prestorage WBC reduction on the rates of febrile nonhemolytic transfusion reactions to platelet concentrates and RBC. Transfusion 2004;44: 10-15

67. Blumberg N, Heal JM, Gettings KF, et al.: An association between decreased cardio-pulmonary complications (transfusion-related acute lung injury and transfusion-asso-ciated circulatory overload) and implementation of universal leukoreduction of blood transfusions. Transfusion 2010;50: 2738-2744

68. Vamvakas EC: Meta-analysis of randomized controlled trials of the efficacy of white cell reduction in preventing HLA-alloimmunization and refractoriness to random-donor platelet transfusions. Transfus Med Rev 1998;12: 258-270

69. Plurad D, Belzberg H, Schulman I, et al.: Leukoreduction is associated with a decreased incidence of late onset acute respiratory distress syndrome after injury. Am Surg 2008;74: 117-123

70. Vlaar AP, Straat M, Juffermans NP: The relation between aged blood products and onset of transfusion-related acute lung injury. A review of pre-clinical data Clin Lab 2011;57: 267-272

71. Rael LT, Bar-Or R, Ambruso DR, et al.: The effect of storage on the accumulation of oxi-dative biomarkers in donated packed red blood cells. J Trauma 2009;66: 76-81

72. Vlaar AP, Kulik W, Nieuwland R, et al.: Accumulation of bioactive lipids during storage of blood products is not cell but plasma derived and temperature dependent. Transfusion 2011;51: 2358-66

73. Koch CG, Li L, Sessler DI, et al.: Duration of red-cell storage and complications after cardiac surgery. N Engl J Med 2008;358: 1229-1239

74. van de Watering L, Lorinser J, Versteegh M, et al.: Effects of storage time of red blood cell transfusions on the prognosis of coronary artery bypass graft patients. Transfusion 2006;46: 1712-1718

75. Weinberg JA, McGwin G, Griffin RL, et al.: Age of transfused blood: an independent predictor of mortality despite universal leukoreduction. J Trauma 2008;65: 279-282

76. Win N, Chapman CE, Bowles KM, et al.: How much residual plasma may cause TRALI? Transfus Med 2008;18: 276-280

77. Silliman CC, Moore EE, Johnson JL, et al.: Transfusion of the injured patient: proceed with caution. Shock 2004;21: 291-299

78. Vlaar AP, Hofstra JJ, Levi M, et al.: Supernatant of aged erythrocytes causes lung inflam-mation and coagulopathy in a “two-hit” in vivo syngeneic transfusion model. Anesthe-siology 2010;113: 92-103

79. Sachs UJ, Kauschat D, Bein G: White blood cell-reactive antibodies are undetectable in solvent/detergent plasma. Transfusion 2005;45: 1628-1631

80. Flesland O, Seghatchian J, Solheim BG: The Norwegian Plasma Fractionation Project--a 12 year clinical and economic success story. Transfus Apher Sci 2003;28: 93-100

81. Riedler GF, Haycox AR, Duggan AK, et al.: Cost-effectiveness of solvent/detergent-treated fresh-frozen plasma. Vox Sang 2003;85: 88-95

82. Yilmaz M, Keegan MT, Iscimen R, et al.: Toward the prevention of acute lung injury: protocol-guided limitation of large tidal volume ventilation and inappropriate transfu-sion. Crit Care Med 2007;35: 1660-1666

Muller.indd 222 25-7-2014 14:40:34


223

83. Hebert PC, Wells G, Blajchman MA, et al.: A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med 1999;340: 409-417

84. Rana R, Afessa B, Keegan MT, et al.: Evidence-based red cell transfusion in the criti-cally ill: quality improvement using computerized physician order entry. Crit Care Med 2006;34: 1892-1897

85. Beekley AC: Damage control resuscitation: a sensible approach to the exsanguinating surgical patient. Crit Care Med 2008;36: S267-S274


87. Shakur H, Roberts I, Bautista R, et al.: Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemor-rhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010;376: 23-32

88. Shore-Lesserson L, Manspeizer HE, DePerio M, et al.: Thromboelastography-guided transfusion algorithm reduces transfusions in complex cardiac surgery. Anesth Analg 1999;88: 312-319


90. The Acute Respiratory Distress Syndrome Network: Ventilation with lower tidal vol-umes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342: 1301-1308


92. Wiedemann HP, Wheeler AP, Bernard GR, et al.: Comparison of two fluid-management strategies in acute lung injury. N Engl J Med 2006;354: 2564-2575

Websites101. Age of Blood Evaluation (ABLE) trial in the resuscitation of critically ill patients: a mul-

ticentre randomised controlled superiority clinical trial (ISRCTN44878718). www.con-trolled-trials.com/ISRCTN44878718

102. Is plasma transfusion beneficial prior to low-risk procedures in hospitalized patients with blood clotting abnormalities? (NCT00953901). http://clinicaltrials.gov/ct2/show/NCT00953901

103. Transfusion of fresh frozen plasma in non-bleeding intensive care unit (ICU) patients (TOPIC) (NCT01143909). http://clinicaltrials.gov/ct2/show/NCT01143909

104. Appropriateness of frozen plasma use in Canada (NCT01187030). http://clinicaltrials.gov/ct2/show/NCT01187030

105. A feasibility randomised trial comparing restrictive and liberal blood transfusion strate-gies in patients requiring four or more days in intensive care (ICU) (ISRCTN24842715). www.controlled-trials.com/ISRCTN24842715

Muller.indd 223 25-7-2014 14:40:34

Muller.indd 224 25-7-2014 14:40:34

Chapter 13

Prevention of immune-mediated transfusion-related

acute lung injury; from bloodbank to patient

Marcella C.A. Müller*, Leendert Porcelijn*, Alexander P.J. Vlaar

*authors contributed equally

Current Pharmaceutical Design 2012;18(22):3241-8

Muller.indd 225 25-7-2014 14:40:34

Chapter 13

226

Abstract

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion related morbidity and mortality. Immune-mediated TRALI is caused by leucocyte and neutrophil antibodies in the transfused blood products that react with white blood cell antigens of the recipient, hereby inducing endothelial damage and lung injury. About two thirds of TRALI cases are thought to be immune-mediated. Both Human Leucocyte Antibodies (HLA Class I and II) and Human Neutrophil Antibodies (HNA) are involved in TRALI. Most antibodies result from allo-exposure of the blood donor, with multiparous donors having the highest incidence of antibodies. Detec-tion of anti-leucocyte and anti-neutrophil antibodies is complex and many uncer-tainties still exist regarding the interpretation of the test results.In this review we discuss the evidence and effectiveness of measurements to pre-vent immune-mediated TRALI from a bloodbank and bedside perspective. From a bloodbank perspective various preventive measures have been implicated. In some countries bloodbanks have successfully implemented donor selection strategies, ranging from testing of allo-exposed donors for leucocyte antibodies to the exclu-sion of all females from donating high plasma volume products. Another strategy involves dilution of antibodies present by pooling of plasma donations of multiple donors.From a bedside view, the most important measure to prevent TRALI is to limit patients’ exposure to allogenic bloodproducts. Furthermore recognition and aware-ness of the syndrome need to be heightened among clinicians.

Muller.indd 226 25-7-2014 14:40:34

Prevention of TRALI; from bloodbank to patient

227

Introduction

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related morbidity and mortality [1-4]. Although traditionally regarded as a rare syn-drome, recent studies show that the incidence is high in specific patient populations such as critically ill patients [5-7] and significantly contributes to adverse outcome [7,8]. A two-event hypothesis has been postulated that may explain the high inci-dence in the critically ill [9,10]. The first event is an underlying inflammatory con-dition, causing priming of the pulmonary neutrophils and the pulmonary vascular endothelium. The second event is the transfusion of a blood product containing either antibodies or factors that accumulate during storage, providing additional signals for neutrophil-mediated endothelial damage and lung injury. In general the mediators of the second hit can be divided into immune-mediated onset of TRALI (human leucocyte antibodies and human neutrophil antibodies) and non-immune-mediated (factors that accumulate during storage of cell containing blood products such as lysophospatidylcholines, sCD40L). In line with the two-event model, criti-cally ill patients are at high risk for acquiring TRALI. Incidences up to 8% of patients transfused have been reported for this patient population [5,7]. Generally stated to have a good prognosis, recent studies show that development of TRALI has a signifi-cant effect on morbidity and mortality. At this time, treatment of TRALI is directed to supportive care as no therapeutic strategy exists. From this point of view, coun-tries started preventive measurements including donor exclusion policies. In the present review the focus will be on the prevention of immune-mediated TRALI, the prevention of non-immune-mediated TRALI will be discussed elsewhere in the issue. The perspective of this review will be from a bloodbank and a bedside view.

Methods and Materials

The Medline database was used to identify medical subject’s headings (MeSH) to select search terms. In addition to MeSH terms, we also used free–text words. Search terms referred to aspects of the condition (“TRALI”, “prevention”) as well as related topics (“human neutrophil antibodies”, “human leucocyte antibodies”, “fresh frozen plasma” and “platelet transfusion”). Relevance of each paper was assessed using the

Muller.indd 227 25-7-2014 14:40:34

Chapter 13

228

on–line abstracts. In addition, the reference lists of retrieved papers were screened for potentially important papers.

Background

In 1983 Popovsky et al. published the first landmark report on the association between the presence of leucocyte antibodies in the donor serum and onset of acute lung injury (ALI) in the recipient of the transfusion [11]. They described 5 cases of transfusion-related ALI in which in all cases leuco-agglutinating and lymphocyto-toxic antibodies were found in plasma of the transfused blood products. In 3 cases, the antibodies corresponded to the HLA antigens of the recipient. It was also recog-nized that multiparous blood donors whose plasma contained these antibodies rep-resented a potential transfusion hazard. A following report of the same group first identified TRALI as a distinct clinical entity, they reported granulocyte-reactive and lymphocytotoxic antibodies in the sera of 89% and 72% of the blood donors impli-cated in 36 cases of TRALI, respectively. Subsequently, many other authors have reported on the association between the presence of antibodies against human leu-cocyte antigens (HLA) or human neutrohpil antigens (HNA) in donor blood and the onset of TRALI in the recipient [2,12,13]. Although elucidation of the pathogenesis of TRALI is not complete, the role of transfused blood donor HLA and HNA antibodies is widely accepted.

Leucocyte antibodies involved in TRALI

Antibodies in transfused blood products against HNA, HLA-class I and HLA-class II can cause TRALI in the recipient. The antibodies involved are mostly IgG alloanti-bodies produced after pregnancy, transfusion or transplantation, but can also be present in a small percentage of apparently non-immunized donors [14-16]. The number of TRALI cases due to antibodies in the recipient binding to the cognate antigens on the white cells in the blood products is reduced after the introduction of leuco-depletion of blood products [17].Up to now, eight HNA are known [18,19]. Furthermore, a small percentage of Caucasians is deficient for FcγRIIIb which can lead to immunization and FcγRIIIb alloantibody production (tables 1 and 2) [20]. Of these antibodies, HNA-1a, 1b, 2

Muller.indd 228 25-7-2014 14:40:34


229

Table 1: Overview of Human Neutrophil Antigens [16].

Antigen System

Carrier glycoprotein CD Antigen Allele Gene

HNA-1 FcγRIIIb CD16b HNA-1a FCGR3B*01

HNA-1b FCGR3B*02 See table 2

HNA-1c FCGR3B*03

FcγRIIIb deficient Null allele FcRIIIb gene deficient

HNA-2 CD177 glycoprotein CD177 HNA-2* CD177*01

CD177 deficient incorrect splicing process**

HNA-3 CTL2 HNA-3a SLC44A2*01 SLC44A2*461G

HNA-3b SLC44A2*02 SLC44A2*461A

HNA-4 MAC-1;CR3;αmβ2-integrin

CD11b HNA-4a ITGAM*01 ITGAM*230G

ITGAM*02 ITGAM*230A

HNA-5 LFA-1;αLβ2-integrin CD11a HNA-5a ITGAL*01 ITGAL*2372G

ITGAL*02 ITGAL*2372C

* In HNA-2 positive individuals negative neutrophil subpopulations are present due to lack of gene transcription caused by three mutations: A793C, G1084A and possibly C49G. Furthermore, atypical HNA-2 expression causing two distinct HNA-2 positive neutrophil populations can be due to the mutations A134T, G156A and G1333A [91].**Incorrect splicing process generating premature stop codons [92].

Table 2: HNA-1 polymorfisms

Nucleotide/AAPosition/position

HNA-1aNucl/AA

HNA-1bNucl/AA

HNA-1cNucl/AA

141/36 G/Arg C/Ser C/Ser227/65 A/Asn G/Ser G/Ser266/78 C/Ala C/Ala A/Asp277/82 G/Asp A/Asn A/Asn349/106 G/Val A/Ile A/Ile

and especially the strong neutrophil agglutinating HNA-3a antibodies have been implicated in TRALI [16,21,22]. This is somewhat biased as antibody detection and identification until recently were mostly limited to these antigens as antibody iden-tification panels (including antigen negative donor neutrophil suspensions) for HNA-3b, 4a and 5a were not available.

Muller.indd 229 25-7-2014 14:40:35

Chapter 13

230

Detection of antibodies involved in TRALI

Guidelines for the detection of granulocyte specific antibodies are given by the Inter-national Granulocyte Immunology Working party [23]. The use of the Granulocyte ImmunoFluorescence Test (GIFT)[24], the Granulocyte Agglutination Test (GAT)[25] and the Monoclonal Antibody Immobilization of Granulocyte Antigens (MAIGA)[26] assay are advised. In the GIFT, antibody binding on complete neutrophils is tested. In the GAT, microscopically visible agglutination of neutrophils in the presence of the investigated serum is scored. Both the GIFT and the GAT can be used to screen for the presence of antibodies and identification of these antibodies depending on the typed donor panels used. Both tests are scored semi-quantitatively with the possible reaction strength variation of weak to strong positive ((+), 1+, 2+, 3+ and 4+). In the GAT agglutination of neutrophils can be scored. This agglutination is seen for all HNA antibodies and for several HLA antibodies. The MAIGA is a glycoprotein specific ELISA based on the Monoclonal Antibody Immobilization of Platelet Anti-gen (MAIPA), a test used to identify platelet-reactive antibodies [27]. MAIGA can be used for the identification of all HNA antibodies, except for HNA-3 antibodies as Choline Transporter-Like protein 2 (CTL2, the carrier for HNA-3), specific mon-oclonal antibodies that can be used in the MAIGA are not available yet. In most laboratories screening and identification of granulocyte specific antibodies in TRALI cases was limited to the GIFT and GAT, after which some laboratories confirmed the detected antibodies in the MAIGA. Both for HLA and granulocyte specific antibodies it is assumed that the amount of antibodies in the blood product is of importance to overcome a certain threshold to induce TRALI. For this reason, in several countries female plasma, with a higher risk of containing antibodies due to immunization dur-ing pregnancies, is limited to less than 30 ml per blood product intended for trans-fusion [28-30]. A decrease in reported TRALI cases in the different hemovigilance schemes shows a positive effect of this intervention [31,32]. However, this measure-ment will not prevent all immune-mediated TRALI as TRALI-cases after transfusions of blood products containing less than 30 ml plasma have been reported [16,33].It seems also plausible that HNA antibody reaction strength plays a key-role in the occurrence of TRALI, but until now no granulocyte specific antibody reac-tion strength studies have been performed. Most granulocyte specific antibodies

Muller.indd 230 25-7-2014 14:40:35


231

implicated in TRALI cases are IgG antibodies. Quite often weak, a-specific granulo-cyte reactive IgM antibodies are detected in sera of TRALI associated donors and random screened donors, but our personal opinion is that these (very) weak a-spe-cific antibodies do not cause TRALI [15].Both HLA-class I and –class II antibodies can cause TRALI. Class I antibodies can acti-vate neutrophils [34,35] by binding to their cognate antigens expressed on neutro-phils or on endothelial cells after which neutrophils can be ‘trapped’ by binding via the Fc receptors. However, a recent study in mice showed that macrophages and complement activation are important in the onset of HLA Class I induced TRALI [36]. HLA class II antigens are not expressed on neutrophils but can cause TRALI by bind-ing to the recipient monocytes leading to activation, release of soluble mediators and activation of neutrophils [37-40]. Frequently detected HLA antibodies in TRALI implicated donors have specificity for HLA-A2 and HLA-DR4 [21,37] which is partly due to the high antigen frequency (in the Caucasian population 48% is HLA-A2 and 23% HLA-DR4 positive) causing frequent exposure to the antigens in pregnancy and after transfusion or transplantation. However, this cannot be the only explanation for the high frequency of these antibodies as other antibodies against equally fre-quent antigens are detected less in TRALI implicated donors.Several different antibody detection techniques, such as the Lymphocyte Immu-noFluorescence Test (LIFT)[41], HLA antibody ELISAs [42], Complement Depend-ent Cytotoxicity (CDC) assay [43,44] and HLA specific flow cytometry beads assays [45,46], are available for detection of HLA antibodies. In many laboratories, the easy to perform, flow cytometric beads assays replaced the CDC in the past five to ten years. Both beads with multiple HLA antigens for the screening of HLA class I and II antibodies and single antigen beads for antibody identification are available [47]. Results are expressed in mean immunofluorescence intensity (MFI). Since the intro-duction of these beads assays many more HLA antibodies are detected. The high sensitivity can be due to the concentration of antigens on the beads but seem to be strongly dependent of the cut-off values used. The specificity and clinical impor-tance of these extra antibodies is still unclear and needs to be studied [48-51].A recently published study by Hashimoto et al. [52] shows significant stronger mean HLA antibody reaction strength in donors implicated in TRALI cases compared with a nonhemolytic transfusion reaction (NHTR) group. This is a first indication that

Muller.indd 231 25-7-2014 14:40:35

Chapter 13

232

antibody reaction strength is important, but although the means differed signifi-cantly, individual antibodies varied in the TRALI group from strong to weak reactive and further studies are necessary.To confirm the imputability of the detected antibodies, a cross match between leu-cocytes (lymphocytes and neutrophils) of the patient and serum of the donors is necessary. Cross matching with neutrophils of the patient can be done in the GIFT, GAT and MAIGA. Cross matching with lymphocytes is not possible in the beads assays, and must be done in the CDC or LIFT.If cross matching is not possible, HLA and HNA genotyping can show if the patient is positive for the cognate antigens for the detected antibodies in donor blood.

Donor characteristics

Important in prevention of immune-mediated TRALI is to understand which donors have a high incidence of HLA or HNA antibodies. Donor risk factors for HLA Class I and HLA Class II antibody formation include allo-exposure to white blood cells. Tak-ing this into account two groups could be at risk; multiparous donors and donors exposed to blood transfusion.

Multiparous donorsDensmore et al. investigated whether the likelihood of HLA allo-immunization increases with the number of pregnancies. In multiple reports it is shown that the likelihood of HLA allo-immunization increases with the number of pregnancies from 1-9% in the absence of previous pregnancies up to 32-38% of multiparous women harboring HLA antibodies [15,53,54]. The clinical significance of donor gender was furthermore demonstrated in two studies in critically ill patients reporting wors-ened oxygenation after FFP transfusion from (multiparous) female donors [6,55]. From these results it can be concluded that exclusion of female or more specific multi-parous women from high volume plasma products may result in prevention of TRALI. Effect of such strategies will be discussed later on in this review.

Muller.indd 232 25-7-2014 14:40:35


233

Donors previously exposed to transfusionAllo-exposure by transfusion can induce antibody formation in donors, however the prevalence of HLA antibodies in previously transfused donors is low [54]. For the prevention of variant Creuzfeldt-Jakobs disease, all individuals transfused after 1980 have been excluded from donation in the UK since 2004 and in The Netherlands since February 2005. However, this is not a wide spread donor exclusion strategy and due to the low prevalence of allo-immunization it is unlikely that a significant percentage of donors harbouring antibodies is excluded.HLA antibody formation after blood transfusion occurs from exposure to HLA antigens present on the transfused leucocytes. From this point of view it could be hypothesized that leucocyte-reduced blood components, would reduce the rates of allo-immunization. A meta-analysis however showed that allo-immunization varied considerably between studies and range from 7% to 44% among recipients of leu-cocyte-reduced blood transfusions and from 20% to 50% among control recipients of non-leucoreduced blood components [56]. Therefore leucoreduction of blood components is not likely to reduce TRALI due to HLA antibodies from previously transfused donors. Other factors that influence the rate of HLA allo-immunization from transfusion include the number of units transfused, the underlying clinical con-dition resulting in transfusion, the time since transfusion and the method used for detecting HLA antibodies [57,58].Until recently, the prevalence of HLA antibodies in transfused donors was not well characterized. However, a recent study obtained from 7,920 donors (2,086 males and 5,834 females) transfusion and pregnancy history and tested them for HLA Class I and Class II antibody presence [53]. Of interest, HLA antibody prevalence did not significantly differ between 895 transfused (1.7%) and 1138 non-transfused males (1.0%), [OR 1.75; 95%CI 0.80-3.82]. Prevalence in 45 transfused nulliparous females (4.4%) was not statistically different from the 1.6% prevalence in 1732 non-trans-fused nulliparous females [OR 2.94, 95% CI 0.68- 12.74]. However, Middelburg et al. showed a positive association between transfusion and the presence of leucocyte antibodies in 148 transfused donors in a total cohort of 6034 tested donors. They showed the highest risk difference (3.0) for granulocyte specific antibodies and an overall risk difference of 5.8 for any leucocyte antibodies after transfusion [15].

Muller.indd 233 25-7-2014 14:40:35

Chapter 13

234

Epidemiology of immune-mediated TRALI

According to the consensus definition TRALI develops within 6 hours of a transfu-sion, however some authors advocate an extension of the definition with a “delayed TRALI syndrome”. This delayed form of TRALI that can arise up to 72 hours after transfusion, is thought to result from non-immune mediated TRALI in the majority of the cases [59]. However, clinically it is impossible to distinguish if a case of TRALI is immune or non-immune mediated.Although, from a blood bank view, immunologic work-up of reported TRALI cases and involved donors reveals the approximate percentage of immune-mediated TRALI, still these data should be considered with caution as reporting is often passive and different definitions of TRALI are applied, sometimes even dependent whether the immunologic work-up was positive. However, in general it is assumed that up to two thirds of TRALI-cases can be explained by presence of HLA or HNA antibodies. In line with this, the exclusion of female donors (the highest risk population among donors to have HLA-HNA antibodies) from donation of high plasma volume products has resulted in a decrease of approximately 66% of reported TRALI cases [60-62]. Probably, these numbers still do not reflect the real incidence of immune- and non-immune-mediated TRALI as most of the reports are based on passive reporting to the bloodbank, which was shown to be non-representative [63].Taking into account these limitations, immune-mediated TRALI reflects approxi-mately two thirds of all TRALI cases. Studies reporting TRALI-cases show donor anti-bodies against cognate antigens present on leucocytes of the affected recipients for approximately 25-41% of HLA Class I antibodies, in 52-68% of HLA Class II antibodies, and 8-28% for granulocyte antibodies [21,64-66].

Prevention of immune-mediated TRALI from a bloodbank perspective

Excluding donorsThe alarming TRALI figures in haemovigilance schemes were reason for blood sup-ply organizations to introduce TRALI preventive measures based on donor exclu-sion strategies [28,32,67]. Based on results of TRALI series and observational cohort

Muller.indd 234 25-7-2014 14:40:35


235

Table 3: Overview of results of male only donor strategies

Reference Type of studyand inclusion

Popula-tion

Country Study year Endpoint Effect size Effec-tive

Palfi [55]

RCTactive

ICU Sweden 1995-1997 P/F N/A yes

Wright [62]

Retrospectiveactive

Surgery UK 1998-2006 onset TRALI

OR (95% CI) 0.39 (0.16-0.90)

yes

SHOT [28]

Retrospective passive

National UK 2002-2005 onset TRALI

N/A yes

Vlaar [7]

Retrospectiveactive

ICU Nether-lands

2004-2007 onset TRALI

RR (95% CI) 0.35 (0.14-0.88)

yes

Eder [78]

Retrospectivepassive

National US 2006-2008 onset TRALI

OR (95% CI) 0.21 (0.08-0.45)

yes

Wiersum [61]

Retrospective passive

National Nether-lands


PAR (95% CI) 0.33 (0.09-0.51)

yes

Vlaar [81]

Retrospectiveactive

Surgery Nether-lands


N/A no

Nakazawa [93]

Prospectiveactive

Surgery Japan 2008-2008 P/F< 300 N/A yes

Toy [94]

Prospective General hospital popula-tion

United States


Incidence (95% CI) before 2.57 (1.72-3.86), after 0.81 (0.44-1.49)*

yes

P/F=PaO2-FiO2 ratio, RCT=randomized controlled trial, PAR= population-attributable risk, RR=relative risk, CI=confidence interval, N/A=non-applicable, * TRALI incidence % (95% CI) per 10.000 units transfused before (2006) and after (2009) introduction of a male only donor strategy.

studies [18,27] exclusion of blood components with HLA and/or HNA antibodies containing high plasma volumes was considered. As large scale antibody screen-ing was laborious and expensive the known high HLA immunization rate for ever pregnant women was reason for many blood centers, starting in the UK in 2003, to preferentially use plasma from male donors for the production of high plasma volume blood components (whole blood, FFP, multi donor buffy-coat platelet con-centrates). This preventive transfusion strategy resulted in a decline in antibody

Muller.indd 235 25-7-2014 14:40:35

Chapter 13

236

mediated TRALI cases as was shown in different haemovigilance schemes (see also table 3)[60-62,68,69].

Testing donors for antibodiesA less rigorous policy than the above mentioned one is testing all donors or at risk donors for presence of HLA or HNA antibodies. Except for the high labor and costs involved, large scale HLA and HNA antibody screening possibilities were not readily available in 2003. In recent years this partly changed with the introduction of the beads based flow cytometry HLA antibody screening techniques. Unfortunately, in most laboratories the introduction of these techniques was not in time to influence the TRALI prevention policy and screening for HLA class I and HLA class II antibodies is only used (in a growing number of laboratories) on top of the female exclusion strategy for the female apheresis blood components. The apheresis donation loss due to positive results in these beads based antibody screening techniques largely depends on the number of pregnancies and assay cut-offs used. A recent study by Carrick et al. showed a donation loss varying from 0.9% to 5.8% [70].Up to now, no large scale antibody screening techniques for the clinically important HNA are available. Only a few blood centers test female apheresis donors for the presence of antibodies against HNA-1a, 1b, 2a and 3a.

PoolingIn order to prevent immune-mediated TRALI, high plasma volume containing blood products can be prepared from a pool of multiple donors. By pooling of up to hun-dreds donations, dilution of any leucocyte antibodies present occurs, in addition antibodies may be neutralized by soluble HLA antigens in the pool. Concerns of pooling are multiple donor exposure and transmission of viruses and prion diseases. Therefore Solvent/Detergent treatment is a prerequisite when pooled plasma is used.Solvent Detergent plasma (S/D plasma) is prepared from pools of 500 to 1600 single donor units. Sachs et al. screened 20 batches of S/D plasma and could not detect HLA class I and II antibodies, nor did they demonstrate granulocyte-agglutinating antibodies [71]. Norway implemented the use of S/D plasma (Octaplas®) in 1993 and no cases of TRALI have been reported after the use of more than 300.000 units

Muller.indd 236 25-7-2014 14:40:35


237

[29,72]. Of note, it has not been determined if S/D plasma also prevents TRALI in recipients with a lowered threshold (e.g. critically ill) to develop TRALI, since even small amounts of antibodies might elicit TRALI in these patients. In addition, the use of S/D plasma is costly. In the UK it was calculated that replacement of FFP by S/D plasma in order to prevent 107 cases of TRALI would cost up to £9.2 million/year [73]. Since this report, a male-only FFP policy was instituted in the UK resulting in a striking reduction of TRALI, making the use of S/D plasma to prevent a case of TRALI even more costly.Platelets can be collected by apheresis from a single donor or from pools of buffy-coat derived platelets. Single donor apheresis is preferred in the US, while in Europe it is more common to use pooled platelets. To prevent immune-mediated TRALI from apheresis platelets, allo-exposed donors are tested and those harboring leu-cocyte antibodies are deferred [74]. The deferral of all female platelet apheresis donors, as with plasma donations, is likely to hamper the adequate supply of plate-lets and is considered not feasible [75]. Alternatively, platelets can be derived from whole blood and pooled, to overcome the risk of TRALI by dilution of potential leucocyte antibodies present in any of the donors’ plasma, the platelets should be re-suspended in male plasma or platelet additive solution (PAS). However, multiple donor exposures increase the risk of viral infection and whether pooled platelets carry a lower TRALI risk compared to apheresis platelets, remains to be determined. Recently, the French Hemovigilance Network reported 18 cases of TRALI due to apheresis platelets and none after the transfusion of pooled platelets [76]. However, a systematic review based on the limited available data demonstrated that the risk was approximately equal for both products [77].In conclusion, the use of S/D plasma may prevent immune-mediated TRALI due to plasma transfusion, but implementation is currently not cost-effective. Further-more, it is not clear whether the beneficial effect also holds true for high risk TRALI patient populations such as critically ill. It remains to be determined if TRALI risk dif-fers among pooled and apheresis platelets. Although, current measures to prevent immune-mediated TRALI by deferral of antibody positive or allo-exposed apheresis donors is proving effective [21,78].

Muller.indd 237 25-7-2014 14:40:35

Chapter 13

238

Prevention of immune-mediated TRALI from a bedside perspective

Recognition and awareness of TRALITo prevent TRALI, identification of donors implicated in a case of TRALI is essential; therefore recognition of TRALI by clinicians is vital. Furthermore, diagnostic work-up needs to be complete and donors should only be deferred if causality is proven by antibody testing including cross-matching or recipient typing. However, among clinicians there is a lack of awareness of TRALI and underreporting is confirmed by look back investigations and retrospective studies [6,12]. Lack of knowledge about TRALI among clinicians has shown to be an important contributor to underreporting [79]. Furthermore, despite the consensus definition [1] differences in local diagnos-tic criteria still exist (e.g., use of imputability criteria in certain countries). Also, even if the consensus definition is applied mild cases and patients who develop symptoms more than 6 hours after transfusion will be missed. It should be kept in mind that the consensus definition does not exclude inter-observer variability in classifying cases of respiratory distress after transfusion. Respiratory deterioration can be attributed to other causes (e.g., trauma, sepsis or pneumonia), which may lead to misclassifica-tion of TRALI cases, in particular in ICU patients. TRALI is considered to be a diag-nosis of exclusion by some physicians and a pre-transfusion inflammatory condition is even a reason to withhold from reporting of a suspected TRALI case [63]. In the near future, increasing knowledge and subsequent recognition of TRALI among cli-nicians is warranted in order to prevent multiple TRALI cases from a single donor. To achieve this it is important to educate physicians on the incidence, pathogenesis and diagnosis of TRALI. Furthermore, insight how to process a suspected TRALI case through the haemovigilance schemes in the hospitals might also be a good source for education purposes.

Identifying patients with first hitAs stated earlier TRALI incidence is higher in critically ill patients [5,7]. These patients are frequently exposed to allogenic blood products, in addition they often have an underlying condition that induces priming of neutrophils (e.g., sepsis, cardiac sur-gery or the need for mechanical ventilation) in pulmonary capillaries. These primed neutrophils may be prone to HLA and/or HNA antibodies in the blood product, which

Muller.indd 238 25-7-2014 14:40:35


239

mediate activation and additional lung injury after transfusion. Patient factors asso-ciated with an increased risk of ALI after transfusion, include sepsis, mechanical ven-tilation, trauma and hematological malignancies [5,80]. This also applies to patients undergoing cardiac surgery, in whom TRALI significantly contributes to an adverse outcome [81]. However, most of these factors cannot be modified, but measures that attenuate ALI possibly raise the threshold for TRALI. Protective mechanical ven-tilation strategies using low tidal volumes reduce mortality with 22% in ALI/ARDS [82] and injurious mechanical ventilation with high tidal volumes aggravates experi-mental TRALI [83]. Therefore, it seems reasonable to apply lung protective ventila-tion strategies in patients subjected to a blood transfusion.Transfusion-related risk factors for the development of TRALI are the amount of blood products, larger volumes of plasma containing products, donor sex and par-ity. In patients with an increased risk for TRALI (e.g., those with a ‘first hit’), it might be plausible to introduce a ‘tailor made’ transfusion policy. This would include transfusion of plasma containing components from male only donors. Of note, such strategy has been shown to reduce the incidence of TRALI in the general patient population [21,61]. However, in a cohort of cardiac surgery patients use of male plasma only was not shown to reduce TRALI incidence [83]. It should be mentioned that this study was not designed to demonstrate such an effect and the number of included patients was relatively low.In conclusion, host factors contribute to increased susceptibility for immune-medi-ated TRALI. Although the underlying condition cannot be modified, supportive care and treatment modalities need to be applied cautiously. To date, evidence for utility of a ‘tailor made transfusion’ strategy in high-risk patients is lacking.

Restrictive transfusion practiceFrom a bedside perspective, the most important measure to prevent TRALI is to limit patient’s exposure to allogenic blood products. In critically ill, the incidence of ALI is reduced when a restrictive transfusion policy is applied [84] and such strategy is associated with decreased mortality [85].Blood products highly associated with the onset of immune-mediated TRALI are high volume plasma products such as platelet concentrates and fresh frozen plasma [5]. Prevention of immune-mediated TRALI should be focused on minimizing the use

Muller.indd 239 25-7-2014 14:40:35

Chapter 13

240

of these blood products. Although it is not feasible to completely avoid the use of these blood products in daily practice, the use of plasma rich blood components can be reduced by using alternatives. A recent trial showed that the use of tranexamic acid in trauma patients significantly reduced mortality [86]. Furthermore, the use of FFP should be limited to bleeding patients, since evidence for prophylactic use of FFP to prevent bleeding is lacking. However, a recent survey among UK ICUs indicated that up to 50% of FFP is administered to non-bleeding patients, either to prevent bleeding in patients with prolonged coagulation parameters or to reverse warfarin therapy [87]. In addition, despite a more restrictive transfusion trigger in the criti-cally ill, multiple unit transfusion to correct for anaemia is still common practice.The inappropriate use of blood products can be limited by the use of computerized transfusion algorithms [88]. In trauma and cardiac surgery, the use of point of care coagulation tests (e.g. ROTEM®) combined with transfusion algorithms also reduces the amount of transfused blood [89,90].In order to prevent immune-mediated TRALI and to limit inappropriate use of blood products, transfusion algorithms need to be implemented and clinicians’ awareness of appropriate and restrictive use of these products needs to be heightened.

Conclusion

TRALI is a serious complication of transfusion and significantly increases morbidity and mortality. Currently, from a blood bank perspective two major strategies exist which have proven to prevent immune-mediated TRALI. The first strategy is aimed at exclusion of allo-exposed donors from donating high volume plasma containing blood products, the second strategy is aimed at dilution of antibodies present in high volume plasma products by means of pooling. From a bedside perspective, TRALI can be prevented by limiting inappropriate use of blood products. A future bedside perspective might involve tailor made blood products for patients at risk for developing TRALI.

Muller.indd 240 25-7-2014 14:40:35


241

References


2. Holness L, Knippen MA, Simmons L, et al.: Fatalities caused by TRALI. Transfus Med Rev 2004;18: 184-188

3. Kleinman S, Caulfield T, Chan P, et al.: Toward an understanding of transfusion-related acute lung injury: statement of a consensus panel. Transfusion 2004;44: 1774-1789

4. Stainsby D, Jones H, Asher D, et al.: Serious hazards of transfusion: a decade of hemov-igilance in the UK. Transfus Med Rev 2006;20: 273-282



7. Vlaar AP, Binnekade JM, Prins D, et al.: Risk factors and outcome of transfusion-related acute lung injury in the critically ill: A nested case-control study. Crit Care Med 2010;38: 771-778




11. Popovsky MA, Abel MD, Moore SB: Transfusion-related acute lung injury associated with passive transfer of antileukocyte antibodies. Am Rev Respir Dis 1983;128: 185-189

12. Kopko PM, Marshall CS, MacKenzie MR, et al.: Transfusion-related acute lung injury: report of a clinical look-back investigation. JAMA 2002;287: 1968-1971


14. Endres RO, Kleinman SH, Carrick DM, et al.: Identification of specificities of antibodies against human leukocyte antigens in blood donors. Transfusion 2010;50: 1749-1760

15. Middelburg RA, Porcelijn L, Lardy N, et al.: Prevalence of leucocyte antibodies in the Dutch donor population. Vox Sang 2011;100: 327-35

16. Reil A, Keller-Stanislawski B, Gunay S, et al.: Specificities of leucocyte alloantibodies in transfusion-related acute lung injury and results of leucocyte antibody screening of blood donors. Vox Sang 2008;95: 313-317

17. Blumberg N, Heal JM, Gettings KF, et al.: An association between decreased cardio-pulmonary complications (transfusion-related acute lung injury and transfusion-asso-ciated circulatory overload) and implementation of universal leukoreduction of blood transfusions. Transfusion 2010;50: 2738-44

Muller.indd 241 25-7-2014 14:40:35

Chapter 13

242

18. Moritz E, Norcia AM, Cardone JD, et al.: Human neutrophil alloantigens systems. An Acad Bras Cienc 2009;81: 559-569

19. Muschter S, Berthold T, Greinacher A: Developments in the definition and clinical impact of human neutrophil antigens. Curr Opin Hematol 2011;18: 452-460

20. de Haas M, Kleijer M, van Zwieten R, et al.: Neutrophil Fc gamma RIIIb deficiency, nature, and clinical consequences: a study of 21 individuals from 14 families. Blood 1995;86: 2403-2413


22. van Stein D, Beckers EA, Sintnicolaas K, et al.: Transfusion-related acute lung injury reports in the Netherlands: an observational study. Transfusion 2010;50: 213-220

23. Bierling P, Bux J, Curtis B, et al.: Recommendations of the ISBT Working Party on Granu-locyte Immunobiology for leucocyte antibody screening in the investigation and pre-vention of antibody-mediated transfusion-related acute lung injury. Vox Sang 2009;96: 266-269

24. Verheugt FW, von dem Borne AE, Decary F, et al.: The detection of granulocyte alloan-tibodies with an indirect immunofluorescence test. Br J Haematol 1977;36: 533-544

25. Jiang AF, Lalezari P: A micro-technique for detection of leukocyte agglutinins. J Immunol Methods 1975;7: 103-108

26. Bux J, Kober B, Kiefel V, et al.: Analysis of granulocyte-reactive antibodies using an immunoassay based upon monoclonal-antibody-specific immobilization of granulocyte antigens. Transfus Med 1993;3: 157-162

27. Kiefel V, Santoso S, Weisheit M, et al.: Monoclonal antibody--specific immobilization of platelet antigens (MAIPA): a new tool for the identification of platelet-reactive antibod-ies. Blood 1987;70: 1722-1726

28. Shot Annual report 2005: www.shotuk.org: SHOT annual report 2005;29. Flesland O: A comparison of complication rates based on published haemovigilance

data. Intensive Care Med 2007;33: S17-S2130. Funk MB, Guenay S, Lohmann A, et al.: Benefit of transfusion-related acute lung injury

risk-minimization measures - German haemovigilance data (2006-2010). Vox Sang 2012;102: 317-23

31. Knowles S, Cohen H: The 2010 Annual SHOT report (2011); in: 2011.32. Schipperus MR, Wiersum-Osselton JC: On behalf of the TransfusieReacties In Patienten

(TRIP) Steering Group. TRIP rapport 2009 hemovigilantie. www.tripnet.nl.; in: 2009.33. Win N, Chapman CE, Bowles KM, et al.: How much residual plasma may cause TRALI?

Transfus Med 2008;18: 276-28034. Kelher MR, Masuno T, Moore EE, et al.: Plasma from stored packed red blood cells and

MHC class I antibodies causes acute lung injury in a 2-event in vivo rat model. Blood 2009;113: 2079-2087

Muller.indd 242 25-7-2014 14:40:35


243

35. Takahashi D, Fujihara M, Azuma H, et al.: Stimulation of human neutrophils with sera containing HLA Class I alloantibody causes preferential degranulation of azurophilic granules and secretory vesicles. Vox Sang 2010;98: 560-566

36. Strait RT, Hicks W, Barasa N, et al.: MHC class I-specific antibody binding to nonhemat-opoietic cells drives complement activation to induce transfusion-related acute lung injury in mice. J Exp Med 2011;208: 2525-44

37. Bux J: Antibody-mediated (immune) transfusion-related acute lung injury. Vox Sang 2011;100: 122-128

38. Nishimura M, Hashimoto S, Takanashi M, et al.: Role of anti-human leucocyte antigen class II alloantibody and monocytes in development of transfusion-related acute lung injury. Transfus Med 2007;17: 129-134

39. Sachs UJ: Recent insights into the mechanism of transfusion-related acute lung injury. Curr Opin Hematol 2011;18: 436-442

40. Sachs UJ, Wasel W, Bayat B, et al.: Mechanism of transfusion-related acute lung injury induced by HLA class II antibodies. Blood 2011;117: 669-677

41. Décary F: A look at HLA antisera in the indirect immunofluorescence technique (LIFT). Histocompatibility testing 1975; 380-90

42. Uboldi de CM, Pratico L, Curtoni ES: Comparison of different techniques for detection of anti-HLA antibodies in sera from patients awaiting kidney transplantation. Eur J Immu-nogenet 2002;29: 379-382

43. Phelan DL, Rodey GE, Anderson CB: The development and specificity of antiidiotypic antibodies in renal transplant recipients receiving single-donor blood transfusions. Transplantation 1989;48: 57-60

44. Terasaki PI, McClelland JD: Microdroplet assay of human serum cytotoxins. Nature 1964;204: 998-1000

45. Pei R, Lee J, Chen T, et al.: Flow cytometric detection of HLA antibodies using a spectrum of microbeads. Hum Immunol 1999;60: 1293-1302

46. Wahrmann M, Exner M, Haidbauer B, et al.: [C4d]FlowPRA screening--a specific assay for selective detection of complement-activating anti-HLA alloantibodies. Hum Immu-nol 2005;66: 526-534

47. Colombo MB, Haworth SE, Poli F, et al.: Luminex technology for anti-HLA antibody screening: evaluation of performance and of impact on laboratory routine. Cytometry B Clin Cytom 2007;72: 465-471

48. Bray RA, Gebel HM: Strategies for human leukocyte antigen antibody detection. Curr Opin Organ Transplant 2009;14: 392-397

49. Claas FH, Doxiadis II: Human leukocyte antigen antibody detection and kidney alloca-tion within Eurotransplant. Hum Immunol 2009;70: 636-639

50. Doxiadis II, Roelen D, Claas FH: Mature wines are better: CDC as the leading method to define highly sensitized patients. Curr Opin Organ Transplant 2010;15: 716-9

51. Zoet YM, Brand-Schaaf SH, Roelen DL, et al.: Challenging the golden standard in defin-ing donor-specific antibodies: does the solid phase assay meet the expectations? Tissue Antigens 2011;77: 225-228

Muller.indd 243 25-7-2014 14:40:35

Chapter 13

244

52. Hashimoto S, Nakajima F, Kamada H, et al.: Relationship of donor HLA antibody strength to the development of transfusion-related acute lung injury. Transfusion 2010;50: 2582-91

53. Kakaiya RM, Triulzi DJ, Wright DJ, et al.: Prevalence of HLA antibodies in remotely trans-fused or alloexposed volunteer blood donors. Transfusion 2010;50: 1328-1334



56. Vamvakas EC: Meta-analysis of randomized controlled trials of the efficacy of white cell reduction in preventing HLA-alloimmunization and refractoriness to random-donor platelet transfusions. Transfus Med Rev 1998;12: 258-270

57. Gleichmann H, Breininger J: Over 95 per cent sensitization against allogeneic leukocytes following single massive blood transfusion. Vox Sang 1975;28: 66-73

58. Lubenko A, Rodi KM: The detection by enzyme-linked immunosorbent assays of non-complement-fixing HLA antibodies in transfusion medicine. Transfusion 1998;38: 41-44

59. Marik PE, Corwin HL: Acute lung injury following blood transfusion: expanding the defi-nition. Crit Care Med 2008;36: 3080-3084

60. Vlaar AP, Binnekade JM, Schultz MJ, et al.: Preventing TRALI: ladies first, what follows? Crit Care Med 2008;36: 3283-3284



63. Vlaar AP, Wortel K, Binnekade JM, et al.: The practice of reporting transfusion-related acute lung injury: a national survey among clinical and preclinical disciplines. Transfu-sion 2009;50: 443-451

64. Curtis BR, McFarland JG: Mechanisms of transfusion-related acute lung injury (TRALI): anti-leukocyte antibodies. Crit Care Med 2006;34: S118-S123

65. Keller-Stanislawski B, Reil A, Gunay S, et al.: Frequency and severity of transfusion-related acute lung injury--German haemovigilance data (2006-2007). Vox Sang 2010;98: 70-77

66. Wiersum-Osselton JC, Porcelijn L, van Stein D., et al.: Transfusion-related acute lung injury (TRALI) in the Netherlands in 2002-2005. Ned Tijdschr Geneeskd 2008;152: 1784-1788

67. Annual Hemovigilance Report Swissmedic 2008, http://www.swissmedic.ch: in: 2011.

Muller.indd 244 25-7-2014 14:40:35


245


69. Lin Y, Saw CL, Hannach B, et al.: Transfusion-related acute lung injury prevention meas-ures and their impact at Canadian Blood Services Transfusion 2012;52: 567-74

70. Carrick DM, Norris PJ, Endres RO, et al.: Establishing assay cutoffs for HLA antibody screening of apheresis donors. Transfusion 2011;51: 2092-2101

71. Sachs UJ, Kauschat D, Bein G: White blood cell-reactive antibodies are undetectable in solvent/detergent plasma. Transfusion 2005;45: 1628-1631

72. Flesland O, Seghatchian J, Solheim BG: The Norwegian Plasma Fractionation Project--a 12 year clinical and economic success story. Transfus Apher Sci 2003;28: 93-100

73. Riedler GF, Haycox AR, Duggan AK, et al.: Cost-effectiveness of solvent/detergent-treated fresh-frozen plasma. Vox Sang 2003;85: 88-95


75. Rios JA, Schlumpf KS, Kakaiya RM, et al.: Blood donations from previously transfused or pregnant donors: a multicenter study to determine the frequency of alloexposure. Transfusion 2011;51: 1197-1206

76. Ozier Y, Muller JY, Mertes PM, et al.: Transfusion-related acute lung injury: reports to the French Hemovigilance Network 2007 through 2008. Transfusion 2011;51: 2102-2110

77. Vamvakas EC: Relative safety of pooled whole blood-derived versus single-donor (apheresis) platelets in the United States: a systematic review of disparate risks. Trans-fusion 2009;49: 2743-2758

78. Eder AF, Herron RM, Jr., Strupp A, et al.: Effective reduction of transfusion-related acute lung injury risk with male-predominant plasma strategy in the American Red Cross (2006-2008). Transfusion 2010;50: 1732-1742

79. Kram R, Loer SA: Transfusion-related acute lung injury: lack of recognition because of unawareness of this complication? Eur J Anaesthesiol 2005;22: 369-372



82. The Acute Respiratory Distress Syndrome Network:Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respira-tory distress syndrome. N Engl J Med 2000;342: 1301-1308

83. Vlaar AP., Wolthuis EK, Hofstra JJ, et al.: Mechanical ventilation aggravates transfu-sion-related acute lung injury induced by MHC-I class antibodies. Intensive Care Med 2010;36: 879-87

Muller.indd 245 25-7-2014 14:40:35

Chapter 13

246

84. Yilmaz M, Keegan MT, Iscimen R, et al.: Toward the prevention of acute lung injury: protocol-guided limitation of large tidal volume ventilation and inappropriate transfu-sion. Crit Care Med 2007;35: 1660-1666

85. Hebert PC, Wells G, Blajchman MA, et al.: A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999;340: 409-417

86. Shakur H, Roberts I, Bautista R, et al.: Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemor-rhage (CRASH-2): a randomised, placebo-controlled trial. Lancet 2010;376: 23-32

87. Walsh TS, Stanworth SJ, Prescott RJ, et al.: L Prevalence, management, and outcomes of critically ill patients with prothrombin time prolongation in United Kingdom intensive care units. Crit Care Med 2010;38: 1939-1946

88. Rana R, Afessa B, Keegan MT, et al.: Evidence-based red cell transfusion in the criti-cally ill: quality improvement using computerized physician order entry. Crit Care Med 2006;34: 1892-1897


90. Shore-Lesserson L, Manspeizer HE, DePerio M, et al.: Thromboelastography-guided transfusion algorithm reduces transfusions in complex cardiac surgery. Anesth Analg 1999;88: 312-319

91. Wolff J, Brendel C, Fink L, et al.: Lack of NB1 GP (CD177/HNA-2a) gene transcription in NB1 GP- neutrophils from NB1 GP-expressing individuals and association of low expres-sion with NB1 gene polymorphisms. Blood 2003;102: 731-733

92. Kissel K, Scheffler S, Kerowgan M, et al.: Molecular basis of NB1 (HNA-2a, CD177) defi-ciency. Blood 2002;99: 4231-4233

93. Nakazawa H, Ohnishi H, Okazaki H, et al.: Impact of fresh-frozen plasma from male-only donors versus mixed-sex donors on postoperative respiratory function in surgical patients: a prospective case-controlled study. Transfusion 2009;49: 2434-2441

94. Toy P, Gajic O, Bacchetti P, et al.: Transfusion related acute lung injury: incidence and risk factors. Blood 2012;119:1757-67

Muller.indd 246 25-7-2014 14:40:35

Chapter 14

Low risk TRALI donor strategies and the impact on the onset

of transfusion-related acute lung injury; a meta-analysis

Marcella C.A. Müller*, Danielle van Stein*, Jan M. Binnekade, Dick J. van Rhenen, Alexander P.J. Vlaar

*authors contributed equally

Accepted for publication in Transfusion

Muller.indd 247 25-7-2014 14:40:35

Chapter 14

248

Abstract

Background: Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related mortality. In the past decade blood banks have implemented low risk TRALI donor strategies, including a male only donor policy for plasma con-taining blood products in order to prevent onset of TRALI. We performed a meta-analysis to determine whether use of low risk TRALI donor strategies for plasma indeed reduces onset of TRALI.Study design and Methods: We searched Medline and Cochrane Central Register of Controlled Trials from January 1995 up to January 2013. Two reviewers indepen-dently extracted data on study characteristics, methods, and outcomes. Primary endpoint was onset of TRALI. Subgroup analyses were performed for patient popu-lations prone to develop TRALI and general patient populations.Results: Ten articles were included. Meta-analysis using a random effects model tak-ing into account all transfused products, showed a significant reduction for the risk of TRALI after implementation of low risk TRALI donor strategies (OR 0.61 95%CI 0.42-0.88). Data from patient populations prone to develop TRALI showed a sig-nificant reduction of TRALI risk (OR 0.51 95%CI 0.29-0.90), while data from general patient populations showed a similar non-significant trend (OR 0.66 95%CI 0.40-1.09). Results were similar when taking only plasma products into account (OR 0.62 95% 0.42-0.92).Conclusion: The introduction of low risk TRALI donor strategies for plasma contain-ing products results in a reduction of TRALI.

Muller.indd 248 25-7-2014 14:40:35

The impact of TRALI prevention strategies

249

Introduction

Transfusion-related acute lung injury (TRALI) is a life-threatening complication of blood transfusion. It causes high morbidity and is the leading cause of transfusion-related mortality for the last five years in the US [1]. TRALI is a clinical diagnosis and defined as a new episode of acute lung injury (ALI) occurring during, or within 6 hours after a blood transfusion in the absence of hydrostatic edema [2,3]. TRALI can be the result of a single event (e.g. transfusion), although most TRALI cases are postulated to be a ‘two-event’ entity. The first event is related to the underlying condition of the patient (e.g., infection or surgery) that causes activation of the pul-monary endothelium, leading to the sequestration and priming of neutrophils in the lung. The second event is the transfusion of a blood product, which activates the primed neutrophils in the lung, causing endothelial damage, and subsequently TRALI [4].The transfusion factors can be divided in antibody and non-antibody mediated TRALI. Non-antibody mediated TRALI is caused by transfusion of stored cell con-taining blood products. Pro-inflammatory mediators which have accumulated dur-ing storage or the aged erythrocyte and platelet themselves have been implicated in non-antibody mediated TRALI [4-7]. Antibody mediated TRALI is caused by pas-sive infusion of antibodies which causes neutrophil activation [8]. The latter form of TRALI is the most prevalent type [9,10] and will be subject of the current meta-analysis. Antibodies are present in plasma containing blood products. Plasma from female donors has been particularly implicated in the pathogenesis of TRALI [11-13]. Donor leukocyte reactive antibodies (HLA class I and II and anti-granulocyte (HNA) antibodies) are mainly found in (multiparous) female donors. This might be explained by allo-immunization during pregnancies [14]. Therefore, plasma derived from male donors should less likely cause TRALI than plasma derived from women. In 2003, the National Blood Service (NBS) in the UK was the first to implement a precautionary measure to reduce TRALI by the preferential use of plasma from male donors for transfusion [15]. Nowadays, more blood collection organizations have taken similar measures (table 1). Most of them use only, or preferentially, plasma from male donors for single donor plasma. Some blood collection organizations also implemented a (pre-dominantly) male only donor strategy for other high volume

Muller.indd 249 25-7-2014 14:40:35

Chapter 14

250

plasma products such as platelet concentrates [16]. However, the implementation of these low risk TRALI donor strategies for plasma containing blood products have serious consequences for the blood supply, in particular for the availability of AB plasma. So far, effects of these high volume plasma donor policies have only been published in observational studies [15,17-25].To quantify the impact of low risk TRALI donor strategies for plasma on the onset of TRALI, a meta-analysis of all trials on this topic since 1995 was performed.

Methods

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations was used for this meta-analysis [26]. We did not register our pro-tocol.

Study Selection CriteriaTo identify and review all studies that met the following criteria: observational con-trolled trials of transfusion, plasma, transfusion-related acute lung injury and acute lung injury were sought.

Search StrategyTo identify literature in electronic databases, MEDLINE from January 1, 1995, through January 1, 2013 was searched, by using the following medical subject head-ing (MeSH) terms: blood transfusion, plasma, ARDS and ALI. The following text words were used: TRALI and transfusion related acute lung injury. To identify observational studies, the MeSH terms case-control study and retrospective study were added.

Table 1 Overview of different low risk TRALI donor strategies [44].

Low risk TRALI donor strategiesDonor deferral based on antibody screening Screening all donors for HLA and/or HNA antibodies Screening previously alloexposed donors for HLA and/or HNA antibodiesDonor deferral based on history of pregnancy or history of transfusionDeferral of all female donors

Muller.indd 250 25-7-2014 14:40:35


251

The Cochrane Library (2012), which contains the CENTRAL Database of Controlled Trials, the Database of Abstracts of Review Effectiveness, and the Cochrane Data-base of Systematic Reviews, was also searched.In addition, the related articles feature of PubMed, which identifies related articles by using a hierarchical search engine that is not solely based on MeSH headings, was used. This search was completed with articles selected by 2 of the authors (AV and MM). Although search was also for non–English-language citations, subsequent article review involved only English-language publications.

Study SelectionAfter all citations based on our search strategy were identified, 2 of the authors (AV and MM) independently reviewed each abstract to confirm eligibility. Eligible stud-ies evaluated use of low risk TRALI donor strategies for plasma and onset of lung injury and mortality. Low risk TRALI donor strategies were defined as predominantly or male only donor policy for plasma products, plasma from donors screened for absence of HLA/HNA antibodies and plasma from female donors without history of pregnancy. Both controlled and observational trials were considered eligible. If an abstract was selected as eligible, the same authors independently reviewed the respective article, if available, to confirm that it met inclusion criteria. To resolve discrepancies, the 2 reviewers either had to reach consensus, or use a third reviewer (JB). Inter-observer agreement for study selection was determined by κ, with a value above 0.80 indicating good agreement.

Data ExtractionUsing a predefined data collection form, data from the studies to describe patient characteristics, study methods, and study findings were extracted. All data were abstracted independently by each of the 2 primary reviewers and verified for accu-racy by the third reviewer, again with discussion used to resolve differences among reviewers. Both primary reviewers were physicians with formal training in clinical epidemiology and biostatistics. Corresponding authors of included studies were requested by email to provide missing data for the meta-analysis.

Muller.indd 251 25-7-2014 14:40:35

Chapter 14

252

Data Synthesis and AnalysisPrimary endpoint was the effect of the introduction of a low risk TRALI donor strat-egy for plasma containing products on occurrence of TRALI. Secondary endpoint was the effect of the introduction of a low risk TRALI donor strategy for plasma con-taining products on mortality among TRALI patients. Two subsets were analyzed: 1) the group of studies with data from local registries (patients prone to develop TRALI); 2) the group of studies with data from nationwide registries (general patient population). As patients often receive different types of blood products we analysed the effect of TRALI risk reduction strategies for all transfused products (denomina-tor: total number of transfusions and numerator: all TRALI cases), in addition we assessed the effects of TRALI risk reduction strategies on plasma-induced TRALI (denominator: total number of plasma transfusions and numerator: plasma-induced TRALI cases).The percentage of agreement before discussion among reviewers in study selec-tion, study design, and data abstraction was measured. For data synthesis, evidence tables were constructed, to present data separately for the primary outcome varia-ble: onset TRALI per transfused blood product and the secondary outcome variable: 30 day mortality. Furthermore data were presented based on study population e.g. at risk patients (ICU and surgery) or general hospital/region population.The study design was classified as a randomized clinical trial, cohort study (prospec-tive, retrospective, or historical control), case-control study, or outcomes study (cross-sectional).The methodological quality of eligible studies was assessed using the modified Newcastle-Ottawa scale, a validated instrument designed to evaluate the quality of observational studies in systematic reviews and meta-analyses [27]. The quality of cohort studies and case control studies were assessed separately. Methodology was evaluated in three domains: selection of study population, comparability of study groups and the quality of outcome assessment. The kappa (κ) statistic was used to assess inter-observer agreement on study quality and a value above 0.80 was defined as good agreement [28].

Muller.indd 252 25-7-2014 14:40:36


253

Statistical analysisThe meta-analysis was performed with a random effects model [29]. A random effect model, instead of a fixed effect model, was chosen to take into account het-erogeneity of the studies. The common study characteristics are the incidence of TRALI cases for the total number of blood products transfused. Effect sizes are expressed as Odds ratio and their 95 % confidence intervals. The odds is the number of TRALI cases by the number of transfused products in a defined context. An Odds ratio lower than one expresses a protective effect of a low risk TRALI donor strat-egy for plasma products to prevent onset of TRALI. An Odds Ratio higher than one shows a protective effect against TRALI of transfused products in the period before the introduction of a male only donor policy. Heterogeneity was expressed by the CHI2 test. A cut-off value p 0.10 was used to consider heterogeneity. A high p value suggests that the heterogeneity is insignificant.Analyses were performed in R: A language and environment for statistical comput-ing. R Core Team (2013) [30].

Results

Study selectionOur pre-defined search strategy identified 118 records after removal of duplicates. After title review 53 citations were excluded and an additional 28 were excluded based on the abstract. Of 37 full text articles, 10 studies were considered eligible (table 2) and 27 studies were excluded after detailed review. An overview of the search is presented in figure 1. Inter-observer agreement for study selection was good (κ = 0.84).

Muller.indd 253 25-7-2014 14:40:36

Chapter 14

254

Table 2: O

verv

iew

of s

tudi

es.

Referenc

eType

of study

and

incl

usio

nPo

pulatio

nCo

untr

yStud

y ye

arEn

dpoi

ntRe

ported

type

of

cases*

Plasma or/and

Platelets

Percen

tage

male

dono

rs after im

ple-

men

tatio

n Pl

asm

aPlatelets

Wrig

ht[2

5]re

tros

pecti

veac

tive

surg

ery

UK

1998

-200

6on

set

TRAL

IM

ixed

ca

ses

Mal

e on

ly p

lasm

a an

d pl

atel

ets

>99%

>93%

Chap

-m

an[1

5]re

tros

pecti

vepa

ssiv

ena

tiona

lU

K19

99-2

006

onse

t TR

ALI

Non

m

ixed

ca

ses

Pred

omin

antly

mal

e on

ly

plas

ma

and

plat

elet

s80

-90%

80-9

0% fo

r BC

der

ived

PL

T po

ols

Vlaa

r[23

]re

tros

pecti

veac

tive

ICU

Net

her-

land

s20

04-2

007

onse

t TR

ALI

Mix

ed

case

sM

ale

only

pla

sma,

pla

te-

let s

trat

egy

not a

ppli-

cabl

e

100%

NA

Eder

[18]

retr

ospe

ctive

pass

ive

natio

nal

US

2008

-201

1on

set

TRAL

IM

ixed

ca

ses

Pred

omin

antly

mal

e on

ly

plas

ma,

pla

tele

t str

ateg

y no

t app

licab

le

>95%

>65%

of

aphe

resis

PL

T

Wie

r-su

m[2

4]re

tros

pecti

ve

pass

ive

natio

nal

Net

her-

land

s20

02-2

009

onse

t TR

ALI

Mix

ed

case

sM

ale

only

pla

sma,

pla

te-

let s

trat

egy

not a

ppli-

cabl

e

100%

NA

Nak

a-za

wa[

21]

pros

pecti

veac

tive

surg

ery

Japa

n20

08-2

008

onse

t TR

ALI

Mix

ed

case

sM

ale

only

pla

sma,

pla

te-

let s

trat

egy

not a

ppli-

cabl

e

100%

**N

A

Lin[

20]

retr

ospe

ctive

pass

ive

natio

nal

Cana

da20

01-2

009

onse

t TR

ALI

Mix

ed

case

sPr

edom

inan

tly m

ale

only

pl

asm

a an

d pl

atel

ets

86-1

00%

90-9

9%

Funk

[19]

retr

ospe

ctive

pass

ive

natio

nal

Ger

man

y20

06-2

010

onse

t TR

ALI

Mix

ed

case

sM

ale

only

pla

sma

and

plas

ma

from

fem

ale

dono

rs w

ithou

t his

tory

of

pre

gnan

cy o

r with

out

WBC

anti

bodi

es. P

late

let

stra

tegy

not

app

licab

le

NA

NA

Muller.indd 254 25-7-2014 14:40:36


255

Referenc

eType

of study

and

incl

usio

nPo

pulatio

nCo

untr

yStud

y ye

arEn

dpoi

ntRe

ported

type

of

cases*

Plasma or/and

Platelets

Percen

tage

male

dono

rs after im

ple-

men

tatio

n Pl

asm

aPlatelets

Arin

s-bu

rg[1

7]re

tros

pecti

vepa

ssiv

eho

spita

lsU

S20

06-2

008

onse

t TR

ALI

Non

m

ixed

ca

ses

Pred

omin

antly

mal

e on

ly

plas

ma,

pla

tele

t str

ateg

y no

t app

licab

le

95-1

00%

NA

Toy[

22]

pros

pecti

veac

tive

hosp

itals

US

2006

-200

9on

set

TRAL

IM

ixed

ca

ses

1 ce

nter

mal

e on

ly

plas

ma

and

pred

omi-

nant

ly m

ale

only

pla

te-

lets

and

1 c

ente

r mal

e or

ne

ver p

regn

ant f

emal

e do

nors

for p

lasm

a an

d pl

atel

ets

NA

NA

*for

ana

lysis

non

-mix

ed c

ases

wer

e co

nver

ted

to m

ixed

cas

es in

ord

er to

com

pare

TRA

LI c

ases

per

tota

l num

ber o

f pro

duct

s tra

nsfu

sed/

dist

ribut

ed.

** d

urin

g th

e st

udy

perio

d (O

ctob

er 2

007-

Janu

ary

2008

)N

on m

ixed

cas

es =

a si

ngle

pro

duct

has

bee

n id

entifi

ed c

ausin

g TR

ALI,

mix

ed c

ases

= a

ll pr

oduc

ts w

ithin

the

6 ho

urs p

rior o

nset

of T

RALI

hav

e be

en ta

ken

into

acc

ount

.N

A =

not a

pplic

able

BC =

buff

y co

atPL

T =

plat

elet

Muller.indd 255 25-7-2014 14:40:36

Chapter 14

256

Figure 1: Inclusion flow. Number of articles identified and process for inclusion in the meta-analysis.

Addi�onal recordsiden�fied through other sources

N=6

Records iden�fied from databases:Pubmed N=113Cochrane N=1

Records a�er duplicatesremoved

N=118

Records screenedN=118

Excluded a�er �tle reviewN=53

Abstracts screened for eligibilityN=65

Full-text ar�cles assessed foreligibility

N=37

Excluded a�er review ofabstracts

N=28

Excluded a�er review of full-text ar�cle

N=27Studies included in qualita�vesynthesis

N=10

Iden

�fica

�on

Scre

enin

gEl

igib

ility

Incl

uded

Muller.indd 256 25-7-2014 14:40:36


257

Study characteristicsNo randomized controlled trials were found on the effect of a low risk TRALI donor strategy for plasma products on the onset of TRALI. Therefore we only included observational studies. Of these, two studies were prospective [21,22]. The remain-ing 8 studies had a retrospective design. Five studies were carried out within national registries using passive reporting of TRALI [15,18-20,24]. Four studies were carried out in various high-risk TRALI populations (e.g. intensive care and surgery patients) and in these studies TRALI was actively screened for [21-23,25]. One study was car-ried out in the general hospital patients, but using passive reporting [17]. All authors used Canadian Consensus Criteria to diagnose TRALI [2,31] In four studies data were also collected before 2004, the year of establishment of the Canadian Consensus Criteria [15,20,24,25]. Of these studies, one used American-European consensus cri-teria for ALI within 6 hours of transfusion[25] and the second defined TRALI cases as “acute dyspnea with hypoxia and bilateral pulmonary infiltrates occurring during or in the 24 hours after transfusion, with no other apparent cause” [15]. The remain-ing two studies retrospectively applied the Canadian Consensus Criteria [20,24]. Majority of the included studies were able to implement a 95 to 100% male only plasma donation policy. Two studies used, in addition to male only plasma, plasma from female donors without history of pregnancy or without HLA/HNA antibodies [19,22]. Four studies reported implementation of a male only donor strategy for PLTs products, in addition to a male only donor strategy for plasma products, as this are high plasma volume products as well [15,18,20,25]. Detailed characteristics of included studies are shown in tables 2, 3a and 3b.

Muller.indd 257 25-7-2014 14:40:36

Chapter 14

258

Table 3a

: Tra

nsfu

sion

dat

a of

coh

ort s

tudi

es re

porti

ng tr

ansf

used

uni

ts.

Referenc

eTR

ALI cases

Total u

nits tran

sfused

Plasma tran

sfused

Platelets t

ransfused

Red bloo

d cells

tran

sfused

Before

After

Before

After

Before

After

Before

After

Before

After

Wrig

ht[2

5]37

1429

1717

5010

2662

913

191

1760

1030

Vlaa

r[23

]17

613

5048

517

413

588

2510

8832

5N

akaz

awa[

21]

32

1596

1480

467

424

545

556

584

500

Arin

sbur

g[17

]9

122

7913

2336

8547

756

5223

050

041

5014

313

0116

1313

12To

y[22

]23

1089

321

1237

31N

AN

AN

AN

AN

AN

A

NA=

not a

vaila

ble

Table 3b

: Tra

nsfu

sion

dat

a of

coh

ort s

tudi

es re

porti

ng d

istr

ibut

ed u

nits

.

Referenc

eTR

ALI cases

Total u

nits distributed

Plasma distrib

uted

Platelets d

istributed

Red bloo

d cells

distrib

uted

Before

After

Before

After

Before

After

Before

After

Before

After

Chap

man

[15]

5812

1655

0000

5897

000

1874

000

6340

0012

6500

051

8000

1341

1000

4745

000

Eder

[18]

5812

783

2043

733

8107

6216

6459

866

9503

763

7751

3172

097

6018

088

2394

3628

Wie

rsum

[24]

6831

4067

000

1377

000

5450

0018

3000

2790

0098

000

3243

000

1096

000

Lin[

20]

105

3176

3356

023

6440

016

4840

047

9050

6380

2021

8330

5347

140

1667

020

Funk

[19]

444

1217

0000

6200

000

2360

000

1080

000

8800

0049

0000

8930

000

4630

000

Muller.indd 258 25-7-2014 14:40:36


259

Figure 2: Meta-analysis for the onset of TRALI expressed over all products transfused before and after introduction of a low risk TRALI donor strategy for plasma containing products.

Effect sizes are expressed as Odds ratio and their 95 % confidence intervals. The Odds is the number of TRALI cases by the number of transfused products. An Odds ratio lower than one expresses a protective effect against TRALI of products transfused in the period after the introduction of a low risk TRALI donor strategy for plasma.

Methodological qualityThe methodological quality of cohort studies is summarized in table 4 and of case-control studies in table 5. Quality assessment revealed that risk of bias of patient selection was low. However, included studies were heterogeneous regarding TRALI surveillance, using passive or active reporting. In addition, in studies carried out within national registries possibly not all transfused patients were observed for the minimum of 6 hours, which may have resulted in underreporting of TRALI. How-

Muller.indd 259 25-7-2014 14:40:37

Chapter 14

260

Table 4. Quality of included cohort studies based on modified Newcastle-Ottowa Scale [27].

Reference Selection Comparability Outcome

Coho

rt

represen

tativ

e?

Non

-exp

osed

co

hort ade

quate?

Ascertainm

ent o

f ex

posure clear?

Outco

me no

t presen

t at s

tart of

stud

y?

Adjustmen

t for

confou

nding/bias

Assessmen

t of

outcom

e blinde

d

Leng

th of follow-up

approp

riate?

Adeq

uate of

follo

w-up of coh

ort?

Toy[22] Yes Yes Yes Yes No Yes Yes YesChapman[15] Yes Yes Yes Yes No No Yes YesEder[18] Yes Yes Yes Yes No No Yes YesWiersum[24] Yes Yes Yes Yes No Yes Yes NoLin[20] Yes Yes Yes Yes No Yes Yes YesFunk[19] Yes Yes Yes Yes No No Yes YesArinsburg[17] Yes Yes Yes Yes No Yes Yes Yes

Table 5. Quality of included case control studies based on modified Newcastle-Ottowa Scale [27]

Reference Selection Comparability Outcome

Coho

rt

represen

tativ

e?

Non

-exp

osed

co

hort ade

quate?

Ascertainm

ent o

f ex

posure clear?

Outco

me no

t presen

t at s

tart of

stud

y?

Adjustmen

t for

confou

nding/bias

Assessmen

t of

outcom

e blinde

d

Leng

th of follow-up

approp

riate?

Adeq

uate of

follo

w-up of coh

ort?

Wright[25] Yes Yes Yes Yes Yes Yes Yes YesVlaar[23] Yes Yes Yes Yes Yes Yes Yes YesNakazawa[21] Yes Yes Yes Yes Yes Yes Yes Yes

ever, this possible disadvantage was applicable for both periods, before and after implementation of TRALI risk reduction strategies. None of the nationwide studies corrected for confounding that may have been a possible source of bias. Blinded endpoint assessment was carried out in 7 studies, leaving the possibility of inter-pretation bias in the other 3 studies [15,18,19]. Inter-observer agreement for the quality assessment was good (median κ = 0.89 (range 0.40 to 1.0)). Heterogeneity was assessed using I2 test. Heterogeneity was high (p<0.01) for all studies combined and low for pre-defined subgroup analysis (p>0.10).

Muller.indd 260 25-7-2014 14:40:37


261

Meta-analysis on the risk of TRALI

All TRALI casesThe effect of a low risk TRALI donor strategy for plasma containing products on the occurrence of TRALI for all transfused products in all studies is summarized in figure 2. Figures 3 and 4 show the effect of a low risk TRALI donor strategy for plasma containing products for high-risk populations and from nationwide regis-tries respectively. Four studies showed an association between the introduction of a low risk TRALI donor strategy for plasma and a reduction of TRALI risk [17-19,22]. In five studies there was a trend to TRALI risk reduction after introduction of a low risk TRALI donor strategy for plasma [15,20,21,23,25] TRALI risk was unaffected in one study [24].Pooled data of all studies showed a significant reduction for the risk of TRALI after implementation of a low risk TRALI donor strategy for plasma containing products (OR 0.61 95%CI 0.42-0.86).Subgroup analysis revealed that data from local registries showed a significant reduction of TRALI risk (OR 0.51 95%CI 0.29-0.90). Patients in local registries con-sisted of critically ill patients [23] and patients undergoing major surgery [21,25]. As confirmed by one of the included studies [22] these patients are more prone to develop a TRALI reaction compared to a general patient population that also includes outpatients. Data from nationwide registries (general patient population) showed only a tendency towards protection of a low risk TRALI donor strategy for plasma containing products against TRALI (OR 0.66 95%CI 0.40-1.09) (figure 4).

Muller.indd 261 25-7-2014 14:40:37

Chapter 14

262

Figure 3: Meta-analysis for the onset of TRALI in at risk patient populations expressed over all products transfused before and after introduction of a low risk TRALI donor strategy for plasma containing products.

Effect sizes are expressed as Odds ratio and their 95 % confidence intervals. The odds is the number of TRALI cases by the number of transfused products. An Odds ratio lower than one expresses a protective effect against TRALI of products transfused in the period after the introduction of a low risk TRALI donor strategy for plasma.

Muller.indd 262 25-7-2014 14:40:37


263

Figure 4: Meta-analysis for the onset of TRALI in general patient populations expressed over all products transfused before and after introduction of a low risk TRALI donor strategy for plasma containing products.

Effect sizes are expressed as Odds ratio and their 95 % confidence intervals. The odds is the number of TRALI cases by the number of transfused products. An Odds ratio lower than one expresses a protective effect against TRALI of products transfused in the period after the introduction of a low risk TRALI donor strategy for plasma.

Muller.indd 263 25-7-2014 14:40:38

Chapter 14

264

Plasma-induced TRALIIncluded studies implemented different TRALI risk reduction strategies which also affected TRALI risk associated with other products than plasma. Therefore we performed an additional subgroup analysis to assess the effect of risk reduction strategies on plasma-induced TRALI. Nine of ten included studies provided data on plasma-induced TRALI [15,17-21,23-25]. The pooled effect of a low risk TRALI donor strategy on occurrence of TRALI per transfused unit of plasma is shown in figure 5. Figures 6 and 7 summarize the effect of low risk TRALI donor strategy on plasma-induced TRALI in for high-risk populations and from nationwide registries respectively. Data for plasma-induced TRALI showed a similar pattern as the data for all transfused products. Plasma-induced TRALI was significantly reduced after introduction of a low risk donor strategy (OR 0.62 95% CI 0.42-0.92). Patients from local registries (prone to develop TRALI) experienced a significant risk reduction (OR 0.51 95% CI 0.31-0.83), while data from nationwide registries (general patient popu-lation) showed a tendency to reduced risk of plasma-induced TRALI (OR 0.69 95% CI 0.42-1.13).

Meta-analysis on mortalityData on 30-day mortality of the total cohort were only available from two studies [23,25]. Both studies showed a non-significant reduction of mortality of the cohort of transfused patients after implementation of a low risk TRALI donor strategy for plasma products (43% before vs. 35% after and 43% before vs. 23%, N.S. respec-tively). As there were only two studies no meta-analysis was performed on these data points. As antibody mediated TRALI is suggested to be the most severe form of TRALI it can be hypothesized that reducing antibody mediated TRALI would also result in a reduction of mortality among TRALI patients. Data on 30-day mortality among TRALI patients were available from four studies [22-25]. None of the stud-ies showed a significant reduction of 30-day mortality among TRALI patients after the introduction of a low risk TRALI donor strategy for plasma containing products. Pooling of data from these four studies showed a trend to a reduction of mortal-ity among TRALI patients after introduction of a low risk TRALI donor strategy for plasma containing products (OR 0.69 95%CI 0.27-1.75). An overview of the effect of introduction of a low risk TRALI donor strategy for plasma containing products on mortality among TRALI patients is given in figure 8.

Muller.indd 264 25-7-2014 14:40:38


265

Figure 5: Meta-analysis for the onset of plasma-induced TRALI expressed over all units of plasma transfused before and after introduction of a low risk TRALI donor strategy for plasma.

Effect sizes are expressed as Odds ratio and their 95 % confidence intervals. The odds is the number of plasma-induced TRALI cases by the number of transfused units plasma. An Odds ratio lower than one expresses a protective effect against TRALI of plasma transfused in the period after the introduction of a low risk TRALI donor strategy for plasma.

Muller.indd 265 25-7-2014 14:40:39

Chapter 14

266

Figure 6: Meta-analysis for the onset of plasma-induced TRALI in at risk populations expressed over units of plasma transfused before and after introduction of a low risk TRALI donor strategy for plasma.

Effect sizes are expressed as Odds ratio and their 95 % confidence intervals. The odds is the number of plasma-induced TRALI cases by the number of transfused plasma. An Odds ratio lower than one expresses a protective effect against TRALI of plasma transfused in the period after the introduction of a low risk TRALI donor strategy for plasma.

Muller.indd 266 25-7-2014 14:40:39


267

Figure 7: Meta-analysis for the onset of plasma-TRALI in general patient populations expressed over units of plasma transfused before and after introduction of a low risk TRALI donor strategy for plasma.

Effect sizes are expressed as Odds ratio and their 95 % confidence intervals. The odds is the number of plasma-induced TRALI cases by the number of transfused plasma. An Odds ratio lower than one expresses a protective effect against TRALI of plasma transfused in the period after the introduction of a low risk TRALI donor strategy for plasma.

Muller.indd 267 25-7-2014 14:40:40

Chapter 14

268

Figure 8: Meta-analysis for mortality among patients developing TRALI before and after introduction of a low risk TRALI donor strategy for plasma containing products.

Effect sizes are expressed as Odds ratio and their 95 % confidence intervals.

Muller.indd 268 25-7-2014 14:40:40


269

Discussion

This study is the first meta-analysis on the impact of low risk TRALI donor strategies for plasma on the onset of TRALI. The main findings of this study are; 1) implementa-tion of a low risk TRALI donor strategy for plasma containing products, mainly male only donor policies, results in a significant reduction of onset of TRALI (OR 0.61, 95%CI 0.42-0.88). 2) introduction of a low risk TRALI donor strategy for plasma con-taining products results into a trend towards a reduction of 30-day mortality among TRALI-patients (OR 0.69, 95%CI 0.27-1.75). 3) at risk patient populations seem to benefit most of an introduction of a low risk TRALI donor strategy for plasma prod-ucts.Our study is able to confirm that the use of a low risk TRALI donor strategy such as a male only donor policy for plasma indeed results in a reduction of TRALI. The major-ity of TRALI cases, up to 89%, is thought to be antibody-mediated TRALI, caused by the passive infusion of donor leukocyte reactive antibodies, present in plasma containing blood products [10]. These antibodies are mainly induced by pregnancies and blood transfusions [14]. This was the basis for the hypothesis that the use of plasma from non-transfused male donors would minimize the antibody-mediated TRALI. Although this precautionary measure is effective it has serious consequences for the blood supply as the number of (potential) plasma donors is decreased by half, leading to a shortage of donors. In particular, a shortage in the availability of group AB plasma for transfusion is a serious concern [18,20]. From this point of view some countries introduced a preferentially male plasma donor strategy, as it was not feasible to use male-only plasma [15]. It should be noted that using a male only donor plasma policy does not totally prevent antibody mediated TRALI. First of all approximately 1% of the male donor population have antibodies [32]. Second, plate-let products are also high plasma volume products. The plasma added to the pooled plasma does not originate from male only plasma in all countries. Third, red blood cells (RBCs) are suspended in an additive solution. Although the final product con-tains only a mean volume of 10 - 20 mL of plasma, RBCs are also known to be impli-cated in TRALI [13,33]. Fourth, not all countries exclude male donors who have a previous history of blood transfusion themselves from donating blood [34]. Of note, an association between previous transfusions and the development of allo-reactive

Muller.indd 269 25-7-2014 14:40:40

Chapter 14

270

antibodies has not unequivocally been demonstrated [32]. Abovementioned rea-sons also partly explain residual TRALI cases in included studies [15,17-21]. In addi-tion, residual TRALI cases were reported to be caused by the use of group AB plasma from female donors [18,20].Theoretically, to completely minimize the risk of antibody-mediated TRALI, all donors should be screened for HLA or HNA antibodies and positive donors should be excluded for all blood products containing plasma, regardless the amount of plasma. Indeed, some blood collection organizations have implemented antibody screening for all donors or all female donors [35]. This results in less donor loss, but accounts for higher costs and is very labour intensive. In addition, the occurrence of HNA antibodies may not be linked to pregnancy or gender. Therefore, further research is warranted before donor antibody-screening can be advocated.There are alternatives that might be equally effective. Other countries (e.g. Norway, Finland, France) use pooled solvent/detergent plasma (SD plasma) instead of single donor plasma. Pooling has the advantage that it reduces the possible antibody load by dilution and by neutralizing antigens in the plasma pool, which minimizes the risk of TRALI [36]. The countries that use pooled SD plasma claim that they do not see TRALI reactions associated with this product. Another approach could be to ask the donors for a history of pregnancies and blood transfusions and subsequently test these donors for HLA and HNA antibodies. This method is a reliable predictor of HLA allo-immunisation [37]. However, it has been suggested that this method may neglect potential HLA and/or HNA positive donors as many pregnancies end in the first weeks after conception, while women weren’t even aware of a pregnancy. Of note, no association was found between early miscarriages and occurrence of allo-immunisation of donors [32]. Also, donors are not always aware of receiving blood transfusions as they are sometimes given during an operation without notification afterwards.The introduction of a low risk TRALI donor strategy for plasma containing products seemed to reduce 30-day mortality among patients developing TRALI. However, the observed reduction in mortality was only a trend that can possibly be explained by small number of studies and patients included. An explanation for mortality reduction among TRALI patients is tempting. Antibody mediated TRALI is reported to induce a more severe form of TRALI compared to non-antibody mediated TRALI

Muller.indd 270 25-7-2014 14:40:40


271

[38,39]. We hypothesize that as a result of a low risk TRALI donor strategy for plasma containing products, the frequency of antibody induced TRALI is reduced hereby contributing to a reduction in mortality among TRALI patients.Data from our meta-analysis showed that the implementation of a low risk TRALI donor strategy for plasma had the highest impact among at risk patient populations such as critically ill and surgery patients. This may be well explained by the threshold model of TRALI [40]. In this model a threshold must be overcome to induce a TRALI reaction. Factors that determine the threshold are the predisposition of the patient that determines priming of the lung neutrophils and the ability of the mediators in the transfusion to cause activation of primed neutrophils. A strong antibody-medi-ated response can cause severe TRALI in an otherwise “healthy” recipient. When activation status is too low, it is possible that priming factors in the transfusion are not strong enough to overcome the threshold. This concept has been proven in ani-mal models [41-43]. This model explains why a low risk TRALI donor strategy for plasma has limited effect in the general patient population as they have no severe underlying condition to overcome the threshold. In this population a high volume of antibodies or a strong antibody antigen match is needed to overcome the threshold for onset of TRALI [22].Our study has several limitations. First of all there are no randomized controlled trials on the effect of a low risk TRALI donor strategy for plasma on the onset of TRALI. It is also not expected that after the implementation of a low risk TRALI donor strategy for plasma in many countries a randomized controlled trial will follow. The observational nature of the studies and reliance on spontaneous reporting to the blood collection organizations or the nationwide registries (which may be far from complete) may have introduced bias. However, this bias would apply equally to the before and after period.Second, implemented donor policies were not fully equal among all studies. Not all studies were able to achieve a 100% male only donor policy for plasma due to scar-city of products with a certain blood group. Some studies reported the implementa-tion of a male only PLT policy in addition to a male only donor plasma policy. Others also report the introduction, next to male only plasma, of plasma from female donors without a history of pregnancy or without HLA/HNA antibodies. However, all

Muller.indd 271 25-7-2014 14:40:40

Chapter 14

272

implemented policies assume the same underlying mechanism of excluding (poten-tial) HLA or HNA positive plasma donors to prevent antibody mediated TRALI.Third, some of the studies included TRALI patients before the consensus criteria for TRALI from the Canadian Consensus Conference in 2004 were established [2,31]. Before that year there were no universal criteria for TRALI and TRALI was most likely underreported. However, this will rather give an underestimation of the effect of the low risk TRALI donor strategy for plasma, as the period before 2004 is included in the before group. Fourth, heterogeneity was high for all studies combined and low for subgroup analysis. This can be explained by differences in population and observation period of included studies. The random effects model used in our study provides an average effect and in this way controls for variation in studies included. From this perspective we believe that the higher heterogeneity in the total group analysis has no or limited impact on our results and conclusions.

Conclusion

This is the first meta-analysis on the effect of implementation of a low risk TRALI donor strategy mainly based on a male only donor policy for plasma containing prod-ucts to prevent TRALI. This study shows that introduction of a low risk TRALI donor strategy for plasma containing products reduces the onset of TRALI, especially in high risk patient populations. Furthermore there is a tendency that reduction of antibody mediated TRALI results in a lower mortality rate among TRALI patients.

Muller.indd 272 25-7-2014 14:40:40


273

References

1. http://www.fda.gov/BiologicsBloodVaccines/SafetyAvailability/ReportaProblem/TransfusionDonationFatalities/ucm346639.htm. Accessed 7 September 2013.

2. Kleinman S, Caulfield T, Chan P, et al.: Toward an understanding of transfusion-related acute lung injury: statement of a consensus panel Transfusion 2004;44: 1774-1789

3. Vlaar AP, Juffermans NP: Transfusion-related acute lung injury: a clinical review Lancet 2013;382: 984-994

4. Silliman CC: The two-event model of transfusion-related acute lung injury Crit Care Med 2006;34: S124-S131

5. Mangalmurti NS, Xiong Z, Hulver M, et al.: Loss of red cell chemokine scavenging pro-motes transfusion-related lung inflammation Blood 2009;113: 1158-1166

6. Vlaar AP, Hofstra JJ, Levi M, et al.: Supernatant of Aged Erythrocytes Causes Lung Inflammation and Coagulopathy in a “Two-Hit” In Vivo Syngeneic Transfusion Model Anesthesiology 2010;113: 92-103

7. Vlaar AP, Hofstra JJ, Kulik W, et al.: Supernatant of stored platelets causes lung inflam-mation and coagulopathy in a novel in vivo transfusion model Blood 2010;116: 1360-1368

8. Popovsky MA, Abel MD, Moore SB: Transfusion-related acute lung injury associated with passive transfer of antileukocyte antibodies Am Rev Respir Dis 1983;128: 185-189

9. Curtis BR, McFarland JG: Mechanisms of transfusion-related acute lung injury (TRALI): anti-leukocyte antibodies Crit Care Med 2006;34: S118-S123

10. Popovsky MA, Moore SB: Diagnostic and pathogenetic considerations in transfusion-related acute lung injury Transfusion 1985;25: 573-577

11. Holness L, Knippen MA, Simmons L, et al.: Fatalities caused by TRALI Transfus Med Rev 2004;18: 184-188

12. Popovsky MA, Davenport RD: Transfusion-related acute lung injury: femme fatale? Transfusion 2001;41: 312-315

13. van Stein D, Beckers EA, Sintnicolaas K, et al.: Transfusion-related acute lung injury reports in the Netherlands: an observational study Transfusion 2010;50: 213-220

14. Densmore TL, Goodnough LT, Ali S, et al.: Prevalence of HLA sensitization in female apheresis donors Transfusion 1999;39: 103-106

15. Chapman CE, Stainsby D, Jones H, et al.: Ten years of hemovigilance reports of transfu-sion-related acute lung injury in the United Kingdom and the impact of preferential use of male donor plasma Transfusion 2009;49: 440-452

16. Muller MC, Porcelijn L, Vlaar AP: Prevention of Immune-mediated Transfusion-related Acute Lung Injury; from Bloodbank to Patient Curr Pharm Des 2012;18: 3241-3248

17. Arinsburg SA, Skerrett DL, Karp JK, et al.: Conversion to low transfusion-related acute lung injury (TRALI)-risk plasma significantly reduces TRALI Transfusion 2012;52: 946-952

18. Eder AF, Dy BA, Perez JM, et al.: The residual risk of transfusion-related acute lung injury at the American Red Cross (2008-2011): limitations of a predominantly male-donor plasma mitigation strategy Transfusion 2013;53: 1442-1449

Muller.indd 273 25-7-2014 14:40:40

Chapter 14

274

19. Funk MB, Guenay S, Lohmann A, et al.: Benefit of transfusion-related acute lung injury risk-minimization measures - German haemovigilance data (2006-2010) Vox Sang 2012;102: 317-323

20. Lin Y, Saw CL, Hannach B, et al.: Transfusion-related acute lung injury prevention meas-ures and their impact at Canadian Blood Services Transfusion 2012;52: 567-574

21. Nakazawa H, Ohnishi H, Okazaki H, et al.: Impact of fresh-frozen plasma from male-only donors versus mixed-sex donors on postoperative respiratory function in surgical patients: a prospective case-controlled study Transfusion 2009;49: 2434-2441

22. Toy P, Gajic O, Bacchetti P, et al.: Transfusion related acute lung injury: incidence and risk factors Blood 2012;119:1757-1767

23. Vlaar AP, Binnekade JM, Prins D, et al.: A nested case-control study Crit Care Med 2010;38: 771-778

24. Wiersum-Osselton JC, Middelburg RA, Beckers EA, et al.: Male-only fresh-frozen plasma for transfusion-related acute lung injury prevention: before-and-after compara-tive cohort study Transfusion 2011;51: 1278-1283

25. Wright SE, Snowden CP, Athey SC, et al.: the effect of excluding donations from females from the production of fresh frozen plasma Crit Care Med 2008;36: 1796-1802

26. Moher D, Liberati A, Tetzlaff J, et al.: Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement Ann Intern Med 2009;151: 264-9, W64

27. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed 19 October 2013

28. Landis JR, Koch GG: The measurement of observer agreement for categorical data Biometrics 1977;33: 159-174

29. DerSimonian R, Kacker R: Random-effects model for meta-analysis of clinical trials: an update Contemp Clin Trials 2007;28: 105-114

30. Lumley, T. Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/. The R package rMeta, version 2.16.: in: 2013.

31. Goldman M, Webert KE, Arnold DM, et al.: Proceedings of a consensus conference: towards an understanding of TRALI Transfus Med Rev 2005;19: 2-31

32. Triulzi DJ, Kleinman S, Kakaiya RM, et al.: The effect of previous pregnancy and transfu-sion on HLA alloimmunization in blood donors: implications for a transfusion-related acute lung injury risk reduction strategy Transfusion 2009;49: 1825-1835

33. Win N, Chapman CE, Bowles KM, et al.: How much residual plasma may cause TRALI? Transfus Med 2008;18: 276-280

34. http://www.redcrossblood.org/donating-blood/eligibility-requirements/eligibility-cri-teria-topic#med_treatments. Accessed 20 October 2013.

35. Reesink HW, Lee J, Keller A, et al.: Measures to prevent transfusion-related acute lung injury (TRALI) Vox Sang 2012;103: 231-259

36. Sachs UJ, Kauschat D, Bein G: White blood cell-reactive antibodies are undetectable in solvent/detergent plasma Transfusion 2005;45: 1628-1631

Muller.indd 274 25-7-2014 14:40:40


275

37. Powers A, Stowell CP, Dzik WH, et al.: Testing only donors with a prior history of preg-nancy or transfusion is a logical and cost-effective transfusion-related acute lung injury prevention strategy Transfusion 2008;48: 2549-2558

38. Marik PE, Corwin HL: Acute lung injury following blood transfusion: expanding the defi-nition Crit Care Med 2008;36: 3080-3084

39. Moore SB: Transfusion-related acute lung injury (TRALI): clinical presentation, treat-ment, and prognosis Crit Care Med 2006;34: S114-S117

40. Bux J, Sachs UJ: The pathogenesis of transfusion-related acute lung injury (TRALI) Br J Haematol 2007;136: 788-799

41. Looney MR, Nguyen JX, Hu Y, et al.: Platelet depletion and aspirin treatment pro-tect mice in a two-event model of transfusion-related acute lung injury J Clin Invest 2009;119: 3450-3461

42. Vlaar A.P., Wolthuis EK, Hofstra JJ, et al.: Mechanical ventilation aggravates transfu-sion-related acute lung injury induced by MHC-I class antibodies. Intensive Care Med 2010;36: 879-87

43. Vlaar AP, Kuipers MT, Hofstra JJ,et al.: Mechanical ventilation and the titer of antibodies as risk factors for the development of transfusion-related lung injury Crit Care Res Pract 2012;2012: 720950

44. Muller MC, Juffermans NP: Transfusion-related acute lung injury: a preventable syn-drome? Expert Rev Hematol 2012;5: 97-106

Muller.indd 275 25-7-2014 14:40:40

Muller.indd 276 25-7-2014 14:40:40

Chapter 15

Methylprednisolone fails to attenuate lung injury in a

mouse model of transfusion-related acute lung injury

Marcella C.A. Müller, Pieter R. Tuinman, Koenraad F. van der Sluijs, Louis Boon, Joris J. Roelofs, Margreeth B. Vroom, Nicole P. Juffermans

Transfusion 2014;54:996-1001

Muller.indd 277 25-7-2014 14:40:40

Chapter 15

278

Abstract

Background: Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related morbidity and mortality. Anecdotally, TRALI patients have been treated with corticosteroids. However, evidence for its therapeutic rationale in TRALI is lacking. We determined the effects of corticosteroids on lung injury in a ‘two hit’ mouse model of antibody mediated TRALI.Methods: BALB/c mice were primed with lipopolysaccharide, after which TRALI was induced by injecting MHC-I antibody against H2Kd. Mice infused with phosphate-buffered saline served as controls. Simultaneously, one group of TRALI mice was infused with methylprednisolone (MPS; 2 mg/kg). Mice were supported by mechan-ical ventilation for 2 hours, after which bronchoalveolar lavage fluid (BALF) and lung homogenate were obtained. Statistics were obtained by one-way analysis of vari-ance or Kruskall Wallis.Results: Injection of MHC-I antibodies resulted in TRALI, indicated by pulmonary edema and increased BALF levels of protein and pro-inflammatory mediators mac-rophage inflammatory protein-2, keratinocyte-derived chemokine, and interleukin (IL)-6. Administration of MPS did not affect the amount of edema nor pulmonary protein and chemokine levels. MPS reduced systemic inflammatory reaction as well as IL-6 levels in the BALF.Conclusion: In a ‘two-hit’ model of antibody-mediated TRALI, MPS attenuated the IL-6 host response, but failed to prevent the development of lung injury.

Muller.indd 278 25-7-2014 14:40:40

Methylprednisolone in a murine model of TRALI

279

Introduction

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related mortality and morbidity [1]. In critically ill and surgical patients, relatively high incidences of TRALI have been reported [2-4], associated with prolonged mechanical ventilation, increased hospital mortality and decreased long-term sur-vival [3-5]. Indeed, the clinical picture of TRALI can be devastating, with fulminant edema and ensuing hypoxia [5].Therapy is supportive including mechanical ventilation and hemodynamic support. Case reports show that patients are often treated with high-dose corticosteroids [5-7].The rationale for treatment with steroids may be that acute respiratory distress syndrome (ARDS) and TRALI show clinical and histopathological similarities, with an inflammatory response leading to the disruption of the lung alveolar-capillary per-meability barrier [8]. In severe ARDS, corticosteroids were shown to reduce mortal-ity [9,10]. However, there are no data that support the use of steroids in TRALI.In the present study, we investigated the effect of methylprednisolone (MPS) on pulmonary inflammation in a murine “two-hit” model of antibody-mediated TRALI. This model corresponds well with human TRALI [11], as an inflammatory condition of the patient at the time of the transfusion increases the susceptibility for a TRALI reaction.


Experiments were performed with male BALB/c mice (Charles River, Someren, the Netherlands), aged 10-12 weeks and weighing 22-27 g, randomly assigned to 3 groups (n=16 per group). The study was approved by the Animal Care and Use Committee of the Academic Medical Center at the University of Amsterdam, the Netherlands (protocols 102190-1 and 102190-2). Animal procedures were carried out in compliance with Institutional Standards for Use of Laboratory Animals.

Muller.indd 279 25-7-2014 14:40:40

Chapter 15

280

Experimental study protocolWe performed a stepwise dose finding experiment in order to assess optimal dose of major histocompatibility complex (MHC)-I antibodies capable of inducing TRALI in a ‘two hit’ TRALI model in mice primed with lipopolysaccharide (LPS, from Escheri-chia Coli 0111:B4) which was administered intraperitoneally at a dose of 0.1 mg/kg. Twenty-four hours after LPS, mice were anesthetized by intraperitoneal injec-tion with a mix containing ketamine (EurovetAnimal Health BV, Bladel, the Nether-lands), medetomidine (Pfizer Animal Health BV, Capelle a/d Ijssel, the Netherlands) and atropine (Pharmachemie, Haarlem, the Netherlands), as described previously [12]. A tracheostomy was inserted and animals were mechanically ventilated in a pressure-controlled mode, with an inspiratory pressure of 12 cm H2O, PEEP of 2 cm H2O (resulting in VT of approx. 7.5 mL/kg), respiratory rate 100 breaths/min and FiO2 of 0.5 (Servo 900 C, Siemens, Sweden). After exposure of the jugular vein, using a 30-gauge sterile needle attached to polyethylene tubing, venous blood was aspi-rated from the jugular vein to verify intravascular placement of the needle. TRALI was induced by injecting MHC-I antibody against H2Kd (IgG2a,k). In previous experi-ments, we found that matched isotype antibody (IgG2a, CRL-1908, American Type Culture Collection) did not differ from vehicle (phosphate buffered saline [PBS]) [13]. Therefore, control mice received PBS. A dose of 2 mg/kg was found to induce lung injury without increased mortality (data not shown), which is in line with previous reports on this model [11].Immediately prior to infusion of antibodies, MPS (Solumedrol®, Pfizer, Capelle a/d Ijssel, The Netherlands) in a dose of 2 mg/kg was infused [14,15]. Previous experi-ments showed no differences between a volume matched injection of NaCl 0.9% instead of treatment, and no treatment of mice experiencing TRALI. Therefore, in order to limit the number of animals, we have omitted the saline group. After 2 hours of mechanical ventilation, mice were bled from the carotid artery. Blood gas analysis was done in a blood gas analyzer (Rapidlab 865, Bayer, Mijdrecht, the Netherlands). Bronchoalveolar lavage fluid (BALF) was obtained from the right lung and cell counts were determined using a hemacytometer (Beckman Coulter, Full-erton, CA, USA). Differential counts were done on cytospin preparations stained with a modified Giemsa stain, Diff-Quick (Dade Behring AG, Düdingen, Switzerland). Supernatant was stored at -80°C for total protein level and cytokine measurement.

Muller.indd 280 25-7-2014 14:40:40


281

The left lung was weighed and dried for three days in an oven at 65°C. The ratio of wet weight to dry weight represents tissue edema.In a second set of experiments, right lungs were used for preparation of homogenate after dilution in saline and Greenberger lysis buffer using a tissue homogenizer (Bio-spec products, Bartlesville, OK, USA). Supernatant was stored at -80°C for cytokine measurements. Left lungs were fixed in 4% formalin and embedded in paraffin for histopathology. Four-micrometer sections were stained with hematoxylin-eosin and analyzed by a pathologist who was blinded for group identity. To score lung injury, we used a modified VILI histology scoring system as previously described [16]. In short, four pathologic parameters were scored on a scale of 0–4: 1) alveolar conges-tion, 2) haemorrhage, 3) leukocyte infiltration, and 4) thickness of alveolar wall/hya-line membranes. A score of 0 represents normal lungs; 1, mild, less than 25% lung involvement; 2, moderate, 25% to 50% lung involvement; 3, severe, 50% to 75% lung involvement; 4, very severe, more than 75% lung involvement. The total histology score was expressed as the sum of the score for all parameters.

AssaysTotal protein levels (Bradford Protein Assay Kit, OZ Bioscience, Marseille, France) were measured in BALF. Keratinocyte-derived chemokine (KC), macrophage inflam-matory protein-2 (MIP-2) and interleukin-6 (IL-6) were measured in BALF, homogen-ate and plasma using enzyme-linked immunosorbent assay (ELISA), according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN, USA).

StatisticsData are expressed as mean ± standard deviation (SD) or median [interquartile range] when appropriate. Comparisons between groups were performed using one-way analysis of variance followed by t-test or Kruskall-Wallis followed by Mann Whitney U-test depending on data distribution. A p value of less than 0.05 was con-sidered statistically significant. Statistical analyses were performed with Prism ver-sion 5.0 (GraphPad Software, San Diego, CA, USA).

Muller.indd 281 25-7-2014 14:40:40

Chapter 15

282

Results

Data preliminary studiesPreliminary experiments did not demonstrate differences between mice infused with isotype antibody or vehicle (PBS). Pulmonary edema expressed as wet/dry ratio was low in both groups (4.2 (±0.3) versus 4.5 (±0.3), p=0.07). Furthermore, total protein levels in BALF did not differ (106 ug/ml [84-162] vs. 81 ug/ml [61-151], p=0.16), nor did KC levels in BALF (54 pg/ml (±16) vs. 56 pg/ml (±12), p=0.76). Lung injury scores were low in mice infused with isotype and those infused with PBS (1.3 (±1.1) vs. (1.9 (±0.8), p=0.21). To determine whether volume of MPS infusion had an effect on outcome, we compared infusion of MPS with a volume matched injection of 0.9% NaCl. No differences were found in wet/dry ratio (4.7 (±0.3) in 0.9% NaCl vs. 4.8 (±0.2) in MPS, p=0.25), total protein levels in BALF (117 ug/ml (±65) in 0.9% NaCl vs. 177 ug/ml (±71) in MPS, p=0.11) and pulmonary levels of KC (65 pg/ml [58-90] in 0.9% NaCl vs. 111 pg/ml [70-113] in MPS, p=0.07).

Induction of TRALI with MHC-I antibodyAfter the induction of TRALI, six of the 32 animals deceased prematurely, of whom four in the control group and two in the group treated with MPS. Mortality was related to instrumentation and did not differ significantly between groups. These animals were excluded from analysis. Injection of MHC-I antibodies resulted in TRALI, indicated by edema as reflected by an increase in wet-to-dry ratio of the lungs compared to controls, with increased BALF protein levels due to leakage of protein in the pulmonary compartment (figure 1). Lung injury was further charac-terized by an increase in both lung and BALF levels of chemokines MIP-2 and KC and the pro-inflammatory cytokine IL-6 (figure 2). TRALI did not result in increased pul-monary neutrophil influx within 2 hours (0.0 [0.0-1.0] in controls vs. 4.5 [0.0-9.3] in TRALI, p=0.08). Lung injury scores were significantly higher in TRALI mice compared to controls (4.6 (±0.7) vs. (1.9 (±0.8), p<0.001). TRALI also induced a systemic inflam-matory reaction illustrated by increased plasma levels of both KC and IL-6 (figure 3).

Muller.indd 282 25-7-2014 14:40:40


283

Figure 1: Wet-to-dry ratio of the lungs and BALF protein levels in mice injected with antibod-ies that induce TRALI, treated with MPS.

Data expressed as mean, whiskers indicate minimum and maximum.TRALI = transfusion related acute lung injuryPBS = phosphate buffered salineMPS = methylprednisoloneBALF = broncho-alveolar fluid.

The effect of MPSAdministration of MPS did not decrease wet-to-dry ratio or BALF protein levels (fig-ure 1). BALF and lung homogenate levels of chemokines MIP-2 and KC were also not affected by administration of steroids, while IL-6 homogenate and BALF levels were reduced (p=0.002 and p=0.03) (figure 2). Lung injury scores did not differ between treated (3.6 (±1.9) and untreated mice (4.9 (±0.7), p=0.12). Although steroids only reduced IL-6 in the lung, steroids reduced plasma levels of both KC and IL-6 (fig-ure 3), validating the anti-inflammatory capacity of the selected dose of 2 mg/kg MPS in this model.

Muller.indd 283 25-7-2014 14:40:40

Chapter 15

284

Figure 2: BALF and lung homogenate levels of MIP-2, KC and IL-6 in mice injected with anti-bodies that induce TRALI, treated with MPS.Data expressed as median, whiskers indicate minimum and maximum.

TRALI = transfusion related acute lung injuryPBS = phosphate buffered salineMPS = methylprednisoloneBALF = broncho-alveolar fluid.

Muller.indd 284 25-7-2014 14:40:41


285

Figure 3: Plasma levels of KC and IL-6 in mice injected with antibodies that induce TRALI, treated with MPS.

Data expressed as median, whiskers indicate minimum and maximum.TRALI = transfusion related acute lung injuryPBS = phosphate buffered salineMPS = methylprednisolone.

Discussion

Steroids did not improve markers of lung injury in a two-hit TRALI model, as pul-monary edema persisted, due to leakage of protein into the alveolar compartment. MPS did attenuate the IL-6 host response in the lung and both IL-6 and KC systemi-cally.In patients with early acute ARDS, MPS reduces lung injury, duration of mechanical ventilation and mortality, which, at least in part, was attributed to reduced plasma levels of IL-6 [17]. The latter is in line with our findings in a TRALI model, which show a clear reduction of pulmonary as well as systemic IL-6 levels after MPS treatment. However, in our study, despite the reduction of the IL-6 host response, inflamma-tory damage persisted. A possible explanation could be related to timing. TRALI is a hyperacute syndrome, developing within hours after transfusion, whereas ARDS takes a more prolonged course. In line with the clinical syndrome, we used a hypera-cute model in which mice were euthanized after 2 hours, when inflammation has shown to be maximal [18]. In ARDS, beneficial effects of steroids were only shown after infusion for 14 days [17].

Muller.indd 285 25-7-2014 14:40:41

Chapter 15

286

Another possible explanation of the lack of effect of MPS on lung injury might be the fact that inflammatory damage in TRALI is mediated via pathways which are not affected by the administration of steroids. In this model of antibody mediated TRALI, complement activation and macrophages were shown to be crucial contribu-tors to the development of TRALI [19]. Of note, MPS can increase C1q production by macrophages [20], hereby enhancing complement activation with subsequent inflammatory injury.The dose of MPS used in this study equals the dose which was shown to be ben-eficial in late ARDS [10] and higher then the dose that was beneficial in early ARDS [9]. Also, the chosen dose has shown to attenuate pulmonary inflammation in mice experiencing LPS induced acute lung injury [14,15]. As MPS attenuated the systemic inflammatory response, we think that under dosing of MPS is an unlikely explanation for the absence of beneficial effects on lung injury. However, the absence of phar-macokinetic data is a limitation to our study.Our study has also other limitations. We used an acute model and cannot exclude that prolonged administration of steroids may be beneficial. Also, our model of antibody-mediated TRALI does not take into account other factors that are also capable of inducing TRALI [21]. Possibly, the response to MPS may be different in non-immune mediated TRALI. Finally, this supported TRALI model does not allow the study whether steroids affect mortality, which would be the most relevant out-come measure.Our findings do not demonstrate a beneficial effect of corticosteroids in murine TRALI. A change in practice in human TRALI, based upon these findings, would be premature. Further research on timing and duration of corticosteroids in different models of TRALI is warranted.

Muller.indd 286 25-7-2014 14:40:41


287

References





5. Wallis JP, Lubenko A, Wells AW, et al.: Single hospital experience of TRALI. Transfusion 2003;43: 1053-1059

6. Jeddi R, Mansouri R, Kacem K, et al.:Transfusion-related acute lung injury (TRALI) during remission induction course of acute myeloid leukemia: a possible role for all-transreti-noic-acid (ATRA)? Pathol Biol 2009;57: 500-502

7. Djalali AG, Moore KA, Kelly E: Report of a patient with severe transfusion-related acute lung injury after multiple transfusions, resuscitated with albumin. Resuscitation 2005;66: 225-230


9. Meduri GU, Golden E, Freire AX, et al.: Methylprednisolone infusion in early severe ARDS: results of a randomized controlled trial. Chest 2007;131: 954-963

10. Meduri GU, Headley AS, Golden E, et al.: Effect of prolonged methylprednisolone ther-apy in unresolving acute respiratory distress syndrome: a randomized controlled trial. JAMA 1998;280: 159-165

11. Looney MR, Nguyen JX, Hu Y, et al.: Platelet depletion and aspirin treatment pro-tect mice in a two-event model of transfusion-related acute lung injury. J Clin Invest 2009;119: 3450-3461

12. Tuinman PR, Gerards MC, Jongsma G, et al.: Lack of evidence of CD40 ligand involve-ment in transfusion-related acute lung injury. Clin Exp Immunol 2011;165: 278-284


14. Silva PL, Garcia CS, Maronas PA, et al.: Early short-term versus prolonged low-dose methylprednisolone therapy in acute lung injury. Eur Respir J 2009;33: 634-645

15. Leite-Junior JH, Garcia CS, Souza-Fernandes AB, et al.: Methylprednisolone improves lung mechanics and reduces the inflammatory response in pulmonary but not in extrapulmonary mild acute lung injury in mice. Crit Care Med 2008;36: 2621-2628

16. Belperio JA, Keane MP, Burdick MD, et al.: Critical role for CXCR2 and CXCR2 ligands dur-ing the pathogenesis of ventilator-induced lung injury. J Clin Invest 2002;110: 1703-1716

Muller.indd 287 25-7-2014 14:40:41

Chapter 15

288

17. Seam N, Meduri GU, Wang H, et al.: Effects of methylprednisolone infusion on markers of inflammation, coagulation, and angiogenesis in early acute respiratory distress syn-drome. Crit Care Med 2012;40: 495-501



20. Faust D, Akoglu B, Zgouras D, et al.: Anti-inflammatory drugs modulate C1q secretion in human peritoneal macrophages in vitro. Biochem Pharmacol 2002;64: 457-462


Muller.indd 288 25-7-2014 14:40:41

Chapter 16

The effect of C1-inhibitor in a murine model

of transfusion-related acute lung injury

Marcella C.A. Müller, Ingrid Stroo, Diana Wouters, Sacha S. Zeerleder, Joris J.T.H. Roelofs, Louis Boon, Margreeth B. Vroom, Nicole P. Juffermans

Vox Sanguinis 2014;107:71-75

Muller.indd 289 25-7-2014 14:40:41

Chapter 16

290

Abstract

Background and objective: Transfusion-related acute lung injury (TRALI) is the lead-ing cause of transfusion-related morbidity and mortality. Specific therapy is lacking. We assessed whether C1-inhibitor attenuates lung injury in a ‘two-hit’ TRALI model.Methods: Mice were primed with lipopolysaccharide, subsequently TRALI was induced by MHC-I antibodies. In the intervention group, C1-inhibitor was infused concomitantly. Mice were supported with mechanical ventilation. After 2 hours, mice were killed, lungs were removed and bronchoalveolar lavage fluid (BALF) was obtained.Results: Injection of MHC-I antibodies induced TRALI, illustrated by an increase in wet-to-dry ratio of the lungs, in BALF protein levels and in lung injury scores. TRALI was further characterized by complement activation, demonstrated by increased BALF levels of C3a and C5a. Administration of C1-inhibitor resulted in increased pul-monary C1-inhibitor levels with high activity. C1-inhibitor reduced pulmonary levels of complement C3a associated with improved lung injury scores. However, levels of pro-inflammatory mediators were unaffected.Conclusion: In a murine model of TRALI, C1-inhibitor attenuated pulmonary levels of C3a associated with improved lung injury scores, but with persistent high levels of inflammatory cytokines.

Muller.indd 290 25-7-2014 14:40:41

C1-inhibitor in a murine model of TRALI

291

Introduction

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related mortality and morbidity [1]. TRALI can result in fulminant oedema and ensu-ing hypoxia, associated with prolonged mechanical ventilation, increased hospital mortality and decreased survival in critically ill patient populations [2]. Therapy is only supportive, including mechanical ventilation and hemodynamic support, while curative therapy is lacking.While neutrophils and antileucocyte antibodies are considered key players, the exact pathophysiology of TRALI is still under debate. Animal studies indicate that complement activation is required for development of TRALI [3,4]. Also, in clinical TRALI cases, complement activation contributes to the inflammatory cascade [5,6]. Antibodies in transfused components may result in activation of the complement system [5,6]. Possibly, the formation of immune complexes activates the comple-ment system via the classical pathway; however, the exact route of complement activation in TRALI remains to be established. Activation of the classical pathway of complement system induces a downstream response with formation of ana-phylatoxins C3a and C5a. C1-inhibitor blocks activation of the classical as well as the mannan-binding lectine (MBL) pathway of the complement system [7]. Administra-tion of extra C1-inhibitor was beneficial in other neutrophil-mediated inflammatory diseases including sepsis and myocardial infarction [7]. We hypothesized that lung injury in TRALI may be attenuated by blocking complement activation of the classical pathway. Therefore, we investigated the effect of administration of C1-inhibitor in murine “two-hit” model of antibody-mediated TRALI.


Experiments were performed with male BALB/c mice (n=48) (Charles River, Some-ren, the Netherlands), aged 10-12 weeks and weighing 22-27 g. Mice were randomly assigned to 3 groups (n=16 per group). The study was approved by the Animal Care and Use Committee of the Academic Medical Center at the University of Amster-dam, the Netherlands. Animal procedures were carried out in compliance with Insti-tutional Standards for Use of Laboratory Animals.

Muller.indd 291 25-7-2014 14:40:41

Chapter 16

292

Experimental study protocolMice were primed with lipopolysaccharide (LPS, from E. Coli 0111:B4, 0.1 mg/kg) intraperitoneally (i.p.). Twenty-four hour later, mice were anesthetized, a trache-ostomy was inserted and all mice were mechanically ventilated in a pressure-con-trolled mode, with an inspiratory pressure of 12 cm H2O, PEEP of 2 cm H2O (resulting in VT ~ 7.5 mL/kg), respiratory rate 100 breaths/min and FiO2 of 0.5 (Servo 900 C, Siemens, Sweden). Mice were mechanically ventilated because this TRALI model has a high mortality when unsupported. TRALI was induced by injecting MHC-I antibody (2 mg/kg) against H2Kd (IgG2a,k) into the jugular vein. Controls received phosphate-buffered saline (PBS). Immediately prior to infusion of antibodies, C1-inhibitor (400 U/kg) was infused (Cetor®, Sanquin, Amsterdam, the Netherlands). In previous experiments, we found that injection of control isotype antibody (IgG2a, (CRL-1908), American Type Culture Collection) did not induce an inflammatory reaction and was equivalent to PBS control [8]. Also, priming of LPS did not induce an inflammatory reaction. Therefore, isotype antibody and LPS control groups were omitted in these experiments. Two hours after injection, mice were bled from the carotid artery. Bronchoalveolar lavage fluid (BALF) was obtained from the right lung and centri-fuged at 2000 g. Supernatant was stored at -80°C for cytokine measurement. The left lung was weighed and dried in an oven at 65°C. The ratio of wet-to-dry weight represents tissue edema. In a second set of experiments, right lungs were used for preparation of homogenate after dilution in saline and Greenberger lysis buffer using a tissue homogenizer (Biospec products, Bartlesville, OK, USA). Supernatant was stored at -80°C for cytokine measurements. Left lungs were fixed in 4% formalin and embedded in paraffin for histopathology.

AssaysTotal protein levels (Bradford Protein Assay Kit, OZ Bioscience, Marseille, France) were measured in BALF. Keratinocyte-derived chemokine (KC), macrophage inflam-matory protein-2 (MIP-2) and interleukin-6 (IL-6) were measured in BALF, homogen-ate and plasma using enzyme-linked immunosorbent assay (ELISA), according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN, USA).Complement activation in BALF was determined by C3a and C5a ELISAs. Purified rat anti-mouse C3a (clone I87-1162) or purified rat anti-mouse C5a (clone I52-1486)

Muller.indd 292 25-7-2014 14:40:41


293

were used as capture antibody. Biotinylated rat anti-mouse C3a (clone I87-419) or biotinylated rat anti-mouse C5a (clone I52-278) was used as detection antibody for C3a and C5a, respectively (all from BD Biosciences). A standard curve for C3a and C5a was generated by serial dilutions of an in-house standard of maximal activated mouse serum, by incubating normal mouse serum (Sanquin Reagents) at 37oC for 1 week in the presence of sodium azide. Purified mouse C3a protein (native) and purified recombinant mouse C5a (both from BD Biosciences) were used to deter-mine concentration of C3a and C5a, respectively, in maximal activated mouse serum.C1-inhibitor levels in BALF were assessed by ELISA using monoclonal mouse anti human C1-inhibitor (clone RII, Sanquin) as capture antibody and biotinylated rabbit anti-human C1-inhibitor (Sanquin) for detection. Normal human plasma with known C1-inhibitor concentration was used as standard.

Lung injury scoreFour-micrometre sections were stained with hematoxylin-eosin (H&E) and analyzed by a pathologist who was blinded for group identity. Four pathologic parameters were scored on a scale of 0–4: (i) alveolar congestion, (ii) hemorrhage, (iii) leuko-cyte infiltration, and (iv) thickness of alveolar wall/hyaline membranes. A score of 0 represents normal lungs; 1, mild, <25% lung involvement; 2, moderate, 25–50% lung involvement; 3, severe, 50–75% lung involvement; 4, very severe, >75% lung involvement. The total histology score was expressed as the sum of the score for all parameters.

StatisticsData are expressed as mean ± standard deviation (SD) or median [interquartile range] when appropriate. Comparisons between groups were performed using one-way ANOVA followed by Student’s t-test or Kruskall-Wallis followed by Mann Whitney U-test, depending on data distribution. A p value of <0.05 was considered statistically significant. Statistical analyses were performed with Prism version 5.0 (GraphPad Software, San Diego, CA, USA).

Muller.indd 293 25-7-2014 14:40:41

Chapter 16

294

Results

Induction of TRALI with MHC-I antibodyInjection of MHC-I antibodies induced TRALI, illustrated by an increase in wet-to-dry ratio of the lungs and elevated BALF protein levels (figure 1). Levels of chemokines MIP-2 and KC and pro-inflammatory cytokine IL-6 were all significantly increased in both the lung homogenate and BALF (table 1). Also, TRALI was characterized by complement activation, as demonstrated by increased levels of C3a and C5a in BALF compared to controls (figure 1). The inflammatory reaction resulted in lower oxy-genation of mice (pO2 156 (±57) mmHg in controls vs. 133 (±67) mmHg in TRALI, p=0.41), which was statistically non-significant. Lung injury scores were higher in TRALI mice compared to controls (4.9 (±0.7) vs. (1.9 (±0.8), p<0.0001). TRALI induced a systemic inflammatory response reflected by increased plasma levels of KC and IL-6 (table 1).

The effect of C1-inhibitorAdministration of C1-inhibitor resulted in pulmonary C1-inhibitor levels of 12 ug/ml (± 6.4 ug/ml), with an activity of 99% (±20%), demonstrating effective levels locally. C1-inhibitor reduced C3a levels in BALF, but did not affect C5a levels compared to TRALI mice that were not treated with C1-inhibitor (figure 1). Injection of C1-inhibi-tor did not improve oxygenation (pO2 133 (±67) mmHg in TRALI vs. 93 (±30) mmHg in mice treated with C1-inhibitor, p=0.09), but tended to reduce pulmonary oedema and total protein levels; however, this was only a trend (figure 1). BALF levels of MIP-2, KC and IL-6 were unaffected. In lung homogenates, C1-inhibitor increased chemokine levels, although IL-6 was unaffected compared to untreated mice expe-riencing TRALI (table 1). Of note, treatment with C1-inhibitor improved lung injury scores from 4.9 (±0.7) in untreated mice to 3.5 (±0.8), (p=0.02) (figure 2). Systemic levels of KC and IL-6 were unaffected (table 1).

Muller.indd 294 25-7-2014 14:40:41


295

Figure 1: Wet-to-dry ratio of the lungs, BALF levels of total protein and complement levels in mice injected with antibodies that induce TRALI and those treated with C1-inhibitor.

Data expressed as mean, whiskers indicate minimum and maximum.*p<0.05**p<0.01***p<0.001TRALI = transfusion related acute lung injuryPBS = phosphate buffered salineC1-inh = C1-esterase inhibitorBALF = broncho-alveolar lavage fluid.

Muller.indd 295 25-7-2014 14:40:41

Chapter 16

296

Table 1: Pulmonary and systemic levels of pro-infl ammatory cytokine interleukin-6 (IL-6) and chemokines macrophage infl ammatory protein-2 (MIP-2) and kerati nocyte-derived chemokine (KC) in mice with TRALI and those treated with C1-inhibitor.

PBS controls TRALI TRALI + C1-inh p valueBALF (pg/ml)MIP-2 39 (7) 164 (74)a 127 (22) <0.01KC 56 (12) 245 (145)a 153 (79) <0.01IL-6 24 (9) 188 (97)a 114 (44) <0.01Lung homogenate (pg/ml)MIP-2 94 (14) 2038 (473)a 3748 (645)b <0.01KC 339 (145) 7337 (3451)a 31727 (12007)b <0.01IL-6 31 (9) 1629 (1188)a 957 (284) <0.01Plasma (pg/ml)KC 559 [285-957] 2103 [1476-6557]a 5064 [2939-5389] 0.01IL-6 220 [164-321] 2440 [1404-3890]a 1894 [832-1947] <0.01

Pulmonary levels are expressed as mean (± standard deviati on) and plasma levels are expressed as median [interquarti le ranges].PBS = phosphate buff ered salineTRALI = transfusion related acute lung injuryC1-inh = C1-inhibitorap<0.01 compared to PBS controlsbp<0.01 compared to TRALI

Figure 2: Representati ve histological secti ons of hematoxylin and eosin stained lungs of con-trol mice infused with PBS

(A), mice experiencing TRALI (B) and mice with TRALI who were treated with C1-inhibitor (C).

Muller.indd 296 25-7-2014 14:40:42


297

Discussion

In a ‘two-hit’ murine TRALI model, complement activation was demonstrated, reflected by elevated BALF levels of C3a and C5a. This is line with findings in other experimental models, which found that complement was involved in the inflamma-tory reaction or even a prerequisite for the development of TRALI [3,4]. In contrast, C5a receptors were not essential to develop an inflammatory reaction in a single hit model of antibody mediated TRALI [9]. Also in humans, data on the role of comple-ment activation in TRALI are conflicting [6,10].Administration of C1-inhibitor resulted in a reduction of C3a levels in BALF, whereas C5a levels were unaffected. The C1-inhibitor detected in the BALF was highly active (99%). Since C1-inhibitor is a highly effective inhibitor of the classical and the MBL activation pathway and the current TRALI model uses anti-MHC-I antibodies, the reduction of C3a levels suggests strong activation of the classical pathway of com-plement in this model. Activation of the classical pathway might also be an explana-tion for the persistent high C5a levels, since C5a can be generated by the activation of alternative pathway, which is not inhibited by C1-inhibitor. Moreover, coagula-tion factors generate C5a independent of the presence of C3 [11]. As markers of coagulation are increased in murine TRALI models as well as in TRALI patients [12], enhanced coagulation may have resulted in persistent high C5a levels.C1-inhibitor reduced C3a complement levels in the lung associated with an improve-ment in lung pathology scores. Also, vascular leakage in the lung tended to decrease, as reflected by a decrease in pulmonary oedema and pulmonary protein levels. In contrast, systemic and pulmonary levels of inflammatory cytokines were unaf-fected or even enhanced. A possible explanation for these opposing effects may be the systemic and pulmonary pro-coagulant state in TRALI, with increased lev-els of thrombin-antithrombin complexes as well as a decrease in fibrinolysis [12], which may have contributed to inflammation. Furthermore, a relatively high dose of C1-esterase inhibitor was used, which resulted in BALF levels of C1-inhibitor with high activity. Interestingly, high doses of C1-inhibitor have been shown to induce a pro-coagulant state [13], which might have further contributed to the enhanced levels of inflammatory parameters.

Muller.indd 297 25-7-2014 14:40:42

Chapter 16

298

The absence of systemic C1-inhibitor levels and pharmacokinetic data are a limita-tion to our study. However, the complete inhibition of C3a in BALF shows the pres-ence of active C1-inhibitor in the lungs. We do not think that use of LPS accounts for the absence of an effect of C1-inhibitor in our model, because we used a low priming dose and C1-inhibitor is beneficial in LPS-induced endotoxemia [7]. Another limita-tion of our study is the absence of data on activation of coagulation and degree of fibrinolysis, as previous reports found an association between the effect of C1-inhib-itor and thrombin-antithrombin levels [13].In conclusion, C1-inhibitor in murine TRALI attenuates complement activation asso-ciated with improved lung injury scores. However, these effects are accompanied by persistent inflammation, possibly due to a high dose of C1-inhibitor, which may have aggravated inflammation. Further research is warranted to elucidate the role of complement activation in TRALI before more targeted therapeutic strategies can be developed.

Muller.indd 298 25-7-2014 14:40:42


299

References

1. Goldman M, Webert KE, Arnold DM, et al.: Proceedings of a consensus conference: towards an understanding of TRALI. TransfusMedRev 2005;19: 2-31

2. Vlaar AP, Binnekade JM, Prins D, et al.: Risk factors and outcome of transfusion-related acute lung injury in the critically ill: a nested case-control study. Crit Care Med 38: 771-8

3. Seeger W, Schneider U, Kreusler B, et al.: Reproduction of transfusion-related acute lung injury in an ex vivo lung model. Blood 1990;76: 1438-44


5. Ambruso DR, Giclas P, Silliman CC, et al.: Complement Activation Associated with Trans-fusion Related Acute Lung Injury (TRALI). ASH Annual Meeting Abstracts 2004;104: 2721

6. Lucas G, Rogers S, Evans R, et al.: Transfusion-related acute lung injury associated with interdonor incompatibility for the neutrophil-specific antigen HNA-1a. Vox sanguinis 2000;79: 112-5

7. Wouters D, Wagenaar-Bos I, van HM, et al.: C1 inhibitor: just a serine protease inhibi-tor? New and old considerations on therapeutic applications of C1 inhibitor. Expert Opin Biol Ther 2008;8: 1225-40

8. Vlaar AP, Wolthuis EK, Hofstra JJ, et al.: Mechanical ventilation aggravates transfusion-related acute lung injury induced by MHC-I class antibodies. Intensive Care Med 36: 879-87



11. Huber-Lang M Fau - Sarma JV, Sarma Jv Fau - Zetoune FS, Zetoune Fs Fau - Rittirsch D, et al.: Generation of C5a in the absence of C3: a new complement activation pathway. Nat Med 2006;12: 682-7

12. Vlaar AP, Hofstra JJ, Determann RM, et al.: Transfusion-related acute lung injury in car-diac surgery patients is characterized by pulmonary inflammation and coagulopathy: a prospective nested case-control study. Crit Care Med 40: 2813-20

13. Horstick G, Berg O, Heimann A, et al.: Application of C1-esterase inhibitor during rep-erfusion of ischemic myocardium: dose-related beneficial versus detrimental effects. Circulation 2001;104: 3125-31

Muller.indd 299 25-7-2014 14:40:42

Muller.indd 300 25-7-2014 14:40:42

Chapter 17

Summary and general discussion

Marcella C.A. Müller and Nicole P. Juffermans

Muller.indd 301 25-7-2014 14:40:42

Chapter 17

302

Introduction

The inflammatory response in critical illness simultaneously affects the coagulation system and the resulting coagulopathy frequently occurs among the critically ill. Activation of the coagulation system with enhanced thrombin generation is accom-panied by attenuated fibrinolysis and reduced levels of anticoagulants. Altogether, this leads to a hypercoagulable state with the formation of (micro-) thrombi, impair-ing organ perfusion and thereby contributing to organ failure. However, as this state persists or occurs at a high rate, consumption of clotting factors and thrombocytes will result in deficiencies and hereby the hemostatic balance will shift to a hypocoag-ulable state with an increased bleeding tendency. As discussed in chapters 2 and 3, most clinically used coagulation assays (e.g. Prothrombin time (PT) and International Normalized Ratio (INR)) only reflect a part of the coagulation process and hereby they lack the ability to detect increased bleeding risk or a hypercoagulable state. In contrast, thromboelastography (TEG/ROTEM) assesses the whole process of clot formation and degradation and has the capability to also detect hypercoagulability and changes in fibrinolysis.Despite the limited value of these conventional coagulation assays, increased PT value is the most common trigger to administer fresh frozen plasma (FFP) to critically ill patients. Indeed many patients are transfused with FFP during their stay at the intensive care unit. Of note, the effects of FFP administration on in vivo hemostatic potential of critically ill patients are largely unknown and whether its administration indeed reduces a presumed enhanced bleeding risk has not been established. Also, administration of FFP is not without risks as it can cause adverse events, of which the most important is (transfusion related) acute lung injury.In this thesis, we studied the value of TEG/ROTEM to assess coagulation status in dif-ferent critically ill patient populations. In addition we investigated the effects of pro-phylactic FFP transfusion in critically ill patients with a coagulopathy. We assessed effects of FFP on different tests that evaluate the coagulation system. Also, the effect of FFP on the occurrence of bleeding and lung injury was studied. In the last part of this thesis, the pathophysiology of transfusion related lung injury (TRALI) is studied in more detail. In addition, we tested two potential therapeutic strategies for TRALI in a murine model. Furthermore, we performed a systematic review and

Muller.indd 302 25-7-2014 14:40:42


303

meta-analysis on the effect of implementing low risk plasma donor strategies on the onset of transfusion related acute lung injury.

Part I: Diagnosing coagulopathy in the critically ill: thromboelastometry

Summary of resultsAs described in chapter 2, TEG and ROTEM have been applied in different types of critically ill patients, including trauma, neurosurgical and post-cardiac arrest patients. Overall, evidence supporting the use of TEG/ROTEM to diagnose a hypo-coagulable state and hereby making an estimation of bleeding risk is limited. This also applies for the diagnosis of a hypercoagulable state and the risk assessment for development of multiple organ failure or thrombo-embolic complications. Main reasons for this low level of evidence are heterogeneity of the studies in design, use of different control groups, a lack of reference standards and variability in chosen endpoints.In chapter 3, the available literature on the application of TEG/ROTEM in sepsis was systemically reviewed. Heterogeneity between studies was large and there was risk of bias. None of the studies performed an assessment of bleeding risk based on TEG/ROTEM results. With respect of diagnosing hypercoagulability, TEG/ROTEM showed to be valuable to discriminate between sepsis patients with and without disseminated intravascular coagulation (DIC). Interestingly, a hypocoagulable, but not a hypercoagulable profile, correlated with DIC in sepsis patients and was associ-ated with more severe organ failure and excess mortality.We additionally studied the value of ROTEM to predict outcome in trauma patients in chapter 4. Of the severely injured trauma patients 50% dies in the acute phase due to bleeding. While organ failure is the main reason for late mortality after trauma. Thereby, we hypothesized that late mortality in trauma may be related to a hypercoagulable profile as detected by ROTEM. In a large cohort of trauma patients, we determined whether ROTEM profiles were predictive for the occurrence of mul-tiple organ failure and late mortality. Results indicated that early hypocoagulability was associated with more severe organ failure and higher mortality. Only 10% of the patients showed hypercoagulability early in trauma.

Muller.indd 303 25-7-2014 14:40:42

Chapter 17

304

Finally, we assessed the correlation of ROTEM with conventional coagulation tests in a cohort of critically ill patients (chapter 7). Correlation with the INR was moderate, which underlines the finding in observational studies that INR is a bad predictor of bleeding risk. ROTEM correlated well with prothrombin levels, platelet count and fibrinogen. Again, ROTEM profiles were more hypocoagulable in patients with DIC compared to those without DIC and ROTEM parameters correlated strongly with ISTH DIC score. Hereby we confirmed the potential value of ROTEM to discriminate critically ill patients with and without DIC.

DiscussionDespite the potential advantages of TEG/ROTEM over conventional coagulation assays, such as the possibility to diagnose hypercoagulability or DIC, application of TEG/ROTEM in critically ill patients has several limitations. First, reference values given by the manufacturer do not appear to discriminate between ‘normal’ and ‘a clinically relevant abnormal coagulation status’ in critically ill patients, when com-pared to conventional tests. Second, assays need to be further standardized and to date results of the two available assays (e.g. TEG and ROTEM) are not interchange-able. Thereby, a uniform definition of hyper- and hypocoagulability by viscoelastic tests is currently not available. These limitations preclude the incorporation of TEG and ROTEM into the standard diagnostic evaluation of critically ill patients with a suspected coagulopathy. However, as indicated in chapter 3 and confirmed in our small cohort described in chapter 7, TEG/ROTEM can be helpful in discriminating patients with and without DIC. Although a score to detect DIC using thromboelasto-metry which has been developed in a heterogeneous patient group [1] awaits valida-tion in critically ill patients, this observation is promising as a potential advantage of an improved and facilitated diagnosis of DIC. Also, TEG/ROTEM diagnosed coagu-lopathies could be helpful for more tailor-made administration of future therapies that interfere with the coagulation system.In addition to a potential diagnostic value, TEG/ROTEM results can be useful in prog-nostication of patients. In patients with sepsis, hypocoagulable profiles were asso-ciated with adverse outcomes. We also found that severely injured patients had more severe organ failure and excess mortality when they were hypocoagulable within the first 24 hours of admission. These findings are in line with results in a

Muller.indd 304 25-7-2014 14:40:42


305

large cohort of general intensive care patients, in whom a hypocoagulable profile at admission was also associated with an increased mortality [2]. One could argue that a hypocoagulable profile merely reflects severity of disease or injury. However, our results in trauma patients and the results of others in patients with sepsis [3] and the general intensive care population [2] showed that presence of hypocoagulability as detected by TEG/ROTEM was an independent risk factor for adverse outcome. These observations suggest that early activation of coagulation with reversal of hypocoagulability has potential beneficial effects. Indeed it has been demonstrated that enhanced coagulation during infection is functional, thereby preventing dis-semination of bacteria [4]. In trauma patients, early hypercoagulability could be an evolutionary response to prevent exsanguination, as was also suggested by oth-ers who have reported similar observations [5]. We speculate that these beneficial effects are lost when activation of coagulation is overwhelming and/or persistent and results in consumption of coagulation factors and platelets with a subsequent hypocoagulable state, hereby explaining the observed association between hypo-coagulability and adverse outcome. Overall, one may wonder whether hypocoagu-lopathy is a common pathway in the development of multiple organ failure or even death. This finding warrants further investigation.In conclusion, TEG/ROTEM seems a promising tool to assess coagulation status in different types of critically ill patients, in particular those suspected to have DIC. However, further standardization, availability of reference values and uniform defi-nitions are required before the technique can be implemented in standard clinical use at the intensive care unit. In addition, TEG/ROTEM has shown to be of additional value in the prognostication of critically ill and trauma patients. Further research is warranted to assess whether therapeutic strategies based on TEG/ROTEM profiles indeed improve patient outcomes and in general, the role of hypocoagulopathy in organ failure and dying.

Future directivesAs stated above, to further establish the value of TEG/ROTEM in critically ill patients, validated universally accepted definitions of hyper- and hypocoagulability in rela-tion to clinically relevant outcomes are necessary. As TEG/ROTEM based models have not been compared directly with the widely used ISTH DIC score, subsequently

Muller.indd 305 25-7-2014 14:40:42

Chapter 17

306

further research can be done on the identification of patients with an increased bleeding tendency and in need for correction of their coagulopathy. In addition, it should be further explored in which patients a hypercoagulable state contributes to adverse events, such as thromboembolic complications, multiple organ failure or even (delayed) cerebral ischemia.In sepsis patients, in addition to the abovementioned needs, further research into timing and changes of TEG/ROTEM profiles is highly needed, as activation of coagu-lation is a dynamic process. Currently, two observational prospective trials are being conducted (NCT00994877 and NCT00299949) on the value of TEG/ROTEM to diagnose DIC and to predict organ failure in sepsis. With regard to potential thera-peutic value of viscoelastic test results, these tests might be of additional value to select specific patient populations who are likely to benefit from therapies aimed at intervention in the coagulation cascade during sepsis. To date, such therapies have failed to improve outcome and for activated protein C this can be partially contributed to bleeding complications related to the treatment [6]. Although it is tempting to speculate that viscoelastic testing might be helpful to improve patient selection for these therapies, coagulation disturbances in sepsis are complex and these tests do not take into account levels of anticoagulant proteins, being the main targets of studied interventions. Therefore, further research is warranted to eluci-date whether there is a certain viscoelastic profile or a change over time in a profile that should trigger the institution of a therapy intervening in the coagulation cas-cade. Finally, if in the future viscoelastic testing would be used to select patients for anticoagulant therapy in sepsis, research should also focus on treatment targets. To date, data on viscoelastic testing and treatment effects of anticoagulant therapy for sepsis are extremely limited and it is questionable whether viscoelastic tests are sensitive enough to monitor treatment effects of such therapies.

Part II: Efficacy of plasma transfusion in critically ill patients

Summary of resultsThe second part of this thesis consisted of studies on the effectiveness of FFP transfu-sion in critically ill patients with a coagulopathy, as measured by INR in common eve-ryday clinical practice. We performed a multicentre randomized clinical trial on the

Muller.indd 306 25-7-2014 14:40:42


307

effectiveness of FFP to prevent bleeding complications in critically ill patients with a coagulopathy and the need to undergo an intervention (TOPIC trial). As described in chapter 5, procedure related bleeding complications did not differ between both groups. However, due to slow inclusion rate, the trial was stopped early and non-inferiority of omitting FFP transfusion could not be demonstrated. Although the secondary endpoint, occurrence and severity of lung injury, did not differ between groups, patients who were transfused with FFP had longer duration of mechanical ventilation. This might by explained by the higher incidence of ventilator-associated pneumonia in patients transfused with FFP, however we cannot preclude that low numbers of patients have contributed to these differences. The other secondary endpoint was correction of INR in response to FFP transfusion. Of note, only half of patients corrected to an INR of <1.5, a value frequently reported to be a trigger for FFP transfusion in clinical practice.In chapter 6 we investigated whether INR prolongation paralleled changes in other tests investigating hemostasis and evaluated the effects of FFP transfusion on these parameters. At baseline, patients indeed had reduced levels of individual coagula-tion factors, accompanied by a reduction of levels of anticoagulants. Also, onset of thrombin generation was impaired. Administration of FFP in a dose of 12 ml/kg resulted in an increase of individual factor levels, and levels of anticoagulants rose concomitantly. Thereby, thrombin generation remained nearly unaffected. Alto-gether, while conventional coagulation tests (INR and aPTT) improved in response to FFP because factor levels improved, FFP transfusion did not result in an altered hemostatic balance in these patients.As discussed above, in chapter 7 we demonstrated that critically ill patients with DIC have more hypocoagulable ROTEM profiles compared to those without DIC. For the first time we describe the effect of FFP on ROTEM variables in critically ill patients with a coagulopathy. Indeed, administration of FFP slightly improved ROTEM extem and fibtem profiles. However, observed increments in our study were only small and the clinical significance of the observed improvement is questionable. Of note, effect of FFP did not differ between patients with and without DIC.In chapter 8 the effects of FFP on inflammation and endothelial host response is discussed. This sub-study of the TOPIC trial showed that patients in this cohort had gene rally mildly elevated cytokine levels. Contrary to the hypothesis that FFP

Muller.indd 307 25-7-2014 14:40:42

Chapter 17

308

can induce lung injury and thereby may elicit an inflammatory response resem-bling ‘subclinical’ TRALI, FFP transfusion reduced systemic TNF-alpha levels. Other cytokine levels were unaffected by FFP. Markers of endothelial activation, factor VIII, von Willebrand factor and syndecan-1 all improved in response to FFP transfusion, suggesting a stabilizing effect of FFP on endothelial function.The last chapter (9) of this part consisted of an evaluation of the TOPIC trial. As the trial was stopped early due to limited inclusion we surveyed participating clinicians to determine why the trial was difficult to conduct. We found that although most physicians expressed a need for more evidence on the prophylactic use of FFP in critically ill patients with a coagulopathy, they were reluctant to include patients in the TOPIC trial. The main reason was personal beliefs about the preferable trans-fusion strategy. Together with a short window of opportunity to obtain informed consent, we concluded that future trials on the efficacy of FFP might not be feasible, at least in the Netherlands.

DiscussionWe carried out the first randomized controlled clinical trial on the efficacy of prophy-lactic FFP transfusion to prevent bleeding in coagulopathic ICU patients undergoing an invasive procedure. Although inclusion targets were not achieved, nevertheless, the study has several strengths. First, in contrast to other trials on the effectiveness of FFP in critically ill patients, we used the clinical relevant endpoint of procedure related bleeding. There was not even a trend towards a possible benefit. If any-thing, patients with FFP had more minor bleedings, which may suggest that most commonly performed invasive procedures can be safely carried out in critically ill patients with a coagulopathy.Increased INR and PT are important triggers to administer FFP transfusion. This is based on the assumption that increased INR values are associated with enhanced bleeding risk and that FFP transfusion prevents bleeding by correcting the INR [7]. Indeed patients with an increased INR indeed had attenuated thrombin generation due to reduced levels of multiple coagulation factors. However, levels of anticoag-ulants were reduced concomitantly, suggesting an unaltered hemostatic balance. In addition, despite reduced factor levels, most ROTEM variables were within ref-erence ranges and INR only correlated moderately with ROTEM. Of note, ROTEM

Muller.indd 308 25-7-2014 14:40:42


309

assesses the whole process of clot formation, while PT and INR only represent a part of it. These findings further confirm the limited value of PT and INR to assess in vivo coagulation status and predict bleeding risk. We believe that measuring PT and INR outside of the context to guide vitamin K antagonist therapy or to monitor the degree of liver failure, should not be done to assess risk of bleeding. Use of these tests should thereby be discouraged.Transfusion of FFP resulted in an increase of both individual factor levels and lev-els of anticoagulant proteins, thereby suggesting that FFP transfusion fails to alter hemostatic balance. Indeed the response of patients with and without bleed-ing complications following FFP transfusion did not differ. Moreover, FFP failed to clearly improve thrombin generation or ROTEM profiles. These observations under-line the lack of rationale in observational studies to administer FFP to ICU patients with an increased INR. The question whether FFP is beneficial in patients with a coagulopathy as diagnosed by other hemostatic tests, is not answered in this thesis. Of note, larger and earlier amounts of FFP improve outcome in observational studies of coagulopathic bleeding trauma patients, suggesting that FFP can increase hemo-static potential.Besides effects on hemostasis, we found intriguing other effects of FFP. In experi-mental studies of hemorrhagic shock, it was shown that FFP prevented disruption of the endothelium and improved microvascular perfusion [8]. In our cohort we also observed beneficial effects of FFP administration on markers of endothelial function. The reduction of syndecan-1 levels in response to FFP was accompanied by atten-uated levels of factor VIII and von Willebrand factor (vWF). As ADAMTS-13 levels increased, we postulate that the ADAMTS-13 present in administered FFPs reduces large vWF monomers. This is also the rationale of plasmapheresis in ADAMTS-13 deficiency. Presumably, this mechanism seems to also hold true for endothelial damage in the critically ill. Of note, levels of inflammatory cytokines remained unal-tered or even improved in response to FFP transfusion. These observations are not in line with reports of associations between FFP transfusion and the occurrence of lung injury in critically ill patients [9,10]. However, antibody-containing plasma might have contributed to these observations. Although we did not assess antibody levels, FFP used in our trial was manufactured after institution of the policy to defer female donors for FFP preparation, a measure known to reduce the risk of lung injury after

Muller.indd 309 25-7-2014 14:40:42

Chapter 17

310

transfusion. Thereby, the association of FFP with TRALI may be limited to antibody mediated lung injury, whereas administration of FFP might have beneficial effects in certain patient categories, not to correct coagulopathy, but to improve endothe-lial function. Further research is warranted to identify these patient categories, to assess the optimal dose and timing of FFP and to define outcome measures and treatment targets.In conclusion, we cannot reject the hypothesis that FFP is unnecessary to prevent bleeding complications in critically ill patients with a coagulopathy, due to limited inclusion in the TOPIC trial. However, administration of FFP failed to alter the hemo-static balance in these patients and only marginally affected thrombin generation and thromboelastometry. Despite this lack of effect on coagulation status, adminis-tration of FFP might have beneficial effects on endothelial function. Further research is warranted to confirm and further explore this observation.

Future directivesClinicians treating critically ill patients with a coagulopathy need a, preferably bed-side, test that assesses coagulation status and identifies patients with an increased bleeding risk. Possibly the establishment of reference values for thromboelasto-metry for critically ill patients can contribute to the value of this test in the ICU population. However, studies on the use of thromboelastometry to assess bleeding risk are warranted first. Alternatively, it would be interesting to determine whether the use of thrombin generation tests in the clinical setting is feasible and contributes to improved assessment of patients with a coagulopathy.If a patient indeed is hypocoagulable and has an enhanced bleeding risk, it is ques-tionable whether a dose of 12 ml/kg FFP is sufficient to induce a more procoagu-lant state. However, results from audits indicate that use of higher doses of FFP is not deemed feasible in clinical practice. Also, despite the expressed need of more evidence on the use of FFP by many experts in the field, we believe that another randomized controlled trial on effectiveness of FFP in critically ill patients with a coagulopathy is not feasible, at least not in the Netherlands. An alternative option for further research is to assess whether patients with an increased bleeding risk could benefit from the administration of four factor prothrombin complex concentrate

Muller.indd 310 25-7-2014 14:40:42


311

or fibrinogen. These products require less volume and are not associated with the occurrence of lung injury.In addition, it would be valuable to re-assess the bleeding risk of the most commonly performed procedures in the ICU, for example central venous catheter placement. Previous studies have already shown that procedure related bleeding complications are limited. However, the increasing use of ultrasound guided placement might have further reduced this risk, as was shown for pleural drainage [11]. Further confirma-tion of these observations could support clinicians to refrain from correction of a coagulopathy in case an intervention is necessary.Finally, the observation that administration of FFP might have some beneficial effects on endothelial function needs to be further elucidated. Mechanisms of action and potential effects on clinical outcomes in different patient categories need to be fur-ther investigated. Of note, in addition to the lack of a pro-inflammatory effect of FFP there are also indications that administration of FFP is associated with enhanced risk of infectious complications in critically ill. All these observations need confirmation and to be studied in further detail before FFP can be suggested as a new therapeutic or resuscitation fluid for critically ill patients.

Part III: Risks of transfusion in the critically ill: TRALI

Summary of resultsIn the last part of this thesis we further investigated the problem of lung injury fol-lowing transfusion. In chapter 10, an overview is given of risk factors for the onset of TRALI in critically ill patients. Main patient related risk factors are sepsis, shock, mechanical ventilation, cardiac surgery, the presence of a haematological malig-nancy and the need for massive transfusion. Blood product related risk factors con-sist of the presence of anti-leukocyte antibodies, bioactive response modifiers and the presence of the red cell storage lesion. Knowledge of these risk factors supports the ICU physician to take a more individualized approach in assessing the risk-bene-fit of the decision to transfuse a patient.We further elucidated the inflammatory changes in TRALI and in particular the con-tribution of damage associated molecular patterns (DAMPs). DAMPs ligate with the receptor for advanced glycation end products (RAGE), which is expressed on

Muller.indd 311 25-7-2014 14:40:42

Chapter 17

312

type I alveolar cells, endothelium and neutrophils. Both high-mobility group box 1 (HMGB1) and S100A12 are DAMPs that have been shown to contribute to inflam-matory insults in the lung. We studied their contribution to the inflammatory response in 14 cardiac surgery patients that developed TRALI (chapter 11). These patients were matched with transfused and non-transfused controls without lung injury. Although specific DAMPs including HMGB1 and sRAGE did not differ between groups, there was a trend towards higher levels of early DAMP S100A12 in patients who developed TRALI. S100A12 levels were associated with prolonged cardiopul-monary bypass, pulmonary inflammation, prolonged ventilation and hypoxemia. Hereby S100A12 may mediate the priming phase of acute lung injury in cardiac sur-gery patients who develop TRALI.Although the pathophysiology of TRALI is not fully elucidated, various preventive measures to reduce the risk of TRALI have been studied and implemented by blood supplying facilities across Europe and Northern America. An overview of TRALI prevention strategies is given in chapter 12 and in chapter 13 specific measures to prevent immune-mediated TRALI are discussed. The most striking TRALI prevention measure has been the deferral of females from donating plasma. However, evidence that the deferral of all female plasma donors indeed reduces TRALI incidence is mainly gathered in studies with a before-after design. In order to further strengthen evidence for low TRALI risk donor strategies, we performed a systematic review and meta-analysis (chapter 14). This meta-analysis made clear that the implementation of low risk TRALI plasma donor strategies indeed reduced onset of TRALI (OR 0.61 95%CI 0.42-0.88). The effect was most clear in patients with an increased risk to develop TRALI (OR 0.51 95%CI 0.29-0.90), while effects in general patient popula-tions showed a similar non-significant trend (OR 0.66 95%CI 0.40-1.09).Despite preventive measures TRALI still occurs, in particular in high-risk popula-tions such as critically ill patients. To date no therapeutic strategies are available. Therefore we studied the effects of two potential therapeutic strategies in a mouse model of TRALI. First, effects of corticosteroids were assessed. Despite an attenu-ated IL-6 host response, administration of methylprednisolone failed to prevent the development of lung injury in this ‘two hit’ murine TRALI model (chapter 15). We also assessed the effect of administration of C1-inhibitor in the same TRALI model (chapter 16). We confirmed that induction of TRALI resulted in activation of

Muller.indd 312 25-7-2014 14:40:42


313

the complement system, demonstrated by increased BALF levels of C3a and C5a. Administration of C1-inhibitor reduced pulmonary levels of C3a and improved lung injury scores. However, C5a and inflammatory cytokines levels were unaffected by C1-inhibitor.

DiscussionIn particular critically ill patients are at increased risk to develop TRALI. In these patients, some factors can be modified in order to reduce susceptibility for TRALI. Mechanical ventilation with injurious tidal volumes aggravate TRALI in a mouse model [12] and contributes to adverse outcome in patients with lung injury [13]. In addition, fluid restrictive management has been shown to improve outcomes in patients with acute lung injury [14]. Considering the similarities between acute lung injury and TRALI, it seems reasonable that these strategies also have beneficial effects in reducing TRALI risk. However, as most patient related risk factors can not be modified, the key point remains an adequate assessment of whether a patient is really likely to benefit from the administration of a blood product.Blood products also harbour risk factors for the development of a TRALI reaction. Anti-leukocyte antibodies are indisputably associated with the occurrence of TRALI, moreover the presence of anti-neutrophil antibodies (HNA-3a and 5b) is associ-ated with a more severe course of TRALI [15]. To prevent antibody mediated TRALI, important preventive measures have been implemented in the past decade. The deferral of female or antibody harbouring donors from donating FFP has resulted in a significant reduction of TRALI risk, in particular in at risk populations. This policy was shown to be feasible without interfering with adequate plasma supply. How-ever, immune mediated TRALI can also result from platelet transfusion and deferral of all female or antibody harbouring donors will hamper adequate supply. Alter-native measures to prevent platelet related antibody mediated TRALI could be the resuspension of platelets in male plasma or the addition of platelet additive solu-tion, although these measures have not been widely adopted yet. Finally, occur-rence of non-immune mediated TRALI is unaffected by abovementioned measures. Evidence regarding the contribution of prolonged storage of red blood cells and resulting accumulation of bioactive substances with neutrophil priming capacity to the occurrence of TRALI is conflicting [16,17]. These opposing results are probably

Muller.indd 313 25-7-2014 14:40:42

Chapter 17

314

not only affected by storage time, but also by storage conditions and remaining plasma content in implicated products. However, to date no evidence is available whether specific measures regarding preparation and storage of blood products indeed affect TRALI risk.Although preventive measures have reduced TRALI incidence, once TRALI occurs morbidity and mortality is considerable. Therefore therapeutic strategies are highly warranted. Corticosteroids are anecdotally used in patients with TRALI and in ARDS use of corticosteroids reduces mortality [18]. As ARDS and TRALI show pathophysi-ological similarities, we hypothesized that corticosteroids could attenuate a TRALI reaction. However, in a murine ‘two hit’ TRALI model, methylprednisolone failed to attenuate occurrence of lung injury, despite a systemic and pulmonary reduction of IL-6 levels. Possible explanations for an absence of effect could be suboptimal timing of methylprednisolone administration and contribution of inflammatory pathways to a TRALI reaction that are steroid independent, for example complement activa-tion. Therefore, we also evaluated the effect of C1-inhibitor in TRALI. In line with others, we indeed demonstrated complement activation in TRALI. Although admin-istration of C1-inhibitor reduced C3a levels, C5a levels remained unaffected, as were cytokine levels. We speculate that routes that are independent from C1-esterase, such as the alternative route and routes that activate coagulation, contribute to persisting high C5a levels and subsequent inflammation.In conclusion, critically ill patients carry an increased risk to develop TRALI, but implementation of strategies to reduce TRALI risk of plasma containing blood prod-ucts by blood supplying facilities, have significantly attenuated TRALI risk. However, therapeutic strategies are still warranted, as these measures do not prevent all TRALI cases. Experimental evidence does not support the use of corticosteroids for the treatment of TRALI. Also, effectiveness of the administration of C1-inhibitor could not be demonstrated. Further understanding of TRALI pathophysiology and the role complement activation is warranted before more targeted therapeutic strategies can be developed.

Future directivesAlthough the implementation of preventive measures has reduced TRALI incidence, there are still several issues that need to be addressed for future prevention of

Muller.indd 314 25-7-2014 14:40:42


315

TRALI. First universal use of the consensus definition is highly warranted, hereby donors implicated in TRALI can be identified. In addition, guidelines on donor defer-ral and antibody screening are warranted. Also, standardization of HLA and HNA tests to screen donors is needed. Further research has to elucidate whether storage conditions and duration indeed affect occurrence of TRALI. Results of a trial compar-ing fresh blood with standard transfusion care in critically ill patients are expected within a year. With regard to plasma products, no TRALI cases have been reported with the use of pooled plasma and it has been successfully used for years in Scandi-navian countries. Also, the Netherlands will change to pooled plasma instead of FFP this year. Whether this change indeed results in a further reduction of plasma asso-ciated TRALI cases, including in high-risk patients such as critically ill, will become clear in the near future. Finally, additional evidence on other factors or substances within blood products that contribute to adverse reactions, in particular in patients at risk, will help to develop a more tailored transfusion policy for critically ill patients.Also clinicians’ awareness about the risk and benefits of transfusion of blood prod-ucts needs to be improved, as many audits have shown high rates of inappropriate use of blood products. A restrictive transfusion policy has shown to be safe and advantageous in different types of critically ill patients [19-21]. However, at least in the Netherlands, this evidence has not yet been incorporated in clinical guidelines or become routine practice in daily care. In addition, as discussed previously, inappro-priate use of plasma products is also still high [22]. Therefore, improved knowledge and awareness among clinicians will hopefully contribute to reduced inappropriate use of blood products, with a concomitant reduction in undesired side effects and costs. The implementation of computer-supported decision making is a promising tool in this perspective.

Muller.indd 315 25-7-2014 14:40:42

Chapter 17

316

References


2. Johansson PI, Stensballe J, Vindelov N, et al.: Hypocoagulability, as evaluated by thrombelastography, at admission to the ICU is associated with increased 30-day mor-tality. Blood Coagul Fibrinolysis 2010;21: 168-74

3. Ostrowski SR, Windelov NA, Ibsen M, et al.: Consecutive thrombelastography clot strength profiles in patients with severe sepsis and their association with 28-day mor-tality: A prospective study. J Crit Care 2013;28: 317

4. Massberg S, Grahl L, von Bruehl ML, et al.: Reciprocal coupling of coagulation and innate immunity via neutrophil serine proteases. Nat Med 2010;16: 887-96

5. Branco BC, Inaba K, Ives C, et al.: Thromboelastogram Evaluation of the Impact of Hypercoagulability in Trauma Patients. Shock 2014;41(3): 200-7

6. Kalil AC, LaRosa SP: Effectiveness and safety of drotrecogin alfa (activated) for severe sepsis: a meta-analysis and metaregression. Lancet Infect Dis 2012;12: 678-86







13. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342: 1301-8

14. Wiedemann HP, Wheeler AP, Bernard GR, et al.: Comparison of two fluid-management strategies in acute lung injury. N Engl J Med 2006;354: 2564-75

15. Davoren A, Curtis BR, Shulman IA, et al.: TRALI due to granulocyte-agglutinating human neutrophil antigen-3a (5b) alloantibodies in donor plasma: a report of 2 fatalities. Trans-fusion 2003;43: 641-5

Muller.indd 316 25-7-2014 14:40:42


317

16. Vlaar AP, Kulik W, Nieuwland R, et al.: Accumulation of bioactive lipids during storage of blood products is not cell but plasma derived and temperature dependent. Transfusion 2011;51(11): 2358-66


18. Meduri GU, Golden E, Freire AX, et al.: Methylprednisolone infusion in early severe ARDS: results of a randomized controlled trial. Chest 2007;131: 954-63

19. Villanueva C, Colomo A, Bosch A, et al.: Transfusion strategies for acute upper gastroin-testinal bleeding. N Engl J Med 368: 11-21

20. Hebert PC, Wells G, Blajchman MA, et al.: A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999;340: 409-17

21. Walsh TS, Boyd JA, Watson D, et al.: Restrictive versus liberal transfusion strategies for older mechanically ventilated critically ill patients: a randomized pilot trial. Crit Care Med 2013;41: 2354-63


Muller.indd 317 25-7-2014 14:40:42

Muller.indd 318 25-7-2014 14:40:42

Chapter 18

Dutch summary

Muller.indd 319 25-7-2014 14:40:42

Chapter 18

320

Inleiding

Bij ernstige zieke patiënten die op de intensive care (IC) moeten worden opge-nomen, is er bijna altijd sprake van een ontstekingsreactie. Een ontstekingsreac-tie heeft effect op het functioneren van het stollingssysteem van deze patiënten (coagulopathie). Coagulopathie komt dan ook zeer frequent voor bij ernstig zieke patiënten. Ontsteking activeert het stollingssysteem waarbij er meer trombine (stol-selvorming) gegenereerd wordt en daarnaast is de afbraak van stolsels (fibrinolyse) verminderd, net als de spiegels van de natuurlijk voorkomende antistollingseiwitten. Deze veranderingen induceren een verhoogde stollingsneiging, hypercoagulabiliteit. Hierbij kunnen (micro-) trombi ontstaan die leiden tot verminderde doorbloeding van de organen en zo falen van orgaan functie kunnen induceren. Uiteindelijk kan hierdoor een beeld van multi-orgaanfalen (MOF) ontstaan. Echter, als deze staat van geactiveerde stolling voortduurt, kan het evenwicht verschuiven van hyper- naar een hypocoagulabiliteit (verminderde stollingsneiging) door uitputting van de voor-raad stollingsfactoren, waarbij juist een verhoogde bloedingsneiging kan ontstaan.De testen die standaard in de kliniek gebruikt worden om de stollingsneiging van een patiënt te evalueren (Protrombine Tijd (PT) en International Normalized Ratio (INR)), weerspiegelen slechts een beperkt deel van het gehele proces van stolsel-vorming en -afbraak. Om een inschatting te maken van het reële bloedingsrisico bij IC patiënten is de waarde van deze bepalingen dan ook zeer beperkt en daar-naast ontberen deze testen de mogelijkheid om een staat van verhoogde stolling te detecteren. Het is wel mogelijk om het hele proces van stolselvorming en -afbraak te monitoren met behulp van visco-elastische testen: thromboelastografie (TEG) of thromboelastometrie (ROTEM). Tevens kan met behulp van deze testen een ver-hoogde stollingsneiging aangetoond worden. Klinische toepassing en ervaring van deze methode bij IC patiënten is echter nog niet wijdverbreid.Ondanks de beperkte waarde van de conventionele stollingstesten om een indicatie te geven van het in vivo bloedingsrisico, is een verhoogde PT waarde de belang-rijkste reden om Fresh Frozen Plasma (FFP) aan IC patiënten toe te dienen. Het is dan ook gebleken dat een substantieel aantal patiënten tijdens hun verblijf op de IC wordt getransfundeerd met FFP. De effecten van FFP op de in vivo stollingsbalans van kritisch zieke patiënten zijn echter grotendeels onbekend. Daarnaast is het niet

Muller.indd 320 25-7-2014 14:40:42

Dutch summary

321

bekend of toediening van FFP aan deze patiënten inderdaad een vermindering van het eventueel verhoogde bloedingsrisico geeft. Daarentegen lijkt het gebruik van FFP bij IC patiënten niet zonder risico te zijn. De belangrijkste schadelijke bijwerking is het ontstaan van (transfusie gerelateerde) longschade.In dit proefschrift hebben we onderzocht wat de waarde van TEG/ROTEM is in het vaststellen van stollingsstoornissen bij verschillende typen kritische zieke patiënten. Daarnaast hebben we de effecten van een profylactische FFP transfusie onderzocht bij IC patiënten met een coagulopathie. De effecten van FFP op het stollingssysteem hebben we met meerdere testen geëvalueerd. Tevens hebben we de effecten van FFP op het voorkomen van bloedingen en het optreden van longschade onderzocht. In het laatste deel van dit proefschrift wordt de pathofysiologie van de belangrijk-ste bijwerking van transfusie, transfusie gerelateerde longschade (TRALI), in meer detail bestudeerd. Tevens hebben we twee potentiele therapeutische strategieën voor TRALI getest in een diermodel. Tenslotte hebben we een systematische samen-vatting en meta-analyse verricht van de beschikbare literatuur met betrekking tot het effect van de implementatie van TRALI risico reductie strategieën door bloed-banken.

Deel I: Diagnostiek van coagulopathie bij IC patiënten: thromboelastometrie

In hoofdstuk 2 wordt een overzicht gegeven van de ervaring met TEG en ROTEM bij verschillende typen IC patiënten, waaronder trauma-, neurochirurgische- en cardiochirurgische patiënten. Samenvattend is er beperkt bewijs dat TEG/ROTEM van toegevoegde waarde is in het opsporen van een hypocoagulabele staat bij deze patiënten indien ze opgenomen zijn op de IC. De waarde van deze testen voor het inschatten van het bloedingsrisico van de individuele patiënt is vooralsnog dan ook beperkt. Hetzelfde geldt voor het vaststellen van een hypercoagulabele staat en een inschatting op de kans op orgaan falen of het optreden van trombo-embolische complicaties. De belangrijkste oorzaken voor deze beperkte mate van bewijs zijn de grote verschillen in studie opzet, gebruik van verschillende controle groepen en een gebrek aan referentie waarden en verschil in gekozen eindpunten.

Muller.indd 321 25-7-2014 14:40:42

Chapter 18

322

In hoofdstuk 3 wordt een systematische samenvatting gegeven van de beschikbare literatuur ten aanzien van het gebruik van TEG/ROTEM bij patiënten met sepsis. Ook hier waren de verschillen tussen de studies groot en was er risico op bias. Er zijn geen studies die een inschatting hebben gemaakt van het bloedingsrisico gebaseerd op TEG/ROTEM resultaten. Wat betreft hypercoagulabiliteit, lijkt TEG/ROTEM wel van additionele waarde om een onderscheid te maken tussen patiënten met en zon-der diffuse intravasale stolling (DIS). Niet een hyper-, maar een hypocoagulabel TEG/ROTEM profiel correleert met het voorkomen van DIS bij patiënten met sepsis. Een hypocoagulabel profiel is daarbij geassocieerd met ernstiger orgaan falen en een toegenomen sterftekans.In hoofdstuk 4 hebben we de waarde van ROTEM bestudeerd om de uitkomst van patiënten na een trauma te voorspellen. Van de patiënten die overlijden ten gevolge van een trauma, overlijdt de helft in de acute fase ten gevolge van verbloeding. Van de patiënten die later alsnog overlijden, is het optreden van MOF de belangrijkste doodsoorzaak. Onze hypothese was dat laat overlijden na trauma samen zou hangen met een hypercoagulabel profiel in de ROTEM test. In een groot cohort van trauma patiënten hebben we gekeken of ROTEM profielen voorspellende waarde hadden voor het optreden van orgaan falen en sterfte in deze patiënten. De resultaten lie-ten zien dat een hypocoagulabel profiel vroeg na opname geassocieerd was met een ernstigere vorm van MOF en een toegenomen sterftekans. Een hypercoagulabel profiel komt voor bij 10% van de patiënten die een ernstig trauma heeft gehad.Tenslotte hebben we in een cohort IC patiënten de correlatie bepaald van ROTEM met de conventionele stollingstesten (Hoofdstuk 7). Hierbij bleek dat de correla-tie met INR matig was. ROTEM uitslagen correleerden goed met protrombine spie-gels, trombocyten aantal en fibrinogeen waarden. Patiënten met DIS hadden ook in dit cohort meer hypocaogulabele ROTEM uitslagen dan patiënten zonder DIS. De ROTEM waarden correleerden daarnaast sterk met de ISTH DIC score, de meest gebruikte score om de diagnose DIS te stellen. Deze bevindingen bevestigen de potentiele waarde van ROTEM om IC patiënten met en zonder DIS van elkaar te onderscheiden.

Concluderend lijkt TEG/ROTEM een veelbelovende test om de stollingsstatus te evalueren bij verschillende soorten ernstig zieke patiënten, met name indien er

Muller.indd 322 25-7-2014 14:40:42

Dutch summary

323

de verdenking op DIS bestaat. Er is echter verdere standaardisatie, vaststelling van referentie waarden en de ontwikkeling van uniforme definities noodzakelijk, voor-dat de techniek geïmplementeerd kan worden in de dagelijkse praktijk. Daarnaast is het gebruik van TEG/ROTEM van aanvullende waarde om de prognose van ernstig zieke en trauma patiënten in te schatten. Er is verder onderzoek nodig om te bepa-len of therapeutische interventies gebaseerd op TEG/ROTEM profielen inderdaad de uitkomsten van deze patiënten verbeteren.

Deel II: Effectiviteit van plasma transfusie bij intensive care patiënten

Het tweede deel van dit proefschrift bestaat uit studies naar de effectiviteit van FFP transfusie bij intensive care patiënten met een coagulopathie. We hebben een gerandomiseerde studie verricht in meerdere ziekenhuizen. Doel van deze studie was het effect van een profylactische FFP transfusie te evalueren op het voorkomen van bloedingscomplicaties bij IC patiënten, met een coagulopathie, die een inter-ventie moesten ondergaan (TOPIC trial). Deze studie liet zien dat er geen verschil was in het aantal procedure gerelateerde bloedingscomplicaties tussen patiënten die wel of geen FFP transfusie kregen (hoofdstuk 5). Echter, door tegenvallende inclusie aantallen is de studie vroegtijdig afgebroken. Hierdoor was het niet moge-lijk geplande statistische analyses uit te voeren en kon non-inferiority van het niet geven van FFP aan deze patiënten niet worden aangetoond. Het secundaire eind-punt van de studie was het optreden van longschade, hierbij was er geen verschil tussen patiënten die wel of geen FFP hadden ontvangen. Wel was de beademings-duur van patiënten die getransfundeerd waren met FFP langer, mogelijk ten gevolge van het vaker voorkomen van beademing gerelateerde longontstekingen in deze patiëntengroep. Deze verschillen kunnen echter ook het gevolg zijn van de kleine patiënten aantallen. Correctie van INR na FFP transfusie was een ander secundair eindpunt, waarbij opviel dat slechts de helft van de patiënten na FFP transfusie een INR van minder dan 1.5 had, terwijl deze waarde een veel gerapporteerde grens is voor het toedienen van FFP.In hoofdstuk 6 hebben we onderzocht of INR verlenging inderdaad gepaard ging met veranderingen in andere stollingstesten en wat het effect van FFP op deze testen

Muller.indd 323 25-7-2014 14:40:42

Chapter 18

324

was. Bij aanvang van de studie hadden patiënten inderdaad verlaagde niveaus van de individuele stollingsfactoren. Dit ging echter ook gepaard met verlaagde niveaus van natuurlijk voorkomende antistollingseiwitten. Hierbij was de trombine genera-tie verminderd. Het toedienen van FFP (dosis 12 ml/kg) aan deze patiënten resul-teerde in een stijging van het aantal stollingsfactoren met daarbij ook een stijging van de natuurlijke antistollingsfactoren. De transfusie had echter nauwelijks effect op de mate van trombine generatie. Samenvattend resulteert FFP transfusie in een verbetering van conventionele stollingstesten (INR en PT) doordat de individuele stollingsfactoren toenemen, maar treedt er geen verschuiving op in de hemostati-sche balans van deze patiënten.Zoals al eerder genoemd, laten we in hoofdstuk 7 zien dat IC patiënten met DIS meer hypocoagulabele ROTEM profielen hadden in vergelijking tot patiënten die geen DIS hebben. Voor het eerst beschrijven we in dit hoofdstuk de effecten van FFP op ROTEM variabelen bij IC patiënten met een coagulopathie. De transfusie van FFP resulteerde in enige verbetering van de ROTEM profielen, maar het effect was minimaal en de klinische relevantie is discutabel. Het effect van FFP op ROTEM vari-abelen verschilde niet tussen patiënten met en zonder DIS.In hoofdstuk 8 bespreken we de effecten van FFP op ontsteking en het endotheel, de binnenbekleding van de vaatwand. Deze sub-studie van de TOPIC trial laat zien dat de ontstekingseiwitten in deze patiënten matig verhoogd waren. In tegenstel-ling tot de hypothese dat FFP transfusie een ontstekingsreactie uitlokt, resulteerde FFP transfusie in een afname van TNF-alfa. De spiegels van de andere ontstekings-eiwitten veranderden niet na toedienen van FFP. Markers van endotheel activatie (factor VIII, von Willebrand factor en Syndecan-1) verbeterden allen na een trans-fusie met FFP, dit wekt de suggestie dat FFP mogelijk een endotheel stabiliserende effect heeft.Het laatste hoofdstuk in dit deel (hoofdstuk 9) beschrijft de evaluatie van de TOPIC studie. Omdat de studie vroegtijdig afgebroken is, in verband met tegenvallende inclusie getallen, hebben we een enquête verricht onder de artsen de betrokken waren bij de inclusie in de studie. Het doel was inzicht te krijgen in eventuele barriè-res die een succes vol verloop van de studie in de weg hebben gestaan. Deze enquête toonde aan dat er wel degelijk behoefte bestaat aan meer bewijs ten aanzien van het profylactisch gebruik van FFP bij IC patiënten met een coagulopathie. Toch

Muller.indd 324 25-7-2014 14:40:43

Dutch summary

325

waren artsen terughoudend om patiënten in de studie te includeren. De belangrijk-ste redenen hiervoor waren persoonlijke voorkeuren ten aanzien van de te volgen transfusie strategie. Daarnaast was er maar een beperkt tijdvak waarin patiënten geïncludeerd konden worden. Op basis van de resultaten van deze enquête, lijkt het vrijwel niet mogelijk om in Nederland in de toekomst nog onderzoek te verrichten naar de effectiviteit van FFP.

Concluderend kunnen we, op basis van bovenstaand onderzoek, de hypothese dat FFP niet nodig is om bloedingen bij IC patiënten met een coagulopathie te voorko-men, niet verwerpen. De oorzaak hiervoor is de beperkt inclusie van patiënten in de TOPIC studie. Wel is duidelijk dat FFP toediening bij deze patiënten de hemo-statische balans niet verandert en dat de effecten op trombine generatie en ROTEM zeer beperkt zijn. Ondanks dit gebrek aan effect van FFP op de stollingsstatus, heeft FFP mogelijk wel gunstige effecten op de functie van het endotheel. Verder onder-zoek is nodig om deze bevindingen te bevestigen en de betekenis ervan verder te exploreren.

Deel III: Transfusie gerelateerde longschade bij intensive care patienten

In het laatste deel van dit proefschrift beschrijven we studies met betrekking tot het optreden van transfusie gerelateerde longschade (TRALI). In hoofdstuk 10 wordt een overzicht gegeven van de risicofactoren voor het ontstaan van TRALI bij patiënten op de intensive care. De belangrijkste patiënt gerelateerde risicofacto-ren zijn sepsis, shock, kunstmatige beademing, hartchirurgie, de aanwezigheid van een hematologische maligniteit en de noodzaak om een massale transfusie toe te dienen. Bloedproduct gerelateerde risicofactoren bestaan uit de aanwezigheid van antileukocyten antistoffen en bioactieve stoffen in het te transfunderen product. Daarnaast spelen beschadigingen die aan de rode bloedcel ontstaan als gevolg van de opslag van het product een rol. Kennis van deze risicofactoren maakt het voor de IC-arts mogelijk om per patiënt een afweging te maken tussen de voor- en nadelen van een eventuele transfusie.

Muller.indd 325 25-7-2014 14:40:43

Chapter 18

326

Indien een patiënt toch TRALI ontwikkelt, gaat dit gepaard met een ontstekingsre-actie in de long. We hebben dit ontstekingsproces verder proberen te ontrafelen en daarbij in het bijzonder gekeken naar de bijdrage van “Damage Associated Molecu-lar Patterns” (DAMPs). Deze moleculen, die vrijkomen bij weefselschade en ontste-king, binden aan de RAGE (Receptor for Advanced Glycation Endproducts) receptor, welke onder meer voorkomt op het oppervlakte van de alveoli, het endotheel en neutrofielen. Zowel “high-mobility group box 1” (HMGB-1) als S100A12 zijn DAMPs die een rol spelen in ontstekingsreacties in de long. Wij hebben hun bijdrage aan ontstekingsreacties bestudeerd bij 14 patiënten die hartchirurgie ondergingen en TRALI ontwikkelden (Hoofdstuk 11). Deze patiënten werden vergeleken met patiën-ten die wel of geen transfusie hadden ontvangen, maar geen longschade hadden ontwikkeld. De DAMPs HMGB1 en oplosbaar RAGE verschilden niet tussen de groe-pen, wel was er een trend naar hogere S100A12 waarden in patiënten die TRALI ontwikkelden. Daarnaast waren S100A12 waarden geassocieerd met langere duur aan de hart-longmachine, mate van ontsteking in de long, langere beademingsduur en lagere zuurstof spanning in het bloed. S100A12 speelt daarom mogelijk een rol in de eerste fase (“priming phase”) van het ontstaan van longschade bij patiënten na hartchirurgie.Ondanks dat er nog veel onduidelijkheden zijn rondom de pathofysiologische ver-anderingen die optreden bij patiënten met TRALI, hebben bloedbanken in Europa en Noord Amerika de laatste jaren verschillende TRALI preventie strategieën ont-wikkeld en geïmplementeerd. In hoofdstuk 12 wordt een overzicht gegeven van deze verschillende preventieve strategieën, daarnaast wordt in hoofdstuk 13 dieper ingegaan op specifieke maatregelen om immuun-gemedieerde TRALI te voorkomen. De meest opvallende TRALI preventie maatregel is het uitsluiten van vrouwen van plasma donatie. De onderbouwing voor deze en andere maatregelen om het donor gerelateerde TRALI risico te reduceren, komt echter met name uit studies die een “voor-na” opzet hebben. Om de onderbouwing voor de effectiviteit van deze maat-regelen te verstevigen hebben we een meta-analyse verricht (hoofdstuk 14). Uit deze meta-analyse komt duidelijk naar voren dat de implementatie donor strate-gieën om het TRALI risico te reduceren, inderdaad leidt tot een reductie van de kans op TRALI (OR 0.61 95%CI 0.42-0.88). Het effect was het grootst in patiënten die een verhoogd risico hadden om TRALI te ontwikkelen (OR 0.51 95%CI 0.29-0.90), terwijl

Muller.indd 326 25-7-2014 14:40:43

Dutch summary

327

de effecten in de algemene populatie een zelfde, maar niet significante, trend ver-toonden (OR 0.66 95%CI 0.40-1.09).Ondanks deze preventieve maatregelen komt TRALI nog steeds voor, met name bij patiënten die een verhoogd risico hebben, zoals ernstig zieke patiënten. Tot op heden is er geen behandeling voor TRALI beschikbaar. Daarom hebben we twee mogelijke behandelingen voor TRALI onderzocht in een muis model. Ten eerste hebben we gekeken naar de effecten van toediening van steroïden. Echter, ondanks een afname van IL-6, konden steroïden niet voorkomen dat er toch longschade ont-stond in dit “two-hit” muis model van TRALI (hoofdstuk 15). In het zelfde model hebben we ook het effect C1-inhibitor, een remmer van het complement systeem, geëvalueerd (hoofdstuk 16). Inderdaad leidt de inductie van TRALI tot complement activatie, te zien aan verhoogde C3a en C5a waarden in de longvloeistof. Toediening van C1-inhibitor zorgde voor een reductie van de verhoogde C3a waarden en lagere long schade scores in histologische preparaten, maar had geen effect op de hoogte van de ontstekings parameters en C5a waarden.

Concluderend kunnen we zeggen dat intensive care patiënten een verhoogde kans hebben om TRALI te ontwikkelen. De implementatie van TRALI risico reductie stra-tegieën voor plasma producten door bloedbanken, hebben het risico op het ont-staan van TRALI duidelijk verminderd. Deze maatregelen voorkomen echter niet alle gevallen van TRALI en daarom is het nog steeds van belang dat er een behande-ling voor TRALI wordt ontwikkeld. Uit experimenteel onderzoek komt naar voren dat de toediening van corticosteroïden of C1-inhibitor niet effectief is. Verdere ver-heldering van de pathofysiologische veranderingen in TRALI en de mogelijk rol van complement activatie hierin zijn nodig, voordat meer gerichte therapeutische stra-tegieën ontwikkeld kunnen worden.

Muller.indd 327 25-7-2014 14:40:43

Muller.indd 328 25-7-2014 14:40:43

AppendicesAuthors and affi liati ons

List of publicati ons

Acknowledgements

Research portf olio

Curriculum vitae

Muller.indd 329 25-7-2014 14:40:43

Appendices

330

Authors and Affiliations

Achmed Achouiti, Center for Experimental and Molecular Medicine (CEMM), Aca-demic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Dr. M. Sesmu Arbous, Department of Intensive Care Medicine, Leiden University Medical Center, Leiden University, Leiden, The Netherlands

Kirsten Balvers, Trauma Unit, Department of Surgery, Academic Medical Center, University of Amsterdam, The Netherlands

Dr. Charlotte J. Beurskens, Department of Intensive Care Medicine and Laboratory of Experimental Intensive Care and Anesthesiology (L.E.I.C.A.), Academic Medical Center, University of Amsterdam, The Netherlands

Dr. Jan M. Binnekade, Department of Intensive Care Medicine, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Dr. Louis Boon, Bioceros, Utrecht, The Netherlands

Prof. Karim Brohi, Trauma Sciences, Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London, United Kingdom

Nicola Curry, National Health Service Blood & Transplant and Oxford Radcliffe Hos-pitals, John Radcliffe Hospital, Oxford, United Kingdom

Dr. Christine Gaarder, Department of Traumatology, Oslo University Hospital, Oslo, Norway

Prof. dr. J. Carel Goslings, Trauma Unit, Department of Surgery, Academic Medical Center, University of Amsterdam, The Netherlands

Muller.indd 330 25-7-2014 14:40:43

Appendices

331

Prof. dr. Rob J. de Haan, Clinical Research Unit, Academic Medical Center, University of Amsterdam, The Netherlands

Prof. dr. Evert de Jonge, Department of Intensive Care Medicine, Leiden University Medical Center, Leiden University, Leiden, The Netherlands

Dr. Nicole P. Juffermans, Department of Intensive Care Medicine and Laboratory of Experimental Intensive Care and Anesthesiology (L.E.I.C.A.), Academic Medical Center, University of Amsterdam, The Netherlands

Atilla Karakus, Department of Intensive Care Medicine, Diakonessenhuis, Utrecht, The Netherlands

Dr. J. Henriette Klinkspoor, Departement of General Clinical Chemistry, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands

Knut M. Kolstadbraaten, Department of Traumatology, Oslo University Hospital, Oslo, Norway

Kelly Maijoor, Center for Experimental and Molecular Medicine (CEMM), Academic Medical Center, Univeristy of Amsterdam, Amsterdam, The Netherlands

Prof. dr. Joost C.M. Meijers, Department of Experimental Vascular Medicine, Aca-demic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Dr. Leendert Porcelijn, Sanquin Diagnostic Services, Department of Diagnostic Immunohematology, Sanquin, Amsterdam, The Netherlands

Prof. dr. Dick J van Rhenen, Sanquin Bloodbank South West Region and Erasmus University, Rotterdam, The Netherlands

Dr. Joris J.T.H. Roelofs, Department of Pathology, Academic Medical Center, Univer-sity of Amsterdam, Amsterdam, The Netherlands

Muller.indd 331 25-7-2014 14:40:43

Appendices

332

Claire Rourke, Trauma Sciences, Blizard Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London, United Kingdom

Prof. dr. Marcus J. Schultz, Department of Intensive Care Medicine and Laboratory of Experimental Intensive Care and Anesthesiology (L.E.I.C.A.), Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Dr. Koenraad F. van der Sluijs, Laboratory of Experimental Intensive Care and Anes-thesiology (L.E.I.C.A.), Academic Medical Center, University of Amsterdam, The Netherlands

Dr. Angelique M.E. Spoelstra-de Man, Department of Intensive Care Medicine, VU Medical Center, VU University, Amsterdam, The Netherlands

Dr. Simon Stanworth, National Health Service Blood & Transplant and Oxford Rad-cliffe Hospitals, John Radcliffe Hospital, Oxford, United Kingdom

Danielle van Stein, Sanquin Bloodbank South West Region, Rotterdam, The Nether-lands

Marleen Straat, Department of Intensive Care Medicine and Laboratory of Experi-mental Intensive Care and Anesthesiology (L.E.I.C.A.), Academic Medical Center, University of Amsterdam, The Netherlands

Ingrid Stroo, Department of Immunopathology, Sanquin Research, Amsterdam, The Netherlands

Dr. Pieter Roel Tuinman, Departement of Intensive Care Medicine, VU Medical Center, VU University, Amsterdam, the Netherlands

Anita M. Tuip, Laboratory of Experimental Intensive Care and Anesthesiology (L.E.I.C.A.), Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Muller.indd 332 25-7-2014 14:40:43

Appendices

333

Dr. Cornelis van t Veer, Center for Experimental and Molecular Medicine (CEMM), Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Dr. Roel Vink, Department of Intensive Care Medicine, Tergooiziekenhuizen, Hilver-sum, The Netherlands

Dr. Alexander P. Vlaar, Department of Intensive Care Medicine and Laboratory of Experimental Intensive Care and Anesthesiology (L.E.I.C.A.), Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Prof. dr. Margreeth B. Vroom, Department of Intensive Care Medicine and Labora-tory of Experimental Intensive Care and Anesthesiology (L.E.I.C.A.), Academic Medi-cal Center, University of Amsterdam, Amsterdam, The Netherlands

Dr. Diana Wouters, Department of Immunopathology, Sanquin Research, Amster-dam, The Netherlands

Dr. Sacha S. Zeerleder, Department of Hematology, Academic Medical Center, Uni-versity of Amsterdam, Amsterdam, The Netherlands

Muller.indd 333 25-7-2014 14:40:43

Appendices

334

Publications

Müller MC, van Stein D, Binnekade JM, van Rhenen DJ, Vlaar APJ: Low risk TRALI donor strategies and the impact on the onset of transfusion-related acute lung injury; a meta-analysis. Transfusion, in press.

Müller MC, Arbous MS, Spoelstra – de Man AM, Vink R, Karakus A, Straat M, Bin-nekade JM, de Jonge E, Vroom MB, Juffermans NP: Transfusion of fresh frozen plasma in critically ill patients with a coagulopathy prior to invasive procedures: a randomized clinical trial. Transfusion 2014; DOI 10.111/trf.12750

Müller MC, Tuinman PR, Vlaar AP, Tuip AM, Maijoor K, Achouiti A, Veer van t C, Vroom MB, Juffermans NP: Contribution of damage-associated molecular patterns to transfusion-related acute lung injury in cardiac surgery. Blood Transfusion 2014;2: 1-8

Balvers K, Müller MC, Juffermans NP, The utility of thromboelastometry (ROTEM) and thromboelastography (TEG) in non-bleeding ICU patients. Annual Update in Intensive Care and Emergency Medicine 2014; Springer International Publishing: 583-591

Müller MC, Meijers JCM, Vroom MB, Juffermans NP: Utility of thromboelastography and/or thromboelastometry in sepsis – a systematic review. Critical Care 2014;18: R30

Müller MC, Tuinman PR, Sluijs van der KF, Boon L, Roelofs JJ, Vroom MB, Juffermans NP: Methylprednisolone fails to attenuate lung injury in a mouse model of transfu-sion related lung injury. Transfusion 2014;54: 996-1001

Müller MC, Stroo I, Wouters D, Zeerleder SS, Roelofs JJ, Boon L, Vroom MB, Juffer-mans NP: The effect of C1-inhibitor in a murine model of transfusion-related acute lung injury. Vox Sanguinis 2014;107: 71-75

Muller.indd 334 25-7-2014 14:40:43

Appendices

335

Jong de IM, Müller MCA, Peterson GM, Polle SW: An uncommon cause of of portal vein thrombosis. Netherlands Journal of Medicine 2013;71: 431,435

Tuinman PR, Müller MC, Juffermans NP: Transfusion in myocardial infarction: blood-curdling? JAMA Internal Medicine. 2013;173: 1156-7

Tuinman PR, Muller MC, Jongsma G, Hegeman MA, Juffermans NP: High-dose Ace-tylsalicylic Acid is Superior to Low Dose as Well as to Clopidogrel in Preventing LPS-induced Lung Injury in Mice. Shock 2013;40: 334-8

Müller MCA, Juffermans NP: Transfusion Related Lung Injury – risk factors in ICU patients. Annual Update in Intensive Care and Emergency Medicine 2013; Springer Berlin Heidelberg: 527-535

Kwakman PH, Müller MC, Binnekade JM, van den Akker JP, de Borgie CA, Schultz MJ, Zaat SA: Medical–grade honey does not reduce skin colonization at central venous catheter insertion sites of critically ill patients: a randomized controlled trial. Critical Care 2012;16: R214

Müller MCA, Porcelijn L, Vlaar AP: Prevention of Immune-mediated Transfusion-related Acute Lung Injury; from Bloodbank to Patient. Current Pharmaceutical Design 2012; 18(22): 3241-8

Müller MC, Juffermans NP: Transfusion-related acute lung injury: a preventable syn-drome? Expert Review of Hematology 2012;5: 97-106

Müller MC, de Jonge E, Arbous MS, Spoelstra-de Man AM, Karakus A, Vroom MB, Juffermans NP: Transfusion of fresh frozen plasma in non-bleeding ICU patients – TOPIC trial: study protocol for a randomized controlled trial. Trials 2011;23: 12:266

Khorsand N, Veeger NJ, Muller M, Overdiek JW, Huisman W, van Hest RM, Meijer K: Fixed versus variable dose of prothrombin complex concentrate for counteracting vitamin K antagonist therapy. Transfusion Medicine 2011;21: 116-23

Muller.indd 335 25-7-2014 14:40:43

Appendices

336

Müller MC, Juffermans NP: Anemia and blood transfusion and outcome on the intensive care unit. Critical Care 2010;14: 438

Müller MC, Lagarde SM, Germans MR, Juffermans NP: Cerebral air embolism after arthrography of the ankle. Medical Science Monitor 2010;16: CS92-4

Müller MC, van der Werf SD: A young woman from Cameroon with rectal blood loss, intestinal schistosomiasis and rectosigmoid carcinoma. Nederlands Tijdschrift voor Geneeskunde 2008;152:951-5

Müller M, Willén R, Stotzer PO: Colonoscopy and SeHCAT for investigation of chronic diarrhea. Digestion 2004;69: 211-8

Schoon EJ, Müller MC, Vermeer C, Schurgers LJ, Brummer RJ, Stockbrügger RW: Low serum and bone vitamin K status in patients with longstanding Crohn’s disease: another pathogenetic factor of osteoporosis in Crohn’s disease? Gut 2001;48:473-7

Muller.indd 336 25-7-2014 14:40:43

Appendices

337

Acknowledgements

De woorden altijd of nooit zijn niet van toepassing in de geneeskunde, maar het doen van onderzoek en de voltooiing van een proefschrift vormen hierop een uit-zondering. Als onderzoeker heb je altijd hulp nodig, of het nu gaat om dier experi-menten, klinische studies, analyses of het schrijven van stukken. Daarnaast kan je als mens en dokter die onderzoek doet nooit zonder de steun van je familie, vrienden en collega’s. Deze plek wil ik dan ook gebruiken om degenen te bedanken die me, op wat voor manier dan ook, geholpen of gesteund hebben bij de totstandkoming van dit proefschrift.

Patiënten en familie van patiënten opgenomen op de Intensive Care, veel dank voor jullie vertrouwen in het wetenschappelijk onderzoek en jullie toestemming tot deel-name hieraan.

Prof. dr. M.B. Vroom, promotorBeste Margreeth, heel veel dank voor het vertrouwen dat je op meerdere momen-ten in me gesteld hebt. Het aanbieden van een fellow-plek en gevolgd door de mogelijkheid om onderzoek te combineren met werk als intensivist bij jou op de afdeling, hebben me veel gebracht. Ik heb het als een groot voorrecht ervaren om met je te werken en van je te leren, nogmaals veel dank.

Dr. N.P Juffermans, co-promotorBeste Nicole, zonder jouw ononderbroken ondersteuning had ik dit proefschrift nooit zo (relatief) snel af kunnen ronden. Op de momenten dat ik geen of alleen negatieve resultaten zag, wist jij me altijd wel weer een positieve kant van het geheel te laten zien. Heel veel dank voor de zeer prettige samenwerking, je enthou-siasme, enorme geduld, laagdrempelige beschikbaarheid en altijd opbouwende kri-tieken. Ik vind het absoluut bewonderenswaardig hoe jij het doen van zoveel, goed en interessant onderzoek weet te combineren met de kliniek en je interesses buiten het ziekenhuis.

Muller.indd 337 25-7-2014 14:40:43

Appendices

338

Prof. dr. M.J. Schultz, prof. dr. M.W. Hollmann, prof.dr. E. De Jonge, prof. dr. J.C. Goslings, prof. dr. P.W. Kamphuisen en dr. S.S. Zeerleder, dank voor uw bereidheid zitting te nemen in mijn promotiecommissie en uw kritische beoordeling van dit proefschrift.

Dank aan alle onderzoekers van de afdeling intensive care en het L.E.I.C.A die ik te allen tijde lastig kon vallen met vragen over het beademen van muizen, afdraaien van lab samples, opmaken van Graphpad plaatjes en natuurlijk de inclusie van patiënten in mijn studie. In het bijzonder veel dank aan Marleen, Lieuwe, Charlotte, Hamid, Friso, Kirsten, Hemmik, Frank, Luuk, Sabrine, Lonneke, Maryse, Roosmarijn, Ilse en Laura.

Bedankt medewerkers van het L.E.I.C.A. voor jullie hulp aan mij als ‘lab-naïeve’ clini-cus. Hierbij ben ik met name veel dank verschuldigd aan Anita en Geartsje, naast dat ik het vooral erg gezellig vond om met jullie samen te werken, waren jullie profes-sionele kennis en kunde onmisbaar voor het slagen van de experimenten.

Alexander, Pieter Roel en Walter, samenwerken met jullie was efficiënt, gezellig, ont-spannend en een goede aanvulling op mijn sport/Ajax kennis.

Onderzoekers betrokken bij de TOPIC trial, Sesmu Arbous, Rob de Wilde, Atilla Karakus, Angelique Spoelstra en Roel Vink, veel dank voor jullie inspanningen om patiënten te includeren en data te verzamelen.

Jan Binnekade, dank voor je vele statistische analyses en geduldige uitleg daarbij. Je kamergenoten Frederique en Annelou, dank voor de goede tips voor de praktische uitvoering van onderzoek en het schrijven van de reviews.

Collega intensivisten in het AMC, Dave, Sander, Erik Jan, Anne-Cornelie, Catherine, Janneke, Robert, Thomas, Wim en Jos, veel dank voor de gezelligheid, prettige samenwerking en leermomenten.

Muller.indd 338 25-7-2014 14:40:43

Appendices

339

Mary-Anne, Erika en alle verpleegkundigen van de intensive- en medium care van het AMC, dank voor alle hulp met het uitvoeren van de studies. Jullie vormden een belangrijke bijdrage aan het plezier dat ik heb gehad tijdens mijn werkjaren in het AMC.

Collega intensivisten van het MCH/Bronovo, Hans, Dirk Jan, Paul, Ron, Trudy, Mathilde, Wim, Thomas, Nienke, Jörn en Stephan, dank dat jullie me de ruimte en tijd hebben gegeven om dit proefschrift af te kunnen ronden. Verheug me op de samenwerking de komende jaren.

Lieve Boukje en Nelien, geweldig dat jullie mijn paranimfen willen zijn. Het lukt door alle drukte, afstand en logistiek met kinderen en mannen (en dit proefschrift) niet altijd om elkaar zo vaak te zien als we zouden willen. Gelukkig staan kwantiteit en kwaliteit los van elkaar en ik hecht dan ook zeer aan onze vriendschap.

Lieve familie Rothbarth, ik denk niet dat ik me een betere schoonfamilie kan wen-sen, Flip en Anneke in het bijzonder heel veel dank voor jullie altijd flexibele en zeer liefdevolle opvang van Freek en Ties.

Bedankt lieve Katinka, Mark en Naud, voor jullie niet aflatende interesse, altijd attente berichtjes en liefde.

Lieve peer en mamma, jullie liefdevolle opvoeding met onvoorwaardelijk vertrou-wen, waarbij jullie altijd voor ons klaarstaan, is de basis voor alles. Veel dank en liefs.

Mijn allerliefste mannen Joost, Freek en Ties, ik kan me niet vaak genoeg realiseren hoe geweldig en bijzonder jullie zijn, ik houd van jullie.

Muller.indd 339 25-7-2014 14:40:43

Appendices

340

Research portfolio

Name Marcella C.A. MüllerPhD period May 2010 – September 2014PhD supervisors Prof. dr. Margreeth B. Vroom and dr. Nicole P. Juffermans

PhD and clinical training

Courses2010 Laboratory animals2010 Good clinical practice/BROK2010 Radiation protection – level 52010 Teach the teacher – communication and feedback2011 Update on neuromonitoring – ESICM, Rome, Italy2011 Simulation Based Airway Management Training in Intensive Care2011 Hospital-MIMS Provider Course2012 Teach the teacher – modernization of education

Conferences and presentations2010 NVIC dagen, Ede, The Netherlands Poster presentation “Fatal abdominal compartment syndrome”2011 European Society Intensive Care Medicine, Berlin, Germany Oral presentation “The effect of medicinal honey on skin colonization

around central venous catheters”2012 American Thoracic Society, San Francisco, USA Oral presentation “Damage-associated molecular patterns in transfu-

sion-related acute lung injury” Poster presentation “Methylprednisolone attenuates pulmonary

inflammation in an in vivo model of transfusion related lung injury”2013 International Symposium on Intensive Care and Emergency Medicine,

Brussels, Belgium Poster presentation “C1 inhibitor attenuates pulmonary inflammation

in an in vivo model of transfusion related lung injury”

Muller.indd 340 25-7-2014 14:40:43

Appendices

341

2014 American Thoracic Society, San Diego, USA Poster presentation “Male only fresh frozen plasma donation policy

and the impact on the onset of transfusion-related acute lung injury; a meta-analysis”

Muller.indd 341 25-7-2014 14:40:43

Appendices

342

Curriculum Vitae

Marcella Müller was born on the 16th of February 1977 in Groningen, the Netherlands. She attended secondary school at the Sint Maartens College in Haren (Gn), after graduating in 1995 she started her medical studies at the University of Maastricht. During her medical training she participated in a research project in the depart-ment of gastroenterology of the Academic Hospital of Maastricht (dr. E.J. Schoon). In her final year she spent 8 months in Göteborg (Sweden) where she did a research project in the department of gastroenterology of the Sahlgrenska Hospital (dr. P.O. Stotzer) and a clinical rotation in Mölsndals Hospital. After graduating in 2002 (cum laude), she started working as resident internal medicine in the Westeinde Hospital, The Hague, under supervision of dr. P.H.L.M. Geelhoed-Duijvestein. Subsequently she continued her training in internal medicine at the Leiden University Medical Center, Leiden in 2007 (prof. dr. J.A. Romijn). In 2008 she started her fellowship at the Department of Intensive Care Medicine of the Academic Medical Center, Amsterdam, under supervision of prof. dr. M.B. Vroom. In 2010 she completed her training as internist-intensivist and obtained the European Diploma in Intensive Care (EDIC). Thereafter she continued her work as an intensivist at the Academic Medical Center and started her PhD project on coagulopathy and plasma transfusion in criti-cally ill patients, which has resulted in this thesis. Since July 2013 she works at the Medisch Centrum Haaglanden/Bronovo in The Hague. Marcella lives in The Hague together with her husband Joost Rothbarth and their two sons Freek and Ties.

Muller.indd 342 25-7-2014 14:40:43

Coagulopathy and plasma transfusion

in critically ill patients

Marcella Müller


Marcella M

üller

UITNODIGINGVoor het bijwonen van de

openbare verdediging van het proefschrift


door

Marcella Müller

op donderdag 4 september 2014

om 12:00 uur

in de Agnietenkapel van de Universiteit van AmsterdamOudezijds Voorburgwal 231

te Amsterdam

Na afloop bent u van hartewelkom op de receptie.

ParanimfenBoukje van DijkNelien Hylkema

[email protected]

Marcella MüllerJohannes Bildersstraat 35

2596 EE Den Haag06-41810633

[email protected]

Proefschrift Muller

Documents

Transcript of Proefschrift Muller