Management of Bleeding Following Major Trauma - A European Guideline

Open Access

Available online http://ccforum.com/content/11/1/R17

Vol 11 No 1Research

Management of bleeding following major trauma: a European guidelineDonat R Spahn1, Vladimir Cerny2, Timothy J Coats3, Jacques Duranteau4, Enrique Fernndez-Mondjar5, Giovanni Gordini6, Philip F Stahel7, Beverley J Hunt8, Radko Komadina9, Edmund Neugebauer10, Yves Ozier11, Louis Riddez12, Arthur Schultz13, Jean-Louis Vincent14 and Rolf Rossaint15

1Department of Anesthesiology, University Hospital Zurich, Rmistrasse 100, 8091 Zurich, Switzerland2Charles University in Prague, Faculty of Medicine in Hradec Krlov, Department of Anaesthesiology and Intensive Care Medicine, University Hospital Hradec Krlov, Sokolska 581, 50005 Hradec Krlov, Czech Republic3Leicester Royal Infirmary, Accident and Emergency Department, Infirmary Square, Leicester LE1 5WW, UK4Department of Anaesthesia and Intensive Care, University of Paris XI Facult de Mdecine Paris-Sud, 63 rue Gabriel Pri, 94276 Le Kremlin-Bictre, France5Department of Emergency and Critical Care Medicine, University Hospital Virgen de las Nieves, ctra de Jan s/n, 18013 Granada, Spain6Department of Anaesthesia and Intensive Care, Ospedale Maggiore, Largo Nigrisoli 2, 40100 Bologna, Italy7Department of Orthopaedic Surgery, Denver Health Medical Center, University of Colorado Medical School, 777 Bannock Street, Denver, CO 80204, USA8Departments of Haematology, Pathology and Rheumatology, Guy's & St Thomas' Foundation Trust, Lambeth Palace Road, London SE1 7EH, UK9Department of Traumatology, General and Teaching Hospital Celje, 3000 Celje, Slovenia10Institute for Research in Operative Medicine, University of Witten/Herdecke, Ostmerheimerstrasse 200, 51109 Kln (Merheim), Germany11Department of Anaesthesia and Intensive Care, Universit Ren Descartes Paris 5, AP-HP, Hopital Cochin, 27 rue du Fbg Saint-Jacques, 75014 Paris, France12Department of Surgery and Trauma, Karolinska University Hospital, 171 76 Solna, Sweden13Ludwig-Boltzmann-Institute for Experimental and Clinical Traumatology, Donaueschingenstrasse 13, 1200 Vienna, Austria14Department of Intensive Care, Erasme Hospital, University of Brussels, Belgium, route de Lennik 808, 1070 Brussels, Belgium15Department of Anaesthesiology, University Hospital Aachen, Pauwelsstrae 30, 52074 Aachen, Germany

Corresponding author: Rolf Rossaint, [email protected]

Received: 8 Nov 2006 Revisions requested: 21 Dec 2006 Revisions received: 8 Jan 2007 Accepted: 13 Feb 2007 Published: 13 Feb 2007

Critical Care 2007, 11:R17 (doi:10.1186/cc5686)This article is online at: http://ccforum.com/content/11/1/R17 2007 Spahn et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Introduction Evidence-based recommendations can be madewith respect to many aspects of the acute management of thebleeding trauma patient, which when implemented may lead toimproved patient outcomes.

Methods The multidisciplinary Task Force for AdvancedBleeding Care in Trauma was formed in 2005 with the aim ofdeveloping guidelines for the management of bleeding followingsevere injury. Recommendations were formulated using anominal group process and the GRADE (Grading ofRecommendations Assessment, Development, and Evaluation)hierarchy of evidence and were based on a systematic review ofpublished literature.

Results Key recommendations include the following: The timeelapsed between injury and operation should be minimised forpatients in need of urgent surgical bleeding control, and patientspresenting with haemorrhagic shock and an identified source ofbleeding should undergo immediate surgical bleeding controlunless initial resuscitation measures are successful. A damagecontrol surgical approach is essential in the severely injuredpatient. Pelvic ring disruptions should be closed and stabilised,followed by appropriate angiographic embolisation or surgicalbleeding control, including packing. Patients presenting withhaemorrhagic shock and an unidentified source of bleedingshould undergo immediate further assessment as appropriateusing focused sonography, computed tomography, serumlactate, and/or base deficit measurements. This guideline also

ACS = American College of Surgeons; aPTT = activated partial thromboplastin time; CT = computerised tomography; DPL = diagnostic peritoneal lavage; FAST = focused abdominal sonography in trauma; FFP = fresh frozen plasma; GRADE = Grading of Recommendations Assessment, Devel-opment, and Evaluation; Hb = haemoglobin; Hct = haematocrit; ICU = intensive care unit; KIU = kallikrein inhibitory units; MeSH = Medical Subject Page 1 of 22(page number not for citation purposes)

Heading; MSCT = multi-slice spiral computed tomography; NIH = National Institutes of Health; PCC = prothrombin complex concentrate; PEEP = positive end-expiratory pressure; PT = prothrombin time; RBC = red blood cell; RCT = randomised controlled trial; rFVIIa = recombinant activated coagulation factor VII; TRALI = transfusion-related acute lung injury; TRICC = Transfusion Requirements in Critical Care.

Critical Care Vol 11 No 1 Spahn et al.

reviews appropriate physiological targets and suggested useand dosing of blood products, pharmacological agents, andcoagulation factor replacement in the bleeding trauma patient.

Conclusion A multidisciplinary approach to the management ofthe bleeding trauma patient will help create circumstances in

which optimal care can be provided. By their very nature, theseguidelines reflect the current state-of-the-art and will need to beupdated and revised as important new evidence becomesavailable.

IntroductionTraumatic injury is the leading cause of death worldwideamong persons between 5 and 44 years of age [1] andaccounts for 10% of all deaths [2]. In 2002, 800,000 injury-related deaths in Europe accounted for 8.3% of total deaths[3]. Because trauma affects a disproportionate number ofyoung people, the burden to society in terms of lost productiv-ity, premature death, and disability is considerable. Despiteimprovements in trauma care, uncontrolled bleeding contrib-utes to 30% to 40% of trauma-related deaths and is the lead-ing cause of potentially preventable early in-hospital deaths [4-6].

Resuscitation of the trauma patient with uncontrolled bleedingrequires the early identification of potential bleeding sourcesfollowed by prompt action to minimise blood loss, to restoretissue perfusion, and to achieve haemodynamic stability. Mas-sive bleeding in trauma patients, defined here as the loss ofone blood volume within 24 hours or the loss of 0.5 blood vol-umes within three hours, is often caused by a combination ofvascular injury and coagulopathy. Contributing factors to trau-matic haemorrhage include both surgical and non-surgicalbleeding, prior medication, comorbidities, and acquired coag-ulopathy [7].

Here, we describe early diagnostic measures to identify haem-orrhage that should trigger surgical or radiological interven-tions in most cases. Specific interventions to manage bleedingassociated with pelvic ring injuries and hypothermia are dis-cussed, as well as recommendations for the optimal applica-tion of fluid, pharmacological, blood product, and coagulationfactor therapy in trauma patients.

These guidelines for the management of the bleeding traumapatient were developed by a multidisciplinary group of Euro-pean experts and designated representatives from relevantprofessional societies to guide the clinician in the early phasesof treatment. The recommendations presented here are basedon a critical survey of the published literature and were formu-lated according to a consensus reached by the author group.Many of the critical issues faced by the treating physician havenot been, and for ethical or practical reasons may never be,addressed by prospective randomised clinical studies, andtherefore the formulation and grading of the recommendationspresented here are weighted to reflect both this reality and thecurrent state-of-the-art.

Materials and methodsThese recommendations were formulated and graded accord-ing the Grading of Recommendations Assessment, Develop-ment, and Evaluation (GRADE) hierarchy of evidence outlinedby Guyatt and colleagues [8] and are summarised in Table 1.Comprehensive computer database literature searches wereperformed using the indexed online databases MEDLINE/PubMed and the Cochrane Library. Lists of cited literaturewithin relevant articles were also screened. The primary inten-tion of the review was to identify prospective randomised con-trolled trials (RCTs) and non-randomised controlled trials,existing systematic reviews, and guidelines. In the absence ofsuch evidence, case control studies, observational studies,and case reports were considered.

Boolean operators and Medical Subject Heading (MeSH) the-saurus keywords were applied as a standardised use of lan-guage to unify differences in terminology into single concepts.Appropriate MeSH headings and subheadings for each ques-tion were selected and modified based on search results. Thescientific questions posed that led to each recommendationand the MeSH headings applied to each search are listed inAdditional file 1. Searches were limited to English languageabstracts and human studies; gender and age were not lim-ited. No time-period limits were imposed on searches unlessthe search result exceeded 300 hits. Original publicationswere evaluated for abstracts that were deemed relevant. In thecase of a guideline update, searches were limited to the timeperiod following the publication of the last version of the guide-line. If an acceptable systematic review or meta-analysis wasidentified, searches to update the data were typically limited tothe time period following the search cutoff date reported in thereview. Original publications were evaluated according to thelevels of evidence developed by the Oxford Centre for Evi-dence-Based Medicine (Oxford, Oxfordshire, UK) [9].

The selection of the scientific inquiries to be addressed in theguideline, screening, and grading of the literature to beincluded and formulation of specific recommendations wereperformed by members of the Task Force for Advanced Bleed-ing Care in Trauma, a multidisciplinary, pan-European group ofexperts with specialties in surgery, anaesthesia, emergencymedicine, intensive care medicine, and haematology. The coregroup was formed in 2004 to produce educational material oncare of the bleeding trauma patient [10], on which a subse-quent review article was based [11]. The Task Force con-Page 2 of 22(page number not for citation purposes)

sisted of the core group, additional experts in haematologyand guideline development, and representatives of relevant


European professional societies, including the EuropeanShock Society, the European Society for Anaesthesia, theEuropean Society for Emergency Medicine, the EuropeanSociety for Intensive Care Medicine, and the EuropeanTrauma Society. The European Hematology Associationdeclined the invitation to send a representative to join the TaskForce. Task Force members participated in a workshop on thecritical appraisal of medical literature. The nominal group proc-ess included four face-to-face meetings supplemented by sev-eral Delphi rounds [12]. The guideline development group metin June 2005 to define the scientific questions to beaddressed in the guideline and again in October 2005 to final-ise the scientific scope of the guidelines. Selection, screening,and grading of the literature and formulation of recommenda-tions were accomplished in subcommittee groups consistingof at least three members via electronic or telephone commu-nication. After distribution of the recommendations to theentire group, a further meeting of the Task Force was held inApril 2006 with the aim of reaching a consensus on the draftrecommendations from each subcommittee. After final refine-ment of specific recommendations among committeemembers, a subset of the Task Force met in July 2006 to final-

the endorsing organisations in September and October 2006.An updated version of the guideline is anticipated in due time.

In the GRADE system for assessing each recommendation,the letter attached to the grade of recommendation reflectsthe degree of literature support for the recommendation,whereas the number indicates the level of support for the rec-ommendation assigned by the committee of experts. Recom-mendations are grouped by category and somewhatchronologically in the treatment decision-making process, butnot by priority or hierarchy.

ResultsI. Initial resuscitation and prevention of further bleedingEvidence to support the initial phase of resuscitation and pre-vention of further bleeding is lacking, and there have been fewstudies on the effect of coagulopathy on outcome. Patientswith a coagulopathic condition have worse outcomes thanpatients of the same injury severity without a clotting distur-bance [13,14], and patients with head injury also have worseoutcomes in association with a coagulopathy [15]; however,contrary to popular belief, there is no evidence that patients

Table 1

Grading of recommendations after Guyatt et al. [8]

Grade of recommendation Clarity of risk/benefit Quality of supporting evidence Implications

1A

Strong recommendation, high-quality evidence

Benefits clearly outweigh risk and burdens, or vice versa

Randomised controlled trials (RCTs) without important limitations or overwhelming evidence from observational studies

Strong recommendations, can apply to most patients in most circumstances without reservation

1B

Strong recommendation, moderate-quality evidence


RCTs with important limitations (inconsistent results, methodological flaws, indirect, or imprecise) or exceptionally strong evidence from observational studies

Strong recommendations, can apply to most patients in most circumstances without reservation

1C

Strong recommendation, low-quality or very low-quality evidence


Observational studies or case series Strong recommendation but may change when higher-quality evidence becomes available

2A

Weak recommendation, high-quality evidence

Benefits closely balanced with risks and burden

RCTs without important limitations or overwhelming evidence from observational studies

Weak recommendation, best action may differ depending on circumstances or patients' or societal values

2B

Weak recommendation, moderate-quality evidence

Benefits closely balanced with risks and burden

RCTs with important limitations (inconsistent results, methodological flaws, indirect, or imprecise) or exceptionally strong evidence from observational studies

Weak recommendation, best action may differ depending on circumstances or patients' or societal values

2C

Weak recommendation, low-quality or very low-quality evidence

Uncertainty in the estimates of benefits, risks, and burden; benefits, risk, and burden may be closely balanced

Observational studies or case series Very weak recommendation, other alternatives may be equally reasonablePage 3 of 22(page number not for citation purposes)

ise the manuscript document. The document was approved by with head injury are more likely to develop a coagulopathy thanother severely injured patients [16].


There is no evidence as to whether the degree of initial bleed-ing affects coagulopathy. Coagulopathy is predicted by asystolic blood pressure of below 70 mm Hg [17], but thiscould be either a direct effect of bleeding or an associatedeffect of injury severity. There is no high-level scientific evi-dence that the initial amount of bleeding affects the patient'soutcome; however, the experience of treating physicians isthat uncontrolled haemorrhage is associated with poor out-come. Common experience is that wound compression pre-vents bleeding, but it is not known whether this reduces theincidence of coagulopathy. There is also no evidence that tellsus whether control of acid-base balance during initial resusci-tation affects outcome.

There is evidence to support expedient care for patients fol-lowing traumatic injury; however, no study has examined therelationship between outcomes in patients transported to dif-ferent types of hospital facilities and the amount of bleeding.Pre-hospital bleeding not controlled by compression andsplintage requires rapid surgical or radiological intervention.

Recommendation 1We recommend that the time elapsed between injury andoperation be minimised for patients in need of urgent surgicalbleeding control (grade 1A).

RationaleTrauma patients in need of emergency surgery for ongoinghaemorrhage demonstrate better survival if the elapsed timebetween the traumatic injury and admission to the operatingtheatre is minimised [18-21]. Although there are no ran-domised control studies to verify this statement, there are ret-rospective studies that provide enough evidence for earlysurgical intervention in these patients. This is particularly truefor patients who present in an exsanguinated state or in severehaemorrhagic shock due to penetrating vascular injuries[18,19]. In accordance with these observations, Blocksomand colleagues [20] concluded that rapid resuscitation andsurgical control of haemorrhage is of utmost importance andone of the prognostic determinants in a retrospective study onduodenal injuries. A retrospective study by Ertel and col-leagues [21] that included 80 polytrauma patients in extremisor with persistent haemodynamic instability also favoured earlysurgical intervention to stabilise a pelvic fracture or to surgi-cally control bleeding.

In addition, studies of different trauma systems indirectlyemphasise the importance of minimising the time between ini-tial care and surgery for those with signs of exsanguination orongoing severe haemorrhage. Hill and colleagues [22]observed a significant decrease in mortality from shock byintroducing an educational program on trauma and by estab-lishing a 60-minute emergency department time limit forpatients in a state of haemorrhagic shock. Others also stress

control of haemorrhage in the exsanguinating or the severelybleeding patient [23,24]. In a retrospective review of 537deaths in the operation room, Hoyt and colleagues [25] drewthe conclusion that delayed transfer to the operating room wasa cause of death that could be avoided by shortening the timerequired for diagnosis and resuscitation prior to surgery.

II. Diagnosis and monitoring of bleedingUpon patient arrival in the emergency room, an initial clinicalassessment of the extent of bleeding should be employed toidentify patients at risk of coagulopathy.

Recommendation 2We recommend that the extent of traumatic haemorrhage beclinically assessed using a grading system such as that estab-lished by the American College of Surgeons (ACS) (grade1C).

RationaleAn evaluation of the mechanism of injury (for example, bluntversus penetrating trauma) is a useful tool for determiningwhich patients are candidates for surgical bleeding control.Table 2 summarises the four classes of physiological responseand clinical signs of bleeding as defined by the ACS [26]. Thistype of grading system may be useful in the initial assessmentof bleeding. The initial assessment can also assist indetermining the next patient management goal to minimiseblood loss and achieve haemodynamic stability.

Recommendation 3We do not suggest hyperventilation or the use of excessivepositive end-expiratory pressure (PEEP) when ventilatingseverely hypovolaemic trauma patients (grade 2C).

RationaleThere is a tendency for rescue personnel to hyperventilatepatients during resuscitation [27,28], and hyperventilatedtrauma patients appear to have increased mortality when com-pared with non-hyperventilated patients [28]. The experimen-tal correlates in animals in haemorrhagic shock may be anincreased cardiac output in hypoventilated pigs [29] and adecrease in cardiac output due to 5 cm PEEP in rats [30]. Incontrast, the elimination of PEEP and, to an even greaterextent, negative expiratory pressure ventilation increases car-diac output and survival of rats in haemorrhagic shock [30].

Recommendation 4We recommend that patients presenting with haemorrhagicshock and an identified source of bleeding undergo an imme-diate bleeding control procedure unless initial resuscitationmeasures are successful (grade 1B).

RationalePage 4 of 22(page number not for citation purposes)

the importance of a well-functioning system capable of timely The source of bleeding may be immediately obvious, and pen-etrating injuries are more likely to require surgical bleeding


control. In a retrospective study of 106 abdominal vascularinjuries, all 41 patients arriving in shock following gunshotwounds were candidates for rapid transfer to the operatingtheatre for surgical bleeding control [19]. A similar observationin a study of 271 patients undergoing immediate laparotomyfor gunshot wounds indicates that these wounds combinedwith signs of severe hypovolaemic shock specifically requireearly surgical bleeding control. This observation is true to alesser extent for abdominal stab wounds [31]. Data on injuriescaused by penetrating metal fragments from explosives orgunshot wounds in the Vietnam War confirm the need for earlysurgical control when patients present in shock [18].

In blunt trauma, the mechanism of injury can determine to acertain extent whether the patient in haemorrhagic shock willbe a candidate for surgical bleeding control. Only a few stud-ies address the relationship between the mechanism of injuryand the risk of bleeding, however, and none of these publica-tions is a randomised prospective trial of high evidence. Wehave found no objective data describing the relationshipbetween the risk of bleeding and the mechanism of injury ofskeletal fractures in general or of long-bone fractures inparticular.

Traffic accidents are the leading cause of pelvic injury. Motorvehicle crashes cause approximately 60% of pelvic fracturesfollowed by falls from great height (23%). Most of the remain-der result from motorbike collisions and vehicle-pedestrianaccidents [32,33]. There is a correlation between 'unstable'pelvic fractures and intra-abdominal injuries [32,34]. An asso-ciation between major pelvic fractures and severe head inju-ries, concomitant thoracic, abdominal, urological, and skeletalinjuries is also well described [32]. High-energy injuries pro-duce greater damage to both the pelvis and organs. Patientswith high-energy injuries require more transfusion units, andmore than 75% have associated head, thorax, abdominal, orgenitourinary injuries [35]. It is well documented that 'unstable'

[34], and haemorrhage is the leading cause of death inpatients with major pelvic fractures. Pelvic fractures accountfor 1% to 3% of all skeletal injuries. In patients with multipletrauma, the incidence of pelvic fracture increases to as muchas 25% [33].

Recommendation 5We recommend that patients presenting with haemorrhagicshock and an unidentified source of bleeding undergo imme-diate further assessment (grade 1B).

A patient in haemorrhagic shock with an unidentified source ofbleeding should undergo urgent clinical assessment of chest,abdominal cavity, and pelvic ring stability using focusedabdominal sonography in trauma (FAST) assessment of thoraxand abdomen and/or computerised tomography (CT) exami-nation in the shock room.

SonographyRecommendation 6We recommend early FAST for the detection of free fluid inpatients with suspected torso trauma (grade 1B).

Recommendation 7We recommend that patients with significant free intra-abdominal fluid according to sonography (FAST) and haemo-dynamic instability undergo urgent surgery (grade 1C).

RationaleBlunt abdominal trauma represents a major diagnostic chal-lenge and an important source of internal bleeding. FAST hasbeen established as a rapid and non-invasive diagnosticapproach for detection of intra-abdominal free fluid in theemergency room [36,37]. Large prospective observationalstudies determined a high specificity (range 0.97 to 1.0) anda high accuracy (range 0.92 to 0.99) but low sensitivity (range0.56 to 0.71) of initial FAST examination for detecting intra-

Table 2

American College of Surgeons Advanced Trauma Life Support classification of haemorrhage severity

Haemorrhage severity according to ACS/ATLS classificationa Class I Class II Class III Class IV

Blood loss (ml) 2,000

Pulse rate (per minute) 100 >120 >140

Blood pressure Normal Normal Decreased Decreased

Pulse pressure (mm Hg) Normal Decreased Decreased Decreased

Respiratory rate (per minute) 1420 2030 3040 >40

Urine output (ml/hour) >30 2030 515 Negligible

Central nervous system (mental status) Slightly anxious

Mildly anxious Anxious, confused

Lethargic

aValues are estimated for a 70-kg adult. Table reprinted with permission from the American College of Surgeons [26]. ACS/ATLS, American College of Surgeons/Advanced Trauma Life Support.Page 5 of 22(page number not for citation purposes)

pelvic fractures are associated with massive haemorrhage abdominal injuries in adults and children [38-45]. Shackford


and colleagues [38] assessed the accuracy of FAST per-formed by non-radiologist clinicians (that is, surgeons andemergency physicians who were certified for FAST by definedstandards) for detecting a haemoperitoneum in 241 prospec-tively investigated adult patients with blunt abdominal trauma(except for n = 2 with penetrating injuries) during a four yearperiod. These findings were confirmed by Richards and co-workers [39] in a four year prospective study of 3,264 adultpatients with blunt abdominal trauma. Similar conclusionswere drawn by the same group of investigators in a paediatricpopulation, based on a prospective study on 744 consecutivechildren 16 years old or younger who underwent emergencyFAST for blunt abdominal trauma [40]. Liu and colleagues [41]conducted a one year prospective comparison on the diag-nostic accuracy of CT scan, diagnostic peritoneal lavage(DPL), and sonography in 55 adult patients with blunt abdom-inal trauma. The authors found a high sensitivity (0.92), specif-icity (0.95), and accuracy (0.93) of initial FAST examination forthe detection of haemoperitoneum. Although CT scan andDPL were shown to be more sensitive (1.0 for DPL, 0.97 forCT) than sonography for detection of haemoperitoneum, thesediagnostic modalities are more time-consuming (CT and DPL)and invasive (DPL) [41].

The hypotensive patient (systolic blood pressure below 90mm Hg) presenting free intra-abdominal fluid according toFAST is a potential candidate for early surgery if he or she can-not be stabilised by initiated fluid resuscitation, according to aretrospective study of 138 patients by Farahmand and col-leagues [46]. A similar conclusion can be drawn from a pro-spective blinded study of 400 hypotensive blunt traumavictims (systolic blood pressure below 90 mm Hg) showingthat specific levels of intra-abdominal fluid detected by FASTin these patients was an accurate indicator of the need forurgent surgery [47]. In addition, a retrospective study by Rozy-cki and colleagues [48] of 1,540 patients (1,227 blunt, 313penetrating trauma) assessed with FAST as an early diagnos-tic tool showed that the ultrasound examination had a sensitiv-ity and specificity close to 100% when the patients werehypotensive.

A number of patients who present free intra-abdominal fluidaccording to FAST can safely undergo further investigationwith multi-slice spiral computed tomography (MSCT). Undernormal circumstances, adult patients need to be haemody-namically stable when MSCT is performed outside of theemergency room. In the retrospective study of 1,540 patients(1,227 blunt, 313 penetrating trauma) who were assessedearly with FAST, a successful non-operative management wasachieved in 24 (48%) of the 50 patients who were normoten-sive on admission and had true positive sonographic examina-tions. These results justified an MSCT scan of the abdomenrather than an immediate exploratory laparotomy [48]. In areview article, Lindner and colleagues [49] also concluded

MSCT scanning regardless of the findings from ultrasound orclinical examination.

Computer tomographyRecommendation 8We recommend that haemodynamically stable patients withsuspected head, chest, and/or abdominal bleeding followinghigh-energy injuries undergo further assessment using CT(grade 1C).

RationaleThe increasing role of MSCT in the imaging concept of acutetrauma patients is well documented [50-55]. The integration ofmodern MSCT scanners in the emergency room area allowsthe immediate examination of trauma victims following admis-sion [52,53].

Using modern 16-slice CT scanners, total whole-body scan-ning time amounts to approximately 120 seconds. Sixty-four-slice CT scanners may reduce scanning time to less than 30seconds. In a retrospective study comparing 370 patients intwo groups, Weninger and colleagues [53] showed that thefull extent of injury was definitively diagnosed 12 9 minutesfollowing application of the MSCT protocol. In the group ofconventionally diagnosed patients, definitive diagnosis waspossible after 41 27 minutes. Faster diagnosis led to shorteremergency room and operating room time and shorter inten-sive care unit (ICU) stay [53]. Compared to MSCT, all tradi-tional techniques of diagnostic and imaging evaluation havesome limitations. The diagnostic accuracy, safety, and effec-tiveness of immediate MSCT is dependent on sophisticatedpre-hospital treatment by trained and experienced emergencypersonnel and short transportation times [56,57].

If an MSCT is not available in the emergency room, the realisa-tion of CT scanning implies transportation of the patient to theCT room, and therefore the clinician must evaluate the implica-tions and potential risks and benefits of the procedure.According to established standards, such as those developedby the ACS, only the haemodynamically stable patient shouldbe considered for CT scanning. During transport to the MSCTand imaging, all vital signs should be closely monitored andresuscitation measures continued.

For those patients in whom haemodynamic stability is ques-tionable, imaging techniques such as ultrasound and chestand pelvic radiography may be useful. Peritoneal lavage israrely indicated if ultrasound or CT is available [58]. Transfertimes to and from all forms of diagnostic imaging need to beconsidered carefully in any patient who is haemodynamicallyunstable. In addition to the initial clinical assessment, near-patient testing results, including full blood count, haematocrit(Hct), blood gases, and lactate, should be readily availableunder ideal circumstances.Page 6 of 22(page number not for citation purposes)

that the haemodynamically stable patient should undergo


HaematocritRecommendation 9We do not recommend the use of single Hct measurementsas an isolated laboratory marker for bleeding (grade 1B).

RationaleHct measurements are part of the basic diagnostic work-up fortrauma patients. The diagnostic value of the Hct for detectingtrauma patients with severe injury and occult bleeding sourceshas been a topic of debate in the past decade [59-61]. A majorlimit of the diagnostic value is the confounding influence ofresuscitative measures on the Hct due to administration ofintravenous fluids and red cell concentrates [61-64]. A retro-spective study of 524 trauma patients determined a low sen-sitivity (0.5) of the initial Hct on admission for detecting thosepatients with an extent of traumatic haemorrhage requiring sur-gical intervention [61].

Two prospective observational diagnostic studies determinedthe sensitivity of serial Hct measurements for detectingpatients with severe injury [59,60]. Paradis and colleagues[59] found that the mean change in Hct between arrival and15 minutes and between 15 and 30 minutes was not signifi-cantly different between patients with serious injuries (n = 21)compared to trauma patients without serious injuries (n = 39).Whereas a decrease in Hct of more than or equal to 6.5% at15 and 30 minutes had a high specificity (0.93 to 1.0) for aserious injury, the sensitivity for detecting severely injuredpatients was very low (0.13 to 0.16) [59]. The authors alsofound that a normal Hct on admission did not preclude a sig-nificant injury [59]. Zehtabchi and colleagues [60] expandedthe time window of serial Hct assessments to fourhours afterarrival. All trauma patients requiring a blood transfusion withinthe first fourhours were excluded from the study. In the remain-ing 494 patients, a decrease in Hct of more than 10%between admission and fou hours was highly specific (0.92 to0.96) for severe injury but was associated with a very low sen-sitivity (0.09 to 0.27) for detecting patients with significantinjuries [60]. The limitation of the high specificity of thedecrease in Hct after fourhours in this study is that it includedonly trauma patients who did not receive any blood transfu-sions during the first fourhours [60]. In summary, decreasingserial Hct measurements may reflect continued bleeding, butthe patient with significant bleeding may maintain his or herserial Hct.

Serum lactateRecommendation 10We recommend serum lactate measurement as a sensitivetest to estimate and monitor the extent of bleeding and shock(grade 1B).

RationaleSerum lactate has been used as a diagnostic parameter and

[65]. The amount of lactate produced by anaerobic glycolysisis an indirect marker of oxygen debt, tissue hypoperfusion, andthe severity of haemorrhagic shock [66-69]. Vincent and col-leagues [70] reported on the value of serial lactate measure-ments in predicting survival in a prospective study on aheterogenic group of 27 patients with circulatory shock. Theauthors concluded that changes in lactate concentrations pro-vide an early and objective evaluation of a patient's responseto therapy and suggested that repeated lactate determinationsrepresent a reliable prognostic index for patients with circula-tory shock [70]. Abramson and colleagues [71] performed aprospective observational study on patients with multipletrauma to evaluate the correlation between lactate clearanceand survival. Patients who died within the first 48 hours (n =25) were excluded from the study. The remaining 76 patientswere analysed with respect to the time of serum lactate nor-malisation compared between survivors and non-survivorswho died after 48 hours [71]. Survival was 100% in thosepatients in whom lactate levels returned to the normal range ( 2 mmol/l) within 24 hours. Survival decreased to 77.8% ifnormalisation occurred within 48 hours and to 13.6% in thosepatients in whom lactate levels were elevated above 2 mmol/lfor more than 48 hours [71]. These findings were confirmed ina study on 129 trauma patients by Manikis and colleagues[72]. The authors found that the initial lactate levels werehigher in non-survivors than in survivors and that the prolongedtime for normalisation of lactate levels of more than 24 hourswas associated with the development of post-traumatic organfailure [72]. Together, both the initial serum lactate and seriallactate levels are reliable indicators of morbidity and mortalityfollowing trauma [71,72].

Base deficitRecommendation 11We recommend base deficit as a sensitive test to estimate andmonitor the extent of bleeding and shock (grade 1C).

RationaleBase deficit values derived from arterial blood gas analysisprovide an indirect estimation of global tissue acidosis due toimpaired perfusion [66,68]. Siegel [73] demonstrated that theinitial base deficit represented an independent single predic-tor of post-traumatic mortality in 185 patients with blunt livertrauma. Two large retrospective studies on 3,791 [74] and2,954 [75] trauma patients have strengthened the utility of theinitial base deficit as a sensitive diagnostic marker of thedegree and duration of inadequate perfusion and as a prog-nostic parameter for post-traumatic complications and death.Davis and colleagues [75] stratified the extent of base deficitinto three categories: mild (-3 to -5 mEq/l), moderate (-6 to -9mEq/l), and severe (less than -10 mEq/l). Based on this strati-fication, they established a significant correlation between theadmission base deficit and transfusion requirements within thefirst 24 hours and the risk of post-traumatic organ failure orPage 7 of 22(page number not for citation purposes)

prognostic marker of haemorrhagic shock since the 1960s death [75]. In a different retrospective study, the same group


of authors showed that the base deficit is a better prognosticmarker of death than the pH in arterial blood gas analyses [76].Furthermore, the base deficit was shown to represent a highlysensitive marker for the severity of injury and the incidence ofpost-traumatic death, particularly in trauma patients older than55 years of age [77]. In paediatric patients, admission basedeficit was also shown to correlate significantly with the extentof post-traumatic shock and mortality, as determined in a ret-rospective study which included 65 critically injured childrenand used a cutoff value of less than -5 mEq/l [78]. However, incontrast to the data on lactate levels in haemorrhagic shock,reliable large-scale prospective studies on the correlationbetween base deficit and outcome are still lacking.

Although both the base deficit and serum lactate levels arewell correlated with shock and resuscitation, these two param-eters do not strictly correlate with each other in severelyinjured patients [79]. Therefore, the independent assessmentof both parameters is recommended for the evaluation ofshock in trauma patients [66,68,79,80]. Composite scoresthat assess the likelihood of massive transfusion and thatinclude base deficit and other clinical parameters have beendeveloped but require further validation [80,81].

III. Rapid control of bleedingRecommendation 12We recommend that patients with pelvic ring disruption inhaemorrhagic shock undergo immediate pelvic ring closureand stabilisation (grade 1B).

Recommendation 13We recommend that patients with ongoing haemodynamicinstability despite adequate pelvic ring stabilisation receiveearly angiographic embolisation or surgical bleeding control,including packing (grade 1B).

RationaleMarkers of pelvic haemorrhage include anterior-posterior andvertical shear deformations, CT 'blush' (active arterial extrava-sation), bladder compression pressure, pelvic haematoma vol-umes greater than 500 ml evident by CT, and ongoinghaemodynamic instability despite adequate fracture stabilisa-tion [82-85]. Initial therapy of pelvic fractures includes controlof venous and/or canellous bone bleeding by pelvic closure[86]. Some institutions use primarily external fixators to controlhaemorrhage from pelvic fractures [82], but pelvic closure mayalso be achieved using a bed sheet, pelvic binder, or a pelvicC-clamp [86-90]. Although arterial haemorrhage from pelvicfractures may be lethal, venous bleeding may be equally dev-astating. Arterial embolisation appears to achieve its effect bycontrolling the arterial bleeding and allowing the tamponadeeffect of the haematoma to control venous bleeding [91,92].

Results of surgery to control pelvic haemorrhage via laparot-

collateral circulation. However, in suboptimal situations (forexample, when embolisation is not possible), extraperitonealpacking of the pelvis may reduce the loss of blood. Extraperi-toneal haemorrhage in patients with haemorrhagic shock andpelvic ring disruption may be attributed to ruptured veins, frac-ture surfaces, and/or arterial sources. The overall mortality rateof patients with severe pelvic ring disruptions and haemody-namic instability remains as high as 30% to 45% [93].Angioembolisation is often applied in patients with ongoinghaemodynamic instability despite adequate fracture stabilisa-tion and the exclusion of extra-pelvic sources of haemorrhage.Repeat angiography may be of value in those selected patients[86]. Patients who require embolisation tend to be older, havea higher injury severity score, and are more likely to be coagu-lopathic and haemodynamically unstable than patients whonot require embolisation [94].

Recommendation 14We recommend that early bleeding control be achieved bypacking, direct surgical bleeding control, and the use of localhaemostatic procedures. In the exsanguinating patient, aorticcross-clamping may be employed as an adjunct to achievebleeding control (grade 1C).

RationaleThe choice of thoracic or abdominal aortic clamping should bedetermined according to the site of bleeding, available surgicalskill, and speed. The patient in haemorrhagic shock in whomimmediate aortic cross-clamping is warranted is characterisedby an injury to the torso and the severity of the blood loss andshock. The hypotensive state will not respond to the intrave-nous resuscitation and may lead to cardiac arrest. The causeof injury is predominantly penetrating (for example, a gunshotwound or a stab wound). Depending on the cause of injury, themortality rate in these situations is extremely high [18,19,95].However, when the source of bleeding is intra-abdominal, tho-racic aortic clamping combined with other measures for haem-orrhage control can be life-salvaging in nearly one third ofpatients, according to Millikan and Moore [96] and Cothrenand Moore [97]. It is unclear whether the thoracic aorticclamping should be performed before or after the abdominalincision [98]. No study has compared thoracic aortic clampingabove the diaphragm with abdominal aortic clamping justbelow the diaphragm, although the latter method is favouredby some surgeons [98].

The cross-clamping of the aorta should be considered as anadjunct to other initial haemorrhage control measures such asthe evacuation of blood, direct surgical bleeding control, orpacking of bleeding sources [99]. When aortic clamping isdeemed necessary due to continuous bleeding or low bloodpressure, the prognosis is generally poor [100].Page 8 of 22(page number not for citation purposes)

omy have remained poor due to the existence of an extensive


Recommendation 15We recommend that damage control surgery be employed inthe severely injured patient presenting with deep haemor-rhagic shock, signs of ongoing bleeding, and coagulopathy.Additional factors that should trigger a damage controlapproach are hypothermia, acidosis, inaccessible major ana-tomic injury, a need for time-consuming procedures, or con-comitant major injury outside the abdomen (grade 1C).

RationaleThe severely injured patient arriving to the hospital with contin-uous bleeding or deep haemorrhagic shock generally has apoor chance of survival unless early control of bleeding, properresuscitation, and blood transfusion are achieved. This is par-ticularly true for patients who present with uncontrolled bleed-ing due to multiple penetrating injuries as well as patients withmultiple injuries and unstable pelvic fractures with ongoingbleeding from fracture sites and retroperitoneal vessels. Thecommon denominator in these patients is the exhaustion ofphysiological reserves with resulting profound acidosis, hypo-thermia, and coagulopathy. In the trauma community, this isalso called the 'bloody vicious cycle' or the 'lethal triad.' In1983, Stone and colleagues [101] described the techniquesof abbreviated laparotomy, packing to control haemorrhageand of deferred definitive surgical repair until coagulation hadbeen established. Since then, a number of authors havedescribed the beneficial results of this concept, which is nowcalled 'damage control' [31,33,87,90,101-104]. Damage con-trol consists of three components. The first component is anabbreviated resuscitative laparotomy for control of bleeding,the restitution of blood flow where necessary, and the controlof contamination. This should be achieved as quickly as possi-ble without spending unnecessary time on traditional organrepairs that can be deferred to a later phase. The abdomen ispacked and temporary abdominal closure is performed. Thesecond component is intensive care treatment, focused oncore rewarming, correction of the acid-base imbalance, andcoagulopathy as well as optimising the ventilation and thehaemodynamic status. Further diagnostic investigations arealso frequently performed during this phase. The third compo-nent is the definitive surgical repair that is performed onlywhen target parameters have been achieved [99,105-107].

Despite the lack of controlled randomised studies comparingdamage control to traditional surgical management, a retro-spective review by Stone and colleagues [101] presents datain favour of damage control for the severely injured patient pre-senting signs of coagulopathy during surgery. Rotondo andcolleagues [102] found similar results in a subgroup ofpatients with major vascular injury and two or more visceralinjuries, and Carrillo and colleagues [103] demonstrated thebenefit of damage control in patients with iliac vessel injury. Inaddition, a cumulative review of 961 patients treated with dam-age control reported overall mortality and morbidity rates of

IV. Tissue oxygenation, type of fluid, and hypothermiaRecommendation 16We suggest a target systolic blood pressure of 80 to 100 mmHg until major bleeding has been stopped in the initial phasefollowing trauma without brain injury (grade 2C).

RationaleTo maintain tissue oxygenation, traditional treatment of traumapatients uses early and aggressive fluid administration torestore blood volume. However, this approach may increasethe hydrostatic pressure on the wound and cause a dislodge-ment of blood clots, a dilution of coagulation factors, andundesirable cooling of the patient. The concept of low-volumefluid resuscitation, so-called 'permissive hypotension,' avoidsthe adverse effects of early aggressive resuscitation whilemaintaining a level of tissue perfusion that, although lower thannormal, is adequate for short periods [108]. Its general effec-tiveness remains to be confirmed in randomised clinical trials,but studies have demonstrated increased survival when a low-volume fluid resuscitation concept was used in penetratingtrauma [109,110]. In contrast, no significant difference wasfound in patients with blunt trauma [111]. One study con-cluded that mortality was higher after on-site resuscitationcompared with in-hospital resuscitation [112]. It seems thatgreater increases in blood pressure are tolerated withoutexacerbating haemorrhage when they are achieved graduallyand with a significant delay following the initial injury [113]. Allthe same, a recent Cochrane systematic review concludedthat there is no evidence from randomised clinical trials for oragainst early or larger volumes of intravenous fluids in uncon-trolled haemorrhage [114]. The low-volume approach is con-traindicated in traumatic brain injury and spinal injuriesbecause an adequate perfusion pressure is crucial to ensuretissue oxygenation of the injured central nervous system. Inaddition, the concept of permissive hypotension should beconsidered carefully in the elderly patient and may be contrain-dicated if the patient suffers from chronic arterial hypertension.

Red blood cell (RBC) transfusion enables the maintenance ofoxygen transport in some patients. Early signs of inadequatecirculation are relative tachycardia, relative hypotension, oxy-gen extraction greater than 50%, and PvO2(mixed venous oxy-gen pressure) of less than 32 mm Hg [115-117]. The depth ofshock, haemdoynamic response to resuscitation, and the rateof actual blood loss in the acutely bleeding and haemodynam-ically unstable patient may also be integrated into the indica-tion for RBC transfusion. In general, RBC transfusion isrecommended to maintain haemoglobin (Hb) between 7 and9 g/dl [118].

Recommendation 17We suggest that crystalloids be applied initially to treat thebleeding trauma patient. Colloids may be added within theprescribed limits for each solution (grade 2C).Page 9 of 22(page number not for citation purposes)

52% and 40%, respectively [106].


RationaleIt is still unclear which type of fluid should be employed in theinitial treatment of the bleeding trauma patient. Although sev-eral meta-analyses have shown an increased risk of death inpatients treated with colloids compared with patients treatedwith crystalloids [119-123] and three of these studies showedthat the effect was particularly significant in a trauma subgroup[119,122,123], a more recent meta-analysis showed no differ-ence in mortality between colloids and crystalloids [124].Problems in evaluating and comparing the use of differentresuscitation fluids include the heterogeneity of populationsand therapy strategies, limited quality of analysed studies, mor-tality not always being the primary outcome, and different(often short) observation periods. It is therefore difficult toreach a definitive conclusion as to the advantage of one typeof resuscitation fluid over the other. The SAFE (Saline versusAlbumin Fluid Evaluation) study compared 4% albumin with0.9% sodium chloride in 6,997 ICU patients and showed thatalbumin administration was not associated with worse out-comes; however, there was a trend toward higher mortality inthe trauma subgroup that received albumin (p = 0.06) [125].Promising results have been obtained with hypertonic solu-tions. One study showed that use of hypertonic saline wasassociated with lower intracranial pressure than with normalsaline in brain-injured patients [126], and a meta-analysis com-paring hypertonic saline dextran with normal saline for resusci-tation in hypotension from penetrating torso injuries showedimproved survival in the hypertonic saline dextran group whensurgery was required [127]. A clinical trial with brain injurypatients found that hypertonic saline reduced intracranial pres-sure more effectively than dextran solution with 20% mannitol[128]. However, Cooper and colleagues [129] found almostno difference in neurological function six months after trau-matic brain injury in patients who had received pre-hospitalhypertonic saline resuscitation compared to conventional fluid.

Recommendation 18We recommend early application of measures to reduce heatloss and warm the hypothermic patient in order to achieve andmaintain normothermia (grade 1C).

RationaleHypothermia, defined as a core body temperature of less than35C, is associated with acidosis, hypotension, and coagulop-athy in severely injured patients. In a retrospective study with122 patients, hypothermia was an ominous clinical sign,accompanied by high mortality and blood loss [130]. The pro-found clinical effects of hypothermia ultimately lead to highermorbidity and mortality, and hypothermic patients require moreblood products [131].

Hypothermia is associated with an increased risk of severebleeding, and hypothermia in trauma patients represents anindependent risk factor for bleeding and death [132]. The

impaired coagulation factor function (a 1C decrease in tem-perature is associated with a 10% decrease in function),enzyme inhibition, and fibrinolysis [133,134]. Body tempera-tures below 34C compromise blood coagulation, but this hasbeen observed only when coagulation tests, prothrombin time[PT] and activated partial thromboplastin time [aPTT] are car-ried out at the low temperatures observed in patients withhypothermia and not when assessed at 37C, the temperaturetypically used for such tests. Steps to prevent hypothermiaand the risk of hypothermia-induced coagulopathy includeremoving wet clothing, covering the patient to avoid additionalheat loss, increasing the ambient temperature, forced airwarming, warm fluid therapy, and (in extreme cases) extracor-poreal re-warming devices [135,136].

Animal and human studies of controlled hypothermia in haem-orrhage have shown some positive results compared with nor-mothermia [137,138]. In 2003, McIntyre and colleagues [139]published a meta-analysis showing a beneficial effect on mor-tality rates and neurological outcome when using mild hypo-thermia in traumatic brain injury. In contrast, in 2004, onemeta-analysis analysed the effect of hypothermia in traumaticbrain injury using the results of eight studies with predefinedcriteria for RCTs; no reduction in mortality rates and only aslight benefit in neurological outcome could be demonstrated[140]. These contradictory results may be due to the differentexclusion and inclusion criteria for the studies used for theanalysis. Henderson and colleagues [140] included two stud-ies in which patients without increased intracranial pressurewere enrolled. Had these two studies been excluded from themeta-analysis, a benefit with respect to improved neurologicaloutcome might have been demonstrated [141]. Moreover, thestudies included differed with respect to the speed of induc-tion and duration of hypothermia, which may be very importantfactors influencing the benefit of this treatment.

If mild hypothermia is applied in traumatic brain injury, coolingshould take place within the first 3 hours following injury andbe maintained for approximately 48 hours, rewarming shouldlast 24 hours, and the cerebral perfusion pressure should bemaintained above 50 mm Hg (70 mm Hg). Patients most likelyto benefit from hypothermia are those with a Glasgow ComaScale of between 4 and 7 at admission [142]. Possible sideeffects are hypotension, hypovolaemia, electrolyte disorders,insulin resistance, reduced insulin secretion, and increasedrisk of infection [143]. Further studies are warranted to inves-tigate the postulated benefit of hypothermia in traumatic braininjury, taking these important factors into account.

V. Management of bleeding and coagulationRBCs, fresh frozen plasma, and plateletsRecommendation 19We recommend a target Hb of 7 to 9 g/dl (grade 1C).Page 10 of 22(page number not for citation purposes)

effects of hypothermia include altered platelet function,


RationaleThere is experimental evidence that erythrocytes are involvedin the biochemical and functional responsiveness of activatedplatelets, suggesting that erythrocytes contribute to haemos-tasis. In addition to the rheological effect on the margination ofplatelets, red cells support thrombin generation [144]. How-ever, the optimal Hct or Hb concentration required to sustainhaemostasis in massively bleeding patients is unclear. Furtherinvestigations into the role of the Hb concentration on hae-mostasis in massively transfused patients are thereforewarranted.

The specific effect of the Hct on blood coagulation is largelyunknown [145]. An acute reduction of the Hct may result in anincrease in the bleeding time [146,147] with restoration uponre-transfusion [146]. This may be related to the presence ofthe enzyme elastase on the surface of RBC membranes, whichmay activate coagulation factor IX, thereby triggering bloodcoagulation [148,149]. However, a moderate reduction of theHct does not increase blood loss from a standard spleen injury[147], and an isolated in vitro reduction of the Hct did notcompromise blood coagulation as assessed by thromboelas-tography [150].

No prospective randomised trial has compared restrictive andliberal transfusion regimens in trauma, but 203 trauma patientsfrom the Transfusion Requirements in Critical Care (TRICC)trial [151] were re-analysed [118]. A restrictive transfusionregimen (Hb transfusion trigger less than 7.0 g/dl) resulted infewer transfusions as compared with the liberal transfusionregimen (Hb transfusion trigger less than 10 g/dl) andappeared to be safe. However, no statistically significant ben-efit in terms of multiple organ failure or post-traumatic infec-tions was observed. It should be emphasised that this studywas neither designed nor powered to answer these questionswith precision. In addition, it cannot be ruled out that thenumber of RBC units transfused reflects merely the severity ofinjury. Therefore, the observed correlation between numbersof RBC units transfused and multiple organ failure [152] mayreflect a correlation between the severity of injury and multipleorgan failure. Adequately powered studies similar to theTRICC trial are therefore urgently needed in post-traumaticpatients.

Despite the lack of high-level scientific evidence for a specificHb transfusion trigger in patients with traumatic brain injury,these patients are currently transfused in many centres toachieve an Hb of approximately 10 g/dl [153]. This may be jus-tified by the recent finding that increasing the Hb from 8.7 to10.2 g/dl improved local cerebral oxygenation [154]. Itremains unclear, however, whether this practice will result inan improved neurological outcome. Although the lowest Hctwas correlated with adverse neurological outcome, RBCtransfusions were equally found to be an independent factor

study [155]. Interestingly, the number of days with an Hctbelow 30% was found to be correlated with an improved neu-rological outcome. Therefore, the authors suggest thatpatients with severe traumatic brain injury should not have anHb transfusion threshold different than that of other critically illpatients [155].

Recommendation 20We recommend treatment with thawed fresh frozen plasma(FFP) in patients with massive bleeding or significant bleedingcomplicated by coagulopathy (PT or aPTT more than 1.5 timescontrol). The initial recommended dose is 10 to 15 ml/kg, butfurther doses may be required (grade 1C).

RationaleThe clinical efficacy of FFP is largely unproven [156]. Never-theless, most guidelines recommend the use of FFP either inmassive bleeding or in significant bleeding complicated bycoagulopathy (PT or aPTT more than 1.5 times control)[7,157,158]. Patients treated with oral anticoagulants (vitaminK antagonists) present a particular challenge, and thawed FFPis recommended [158] only when prothrombin complex con-centrate (PCC) is not available [157]. The most frequently rec-ommended dose is 10 to 15 ml/kg [157,158], but furtherdoses may be required [159].

As with all products derived from human blood, the risks asso-ciated with FFP treatment include circulatory overload, ABOincompatibility, transmission of infectious diseases (includingthe prion diseases), mild allergic reactions, and (particularly)transfusion-related acute lung injury (TRALI) [157,160,161].FFP and platelet concentrates appear to be the most fre-quently implicated blood products in TRALI [160-163].Although the formal link between the administration of FFP,control of bleeding, and an eventual improvement in the out-come of bleeding patients is lacking, most experts wouldagree that FFP treatment is beneficial in patients with massivebleeding or significant bleeding complicated by coagulopathy.

Recommendation 21We recommend that platelets be administered to maintain aplatelet count above 50 109/l (grade 1C). We suggest main-tenance of a platelet count above 100 109/l in patients withmultiple trauma who are severely bleeding or have traumaticbrain injury (grade 2C). We suggest an initial dose of 4 to 8platelet concentrates or one aphaeresis pack (grade 2C).

RationaleIn medical conditions leading to thrombocytopaenia, haemor-rhage does not often occur until the platelet count falls to lessthan 50 109/l, and platelet function decreases exponentiallybelow this point [164-167]. There is no direct evidence to sup-port a particular platelet transfusion threshold in the traumapatient. A consensus development conference sponsored byPage 11 of 22(page number not for citation purposes)

for adverse neurological outcome in a recent retrospective the National Institutes of Health (NIH) (Bethesda, MD, USA) in


1986 determined that bleeding is unlikely to be caused bythrombocytopaenia at platelet counts of 50 109/l or greaterand agreed that platelet transfusion is appropriate to preventor control bleeding associated with deficiencies in plateletnumber or function [168,169]. The NIH consensus did notconsider trauma, but it seems reasonable to recommend thata platelet count of at least 50 109/l be maintained followinginjury.

An argument can be made for maintaining a higher level ofplatelets, perhaps up to 100 109/l, following injury. If apatient has increased fibrin degradation products (for exam-ple, in patients with massive bleeding), disseminated intravas-cular coagulation, or hyperfibrinolysis, this will interfere withplatelet function and a higher threshold of 75 109/l has beensuggested by consensus groups [170,171]. Transfusionthreshold levels of up to 100 109/l have been suggested fortreatment of severe brain injury and massive haemorrhage, butthe evidence for the higher threshold is weak [170,171].

When platelet transfusion was introduced in the 1950s, noclinical trials were employed to assess the utility of platelettherapy compared to placebo, and such trials today might beconsidered unethical. The appropriate dose of platelets istherefore uncertain. Platelet concentrate produced from a unitof whole blood contains 7.5 1010 platelets on average andshould increase the platelet count by 5 to 10 109/l in a 70-kg recipient. Aphaeresis platelet concentrates generally con-tain approximately 3 to 6 1011 platelets, depending on localcollection practice, and physicians should be cognizant of thedoses provided locally. A pool of 4 to 8 platelet concentratesor a single-donor aphaeresis unit is usually sufficient to providehaemostasis in a thrombocytopaenic, bleeding patient.

If required, the dose of platelets ( 109) can be calculated inmore detail from the desired platelet increment, the patient'sblood volume in litres (estimated by multiplying the patient'sbody surface area by 2.5, or 70 ml/kg in an adult), and a cor-rection factor of 0.67 to allow for pooling of approximately33% of transfused platelets in the spleen.

Recommendation 22We recommend treatment with fibrinogen concentrate or cry-oprecipitate if significant bleeding is accompanied by aplasma fibrinogen level of less than 1 g/l. We suggest an initialfibrinogen concentrate dose of 3 to 4 g or 50 mg/kg of cryo-precipitate, which is approximately equivalent to 15 to 20 unitsin a 70-kg adult. Repeat doses should be guided by laboratoryassessment of fibrinogen levels (grade 1C).

RationaleCryoprecipitate or fibrinogen is used for the correction of bothinherited and acquired hypofibrinogenaemia. Their use isbased on the assumptions that low fibrinogen levels are asso-

higher levels of fibrinogen decreases that risk. The evidencefor the clinical efficacy of cryoprecipitate and fibrinogen intrauma patients is limited; no clinical randomised studies havebeen performed to determine whether the administration ofcryoprecipitate or fibrinogen improves clinical outcome inseverely bleeding trauma patients. Only indirect observationalstudies are available, but this evidence suggests that clinicallysignificant bleeding decreases in a variety of clinical scenariosfollowing treatment with both agents. Hypofibrinogenaemiaresponds well to treatment with cryoprecipitate concentrate[172]. Administration of fibrinogen was associated with bleed-ing control in patients with generalised, mostly traumatic,bleeding [173]. Administration of 4 g of fibrinogen raisedfibrinogen levels from 0.1 to 1 g/l, and bleeding control wasachieved in patients with bleeding associated with uterine rup-ture and abortion [174]. A few observational studies report thesuccessful use of fibrinogen in patients with congenital afibrin-ogenaemia [175-177]. The optimal initial dose has not beendefined, and regional differences in cryoprecipitate and fibrin-ogen preparations exist, but available evidence suggests thatan initial dose of cryoprecipitate or fibrinogen that raises fibrin-ogen plasma level above 1 g/l will provide sufficient haemos-tasis [174,176,178].

There are no specific risks related to administration of fibrino-gen or cryoprecipitate other than the risks associated withother blood components and the increased risk associatedwith pooled versus single-donor blood products. Fibrinogen orcryoprecipitate can have unpredictable adverse effects. Ofparticular concern are allergic reactions and anaphylaxis.There are no reported specific adverse events related toadministration of fibrinogen or cryoprecipitate in patients withhypofibrinogenaemia.

Pharmacological agentsA large body of evidence supports the use of antifibrinolyticagents for the management of bleeding in elective surgery andcardiac surgery patients. For the purpose of these guidelines,we have assumed that these effects are transferable to traumapatients, and our recommendation is based upon thisunproven assumption.

Recommendation 23We suggest that antifibrinolytic agents be considered in thetreatment of the bleeding trauma patient. Suggested dosagesare tranexamic acid (trans-4-aminomethylcyclohexane-1-car-boxylic acid) 10 to 15 mg/kg followed by an infusion of 1 to 5mg/kg per hour, -aminocaproic acid 100 to 150 mg/kg fol-lowed by 15 mg/kg per hour, or (after a test dose) aprotinin 2million kallikrein inhibitory units (KIU) immediately followed by500,000 KIU/hour in an intravenous infusion. Antifibrinolytictherapy should be stopped once bleeding has been ade-quately controlled (grade 2C).Page 12 of 22(page number not for citation purposes)

ciated with a risk of bleeding and that the achievement of


RationaleTranexamic acid is a synthetic lysine analogue that is a com-petitive inhibitor of plasmin and plasminogen. Tranexamic acidis distributed throughout all tissues and the plasma half-life is120 minutes. There is large variation in the dose employed. Invitro studies have suggested that a dose of 10 g/ml isrequired to inhibit fibrinolysis [179]. Studies of plasma levels[180] confirmed that the Horrow regimen (10 mg/kg followedby 1 mg/kg per hour) [181], shown to reduce blood loss in car-diac surgery, attained these levels. Other studies have usedboluses of up to 5 g per patient with no ill effect [182].

-Aminocaproic acid is also a synthetic lysine analogue thathas a potency 10-fold weaker than that of tranexamic acid. Itis therefore administered in a loading dose of 150 mg/kg fol-lowed by a continuous infusion of 15 mg/hour. The initial elim-ination half-life is 60 to 75 minutes and it must therefore beadministered by continuous infusion in order to maintain ther-apeutic drug levels until the bleeding risk has diminished.

Aprotinin is a broad-spectrum serine protease inhibitor iso-lated from bovine lung and forms irreversible inhibitory com-plexes with a number of serine proteases. In particular, it is apowerful antiplasmin agent, and the initial elimination of apro-tinin is 1.5 to 2 hours [183]. The 'high-dose' regimen [184] (2MKIU to patient and cardiopulmonary bypass prime and aninfusion of 500,000 KIU/hour) has been shown to reduce peri-operative bleeding in open cardiac surgery. However, lowerdoses do produce adequate antiplasmin effects. A dose of 2M units is approved for the treatment of hyperfibrinolysis.

The clear efficacy of antifibrinolytic agents in elective surgeryand especially in cardiac surgery has been shown in numerousclinical trials [184,185]. A larger number of trials to evaluatethe efficacy of aprotinin have been published than assess-ments of lysine analogue efficacy. It may be possible to extrap-olate the benefits of antifibrinolytic agents to bleedingsecondary to trauma, but this assumption is not backed by anypublished data that suggest that the haemostatic response totrauma is similar to the haemostatic response to elective sur-gery. There is insufficient evidence from RCTs of antifibrino-lytic agents in trauma patients to either support or refute aclinically important treatment effect. Further RCTs of antifibri-nolytic agents in trauma patients are required [186]. The effi-cacy of tranexamic acid in trauma will be assessed by theongoing CRASH (Clinical Randomisation of an Antifibrinolyticin Significant Haemorrhage) II study, in which 20,000 traumapatients worldwide are being randomly assigned to 1 g of tran-examic acid for a period of 10 minutes followed by 1 g infusedfor a period of 8 hours [187].

The risk of precipitated thrombosis with the use of antifibrino-lytic agents has been of major theoretical concern; however,the Cochrane review of antifibrinolytics cites studies that

increased risk of either arterial or venous thrombotic events[188]. All antifibrinolytics are renally excreted and accumulatein individuals with renal failure, and therefore dosage shouldbe reduced in patients with renal failure. In practice, milddegrees of renal failure do not seem to affect outcome.

Because aprotinin is a bovine protein with an associated riskof anaphylaxis, a test dose must be given. After high-doseaprotinin, as many as 50% of patients develop specific immu-noglobulin G antibodies within three months of exposure. Themanufacturer (Bayer Pharmaceuticals Corporation, WestHaven, CT, USA) estimates a 0.5% overall risk of anaphylacticreactions following aprotinin treatment, which may increase to6% to 9% following re-exposure [183].

An open study by Mangano and colleagues [189] suggestedthat aprotinin usage in cardiac surgery was associated with anincreased risk of myocardial infarction, stroke, and renal failure.A further publication cited an increased risk of renal problemsin patients receiving aprotinin compared to tranexamic acid[190]. Because the study of Mangano and colleagues [189]was open, it remains unclear whether sicker patients in thestudy may have preferentially received aprotinin. A blindedcomparative study of aprotinin versus tranexamic acid versus-aminocaproic acid [191] which aims to recruit 3,000patients will assess safety and efficacy issues in cardiac sur-gery. At present, in light of the current US Food and DrugAdministration warning against the use of aprotinin [192], thegreater cost associated with aprotinin use, and the need togive a test dose (often impractical in an emergency situation),we favour the use of tranexamic acid or -aminocaproic acid intrauma patients.

Factor replacementRecommendation 24We suggest that the use of recombinant activated coagulationfactor VII (rFVIIa) be considered if major bleeding in blunttrauma persists despite standard attempts to control bleedingand best-practice use of blood components. We suggest aninitial dose of 200 g/kg followed by two doses of 100 g/kgadministered at 1 and 3 hours following the first dose (grade2C).

RationalerFVIIa is not a first-line treatment for bleeding and will be effec-tive only once sources of major bleeding have been controlled.Once major bleeding from damaged vessels has beenstopped, rFVIIa may be helpful to induce coagulation in areasof diffuse small vessel coagulopathic bleeding. rFVIIa shouldbe considered only if first-line treatment with a combination ofsurgical approaches, best-practice use of blood products(RBCs, platelets, FFP, and cryoprecipitate/fibrinogen resultingin Hctabove 24%, plateletsabove 50,000 109/l, and fibrino-genabove 0.5 to 1.0 g/l) and correction of severe acidosis,Page 13 of 22(page number not for citation purposes)

included more than 8,000 patients and demonstrated no severe hypothermia, and hypocalcaemia (resulting in pHabove


7.20, temperatureabove 32C, and ionised Ca++above 0.8mmol/l, respectively) fail to control bleeding. Because rFVIIaacts on the patient's own clotting system, a sufficient numberof platelets are needed to allow a thrombin burst to be inducedby the pharmacological, supraphysiological doses of rFVIIathrough direct binding to activated platelets [193,194].Reduction in platelet count may lead to impaired thrombingeneration [195]. Moreover, fibrinogen is required to ensureformation of a stable clot [158,196]. A recent study showedthat a pH below 7.20 substantially reduced rFVIIa activity butthat a temperature above 32C only slightly improved rFVIIaactivity [197]. Independent of rFVIIa activity, however, pH andbody temperature should be restored as near to physiologicallevels as possible since even small reductions in pH and tem-perature may result in slower coagulation enzyme kinetics[133,134,198]. Moreover, hypocalcaemia is frequentlypresent in severely injured patients [199] and may require theadministration of intravenous calcium with frequent ionisedserum calcium measurement [200].

A number of case studies and case series have reported thattreatment with rFVIIa can be beneficial in the treatment ofcoagulopathic bleeding following trauma [201-204]. Arecently published multi-centre, randomised, double-blind, pla-cebo-controlled study examined the efficacy of rFVIIa inpatients with blunt or penetrating trauma [205]. Patients wererandomly assigned to receive either three doses of rFVIIa(200, 100, and 100 g/kg) or placebo after they had received6 units of RBCs. The first dose of their assigned medicationwas administered after transfusion of a further 2 units of RBCs(8 units in total), and a second and third dose were adminis-tered 1 and 3 hours later. Treatment with rFVIIa in blunt traumaproduced a significant reduction in RBC transfusion require-ments and the need for massive transfusions (>20 units ofRBCs) in patients with blunt trauma surviving for more than 48hours and also significantly reduced the incidence of acuterespiratory distress syndrome in all patients with blunt trauma.In contrast, no significant effects were observed on RBCtransfusion requirements in the penetrating trauma patients inthis study, although trends toward reduced RBC requirementsand fewer massive transfusions were observed. Therefore, norecommendation to use the drug in this group can be made.

The required dose(s) of rFVIIa is still under debate. Whereasthe above dosing recommendation is based on the only pub-lished RCT available in trauma patients and is also recom-mended by a group of European experts [206], Israeliguidelines based on findings from a case series of 36 patientswho received rFVIIa on a compassionate-use basis in Israel[201] propose an initial dose of 120 g/kg (between 100 and140 g/kg) and (if required) a second and third dose. Furthersupport for the dose regimen recommended here comes frompharmacokinetic modelling techniques, which have shownthat the dose regimen for rFVIIa treatment used in the above-

drug to support haemostasis [207]. If rFVIIa is administered,the patient's next of kin should be informed that rFVIIa is beingused outside the currently approved indications (off-label use),especially since the use of rFVIIa may increase the risk ofthromboembolic complications [208].

Recommendation 25We recommend the use of PCC according to the manufac-turer's instructions only for the emergency reversal of vitaminK-dependent oral anticoagulants (grade 1C).

RationaleDespite the common use of PCC, there is no clear indicationfor its use in bleeding non-haemophilia patients. The evidenceof clinical efficacy of PCC in patients without haemophilia islimited, and no clinical randomised studies have been per-formed to determine whether administration of PCC improvesclinical outcome in severely bleeding trauma patients. PCChas been used to control bleeding in haemophilia patients[209-211] or to reverse the effect of oral anticoagulant agents[212,213]. The American Society of Anesthesiology recom-mends the use of PCC in patients with clinical coagulopathyand prolonged PT more than 1.5 times normal [158]. Becausethere are variations in the production of PCC, the dosageshould be determined according to the instructions of the indi-vidual manufacturer [214].

Administration of PCC may carry the risk of venous and arterialthrombosis or disseminated intravascular coagulation[215,216]; however, the type of surgery has no influence onthe type and severity of these complications [217]. Decreasedclearance of activated clotting factor complexes increases thelikelihood of these complications in patients with liver disease[218].

Recommendation 26We do not recommend the use of antithrombin III in the treat-ment of the bleeding trauma patient (grade 1C).

Antithrombin concentrates are indicated in inherited andacquired antithrombin deficiency. Although antithrombin defi-ciency does occur in consumptive coagulopathy, this is not anisolated condition; all coagulation factors and physiologicalanticoagulants undergo consumption under these circum-stances. The best replacement therapy is FFP. Clinical studiesof antithrombin concentrate in severe blunt trauma and in crit-ical care have shown no benefit [219,220].

DiscussionThese guidelines for the management of the bleeding traumapatient are based on a critical appraisal of the published liter-ature and were formulated according to a consensus reachedby the author group and the professional societies involved.We have attempted in an evidence-based manner to addressPage 14 of 22(page number not for citation purposes)

cited RCT is capable of providing adequate plasma levels of a number of critical issues faced by the treating physician con-


fronted with a critically bleeding patient. Figure 1 graphicallysummarises the recommendations included in this guideline.Unfortunately in emergency medicine, a number of pivotalissues have not been, and due to ethical and practical consid-erations may never be, addressed in randomised clinical trials.This reality renders the need for best-practice guidelines evenmore acute.

Although the emphasis in these guidelines has been on themanagement of critical bleeding in the trauma patient, the highrisk of venous thromboembolism in trauma patients shouldalways be kept in mind and thromboprophylactic treatmentconsidered once bleeding has been controlled [221]. In addi-tion, a conscious decision was made to exclude animal studiesfrom the literature reviewed for the development of theseguidelines. Because no animal model has accurately mimickedthe human coagulation system, RCTs in humans provide thestrongest evidence for the management of bleeding in

We have made an effort to consider a number of specificpatient subgroups that may require treatment that has beenadapted to their physiological condition. There is very littleevidence, however, to support specific recommendations forthese special patient groups. The physiology of elderlypatients with respect to coagulopathy is probably not differentthan that of younger adults; however, the bleeding patient whohas been treated with an anticoagulant or an antiplatelet agentmay present with a greater risk of coagulopathic bleeding. Wehave also considered the management of bleeding followinginjury in children. A more conservative approach to the surgicalmanagement of bleeding in children is generally taken, andthere is some evidence that commonly quoted age-relatedphysiological norms are not applicable to the injured child.There is little evidence, however, for specific differences inbleeding and coagulation management in children; therefore,we suggest that these guidelines be applied to both adultsand children until research data that are more specific become

Figure 1

Flowchart of treatment aspects for the bleeding trauma patient which are discussed in this guidelineFlowchart of treatment aspects for the bleeding trauma patient which are discussed in this guideline. aPTT, activated partial thromboplastin time; AT III, antithrombin III; CT, computerised tomography; FAST, focused abdominal sonography in trauma; FFP, fresh frozen plasma; Hb, haemoglobin; KIU, kallikrein inhibitory units; PCC, prothrombin complex concentrate; PT, prothrombin time; RBC, red blood cell; rFVIIa, recombinant activated coagula-tion factor VII.

R5Further assessment

***Patients presenting with haemorrhagic

shock and an unidentified source of

bleeding should undergo immediate

further assessment.

I.Initial resuscitation and prevention of

further bleeding

II.Diagnosis and monitoring of

bleeding

III.Rapid control of bleeding

V.Management of bleeding and

coagulation

IV.Tissue oxygenation, fluid and

hypothermia

R3Ventilation

***Severely hypovolaemic trauma patients

may not be hyperventilated or subjected to

excessive positive end-expiratory pressure.

R2Initial assessment

***The extent of traumatic haemorrhage

should be clinically assessed using an

established grading system.

R1Minimal elapsed time

***The time elapsed between injury and

operation should be minimised.

Surgical intervention

R13Angiographic embolisation or surgery

***Patients with ongoing haemodynamic instability despite

adequate pelvic ring stabilisation should receive early

angiographic embolisation or surgical bleeding control,

including packing.

R12Pelvic ring closure & stabilisation

***Patients with pelvic ring disruption in

haemorrhagic shock should undergo immediate

pelvic ring closure and stabilisation.

R14Cross clamping

***Early bleeding control should be achieved using

packing, direct surgical bleeding control and local

haemostatic procedures; aortic cross clamping may

be employed as an adjunct bleeding control in the

exsanguinating patient.

R15Damage control surgery

***Damage control surgery should be employed in

the severely injured patient presenting with deep

hemorrhagic shock, signs of ongoing bleeding

and coagulopathy, hypothermia, acidosis,

inaccessible major anatomic injury, a need for

time-consuming procedures or concomitant major

injury outside the abdomen.

Coagulation management

R23Antifibrinolytic agents

***Antifibrinolytic agents may be considered in the

treatment of the bleeding trauma patient at the following

dosages: tranexamic acid 1015 mg/kg followed by an

infusion of 15 mg/kg/h; -aminocaproic acid 100150 mg/kg followed by 15 mg/kg/h; or, after a test dose,

aprotinin 2 million KIU immediately followed by

500,000 KIU/h in an intravenous infusion.

Antifibrinolytic therapy should be stopped once

bleeding has been adequately controlled.

R21Platelets

***Platelets should be administered to maintain a platelet

count above 50109/l. A platelet count above 100109/l

in patients with multiple trauma who are severely

bleeding or have traumatic brain injury may be

maintained with an initial dose of 48 platelet

concentrates or one apheresis pack.

R26AT III

***Antithrombin III should not be employed in the

treatment of the bleeding trauma patient.

R22Fibrinogen or cryoprecipitate

***Treatment with fibrinogen concentrate or

cryoprecipitate should be employed if significant

bleeding is accompanied by a plasma fibrinogen level

less than 1 g/l at an initial fibrinogen concentrate

dose of 34 g or 50 mg/kg of cryoprecipitate

approximately equivalent to 1520 units in a 70 kg

adult. Repeat doses should be guided by laboratory

assessment of fibrinogen levels.

R20FFP***

Patients with massive bleeding or significant bleeding

complicated by coagulopathy (PT or aPTT >1.5 times

control) should be treated with thawed FFP at an initial

dose of 1015 ml/kg; further doses may be required.

R24rFVIIa

***Treatment with rFVIIa may be considered if major

bleeding in blunt trauma persists despite standard

attempts to control bleeding and best practice use of

blood components at an initial dose of 200 g/kg

followed by two 100 g/kg doses administered at 1 and

3 h following the first dose.

R25PCC***

Treatment with prothrombin complex concentrate

according to the manufacturers instructions should be

employed only for the emergency reversal of vitamin K-

dependent oral anticoagulants.

Resuscitation

R17Fluid therapy

***Crystalloids may be applied initially to treat the

bleeding trauma patient. Colloids may be added within

the prescribed limits for each solution.

R16Volume replacement

***A target systolic blood pressure of 80100 mmHg may

be appropriate until major bleeding has been stopped

in the initial phase following trauma without brain injury.

R18Normothermia

Early application of measures to reduce heat loss and

warm the hypothermic patient should be employed to

achieve and maintain normothermia.

R19RBCs

***Treatment should aim to achieve a target Hb of 79 g/dl.

Extent of bleeding

R10Serum lactate

***Serum lactate should be employed to

estimate and monitor the extent of

bleeding and shock.

R8Computed tomography

***Haemodynamically stable patients with

suspected head, chest and/or abdominal

bleeding following high-energy injuries should

undergo further assessment using CT.

Source of bleeding

R7Sonography

***Patients with significant free

intraabdominal fluid and haemodynamic

instability should undergo urgent surgery.

R6Sonography

***Early focused sonography (FAST) should

be employed for the detection of free fluid

in patients with suspected torso trauma.

R9Haematocrit

***Single haematocrit measurements

should not be employed as an isolated

laboratory marker for bleeding.

R11Base deficit

***Base deficit should be employed to

estimate and monitor the extent of

bleeding and shock.

R4Immediate intervention

***Patients presenting with haemorrhagic

shock and an identified source of bleeding

should undergo an immediate bleeding

control procedure unless initial

resuscitation measures are successful. Page 15 of 22(page number not for citation purposes)

humans. available.


The apparent weakness of much of the published evidencecited in this work highlights the need for further clinical stud-ies, and underlines the importance of future research that maylead to more clear-cut evidence-based clinical guidelines. TheGRADE system employed in developing these guidelines [8]is appropriate in that it allows strong recommendations to besupported by weak clinical evidence in a field in which many ofthe ideal randomised controlled clinical trials may never beperformed. Other systems, such as the grades of recommen-dation developed by the Oxford Centre for Evidence-BasedMedicine [9], may be less dependent on the weight of expertopinion, and therefore the process employed between thepublished evidence and guideline recommendation must betransparent in order to ensure acceptance and implementationof the guideline. To minimise the bias introduced by individualexperts, this guideline employed a nominal group process todevelop each recommendation and several Delphi rounds toreach an agreement on the questions to be considered and toreach a final consensus on each recommendation, and thegroup was composed of a multidisciplinary pan-Europeangroup of experts, including the active involvement of repre-sentatives from five of the most relevant European professionalsocieties.

ConclusionWe have made every effort to make these guidelines applica-ble in daily practice in a variety of clinical settings. We wish toemphasise our conviction that a multidisciplinary approach tothe management of the bleeding trauma patient will help cre-ate circumstances in which optimal care can be provided. Bytheir very nature, these guidelines reflect the current state-of-the-art and will need to be updated and revised regularly asimportant new evidence

Management of Bleeding Following Major Trauma - A European Guideline

Documents

Transcript of Management of Bleeding Following Major Trauma - A European Guideline