Massive Transfusion and Control of Hemorrhage in the...

IN

TE

RN

AT I O N A L T R A U M

A

CA

RE

� � �

Based on Special ITACCS Seminar Panels.The International Trauma Anesthesia and Critical Care Society (ITACCS)

is accredited by the Accreditation Council for

Continuing Medical Education (ACCME) for physicians.

This CME activity was planned and produced in

accordance with the ACCME Essentials.

ITACCS designates this CME activity for 15 credit hours in

Category 1 of the Physicians Recognition Award

of the American Medical Association.

� � �

CME QUESTIONS INCLUDEDJANUARY 2003

Massive Transfusionand Control of Hemorrhage

in the Trauma Patient

2 Massive Transfusion and Control of Hemorrhage in the Trauma Patient

LEARNING OBJECTIVES OF THE MONOGRAPH

After completion of this activity, the participant will be able to:

1. Evaluate the etiology and pathophysiology of traumatic shock.2. Describe the management of massive transfusion in the trauma patient.3. Discuss the clinical indications and problems related to the use of

blood, blood components, hemostatic agents, oxygen-carrying vol-ume expanders, and venous thromboembolism prophylaxis.

EDITORS

Charles E. Smith, MD, FRCPC, Professor of Anesthesiology,MetroHealth Medical Center, Case Western Reserve University Schoolof Medicine, Cleveland, Ohio; Chair, ITACCS Special Equipment/Tech-niques Committee

Andrew D. Rosenberg, MD, Chairman, Department of Anesthesi-ology, Hospital for Joint Diseases Orthopaedic Institute, Associate Pro-fessor of Clinical Anesthesiology, New York University School of Medi-cine, New York, New York

Christopher M. Grande, MD, MPH, Lecturer, Department ofAnesthesiology, Perioperative and Pain Medicine, Brigham and Women’sHospital, Harvard Medical School, Boston, Massachusetts; Professor,Department of Anesthesiology, State University of New York, Buffalo,Buffalo, New York; Professor of Anesthesiology, West Virginia UniversitySchool of Medicine, Morgantown, West Virginia; Executive Director,International Trauma Anesthesia and Critical Care Society (ITACCS),World Headquarters Baltimore, Maryland

CONTENTS AND CONTRIBUTORS

Section I: Etiology and Pathophysiology

Chapter 1 Trauma, a disease of bleeding ......................... Page 3Thomas M. Scalea, MD, Physician-in-Chief, Professor ofSurgery, R Adams Cowley Shock Trauma Center, Baltimore,Maryland

Chapter 2 Pathophysiology of traumatic shock .............. Page 5Richard P. Dutton, MD, Associate Director, Division ofAnesthesiology, R Adams Cowley Shock Trauma Center,Baltimore, Maryland

Section II: Therapeutic Strategies

Chapter 3 Surgical perspectives to controlbleeding in trauma .......................................... Page 7Brian R. Plaisier, MD, Department of Surgery, BronsonMethodist Hospital, Kalamazoo, Michigan

Chapter 4 Hemostatic drugs in trauma andorthopaedic practice ..................................... Page 11David Royston, MB, FRCA, Consultant Anaesthetist, RoyalBrompton and Harefield NHS Trust, Harefield, Middlesex,United Kingdom

Chapter 5 Antithrombotics in Trauma Care:Benefits and Pitfalls ...................................... Page 14John K. Stene, MD, PhD, Past President, ITACCS, AssociateProfessor of Anesthesia and Director of Trauma Anesthesia,Milton S. Hershey Medical Center, Hershey, Pennsylvania

Chapter 6 Atraumatic blood salvage and autotransfusionin trauma and surgery .................................. Page 17Sherwin V. Kevy, MD, and Robert Brustowicz, MD, Trans-fusion Service, Children’s Hospital Department of Anes-thesia, Harvard Medical School, Boston, Massachusetts

Section III: Transfusion: Clinical Practice

Chapter 7 Current practices in blood and bloodcomponent therapy ....................................... Page 18Charles E. Smith, MD, FRCPC, Department of Anesthesi-ology, MetroHealth Medical Center, Case Western ReserveUniversity School of Medicine, Cleveland, Ohio

Chapter 8 Immunomodulatory effects of transfusion .. 2David T. Porembka, Do, FCCM, FCCP, Associate Professorof Anesthesia and Surgery, Associate Director of SurgicalIntensive Care, University of Cincinnati Medical Center,Cincinnati, Ohio

Chapter 9 Blood transfusions ........................................ Page 27Andrew D. Rosenberg, MD, Department of Anesthesiol-ogy, Hospital for Joint Diseases/Orthopaedic Institute, NewYork, New York

Chapter 10 Vascular access in trauma: options, risks,benefits, complications ................................. Page 28Maureen Nash Sweeney, MD, Attending Anesthesiologist,Department of Anesthesiology, Department of VeteransAffairs Medical Center, New York, New York

Chapter 11 Principles of fluid warming .......................... Page 30Charles E. Smith, MD, Department of Anesthesiology,MetroHealth Medical Center, Case Western Reserve Uni-versity, Cleveland, Ohio

Chapter 12 Management of massive hemorrhage andtransfusion in trauma ................................... Page 34Georges Desjardins, MD, FRCPC, Division of Trauma Anes-thesia and Critical Care, Ryder Trauma Center, Universityof Miami/Jackson Memorial Medical Center, Miami, Florida

Chapter 13 Rapid infusion and point-of-care chemistrytesting monitoring in massive transfusion:avoiding common pitfalls ............................. Page 38Jeffrey R. Jernigan, MD, and John G. D’Alessio, MD, De-partment of Anesthesiology, Elvis Presley Memorial TraumaCenter, Memphis, Tennessee

Section IV: New Horizons in Synthetic Blood Substitutes

Chapter 14 Hemoglobin-based oxygen-carryingsolutions and hemorrhagic shock ............... Page 40Colin F. Mackenzie, MB, ChB, FRCA, FCCM, Director,National Study Center for Trauma and Emergency Medi-cal Systems, University of Maryland School of Medicine,Baltimore, Maryland

Chapter 15 Hemoglobin therapeutics, blood substitutes,and high-volume blood loss .......................... Page 44Armin Schubert, MD, MBA, Chairman, Department of Gen-eral Anesthesia, Cleveland Clinic Foundation, Cleveland, Ohio

CME Questions .................................................................. Page 48

The drug and dosage information presented in this publication isbelieved to be accurate. However, the reader is urged to consult the fullprescribing information on any product mentioned in this publicationfor recommended dosage, indications, contraindications, warnings,precautions, and adverse effects. This is particularly important for drugsthat are new or prescribed infrequently.

Massive Transfusion and Control of Hemorrhage in the Trauma Patient 3

IntroductionPriorities in trauma patient management

are to ensure adequate ventilation and oxygen-ation, control hemorrhage, and restore tissueperfusion to vital organs. The most familiarmeans to control hemorrhage are surgical liga-tures and clips. Other means includetranscatheter embolization, appropriate bloodcomponent therapy, maintenance of normo-thermia, and pharmacologic agents. Finally,attention must also be directed toward treat-ment of the hypercoaguable state that followsmajor traumatic injury and can lead to deepvenous thrombosis and pulmonary embolism.

The management of massive transfusionand control of hemorrhage in the trauma pa-tient were discussed during two specialITACCS seminars. The 15 reports in thismonograph summarize the state-of-the artknowledge and clinical practice issues regard-ing surgical and nonsurgical management ofmassive transfusion and control of hemor-rhage in the injured patient.

In the section on “Etiology and Patho-physiology,” Dr. Scalea reviews the physiologicimportance of recognizing and restoring he-mostasis following injury and discusses theAmerican College of Surgeons classificationscheme for hemorrhage, as well as operativeand nonoperative (e.g., embolization) tech-niques for treatment of ongoing blood loss.Dr. Dutton discusses the four phases of trau-matic shock and reviews the macro- and mi-cro-circulatory responses to traumaticshock—responses that ultimately determinepatient outcome.

The “Therapeutic Strategies” section be-gins with a report on surgical perspectives tocontrol bleeding in trauma. In that article, Dr.Plaisier describes the benefits and risks of topi-cal hemostatic agents such as oxidized cellu-lose, collagen sponges, thrombin, denaturedgelfoam, and fibrin glue. Dr. Royston reviewsthe hemostatic and anti-inflammatory effectsof a variety of drugs in trauma. There appearsto be a significant benefit of high-doseaprotinin therapy to reduce blood loss and theneed for blood and blood product transfusion.Major post-traumatic morbidity and mortalitymay result from venous thromboembolism,and Dr. Stene discusses therapeutic strategiesto prevent and treat deep venous thrombosisand pulmonary embolism in the injured pa-tient. In the article on atraumatic blood sal-vage and autotransfusion, Drs. Kevy andBrustowicz critique the use of surgical suctionsystems as a means of reducing (or supple-menting) allogeneic blood use. Dr. Smith ana-lyzes the use of fluid and blood componenttherapy in trauma and addresses various issuessuch as delayed fluid resuscitation, hypertonic

fluids, endpoints of fluid and blood resuscita-tion, complications of transfusion therapy, andclinical strategies to reduce complications.

The section on “Transfusion: Clinical Prac-tice” begins with a discussion on the immuno-logic consequences of transfusions and con-cludes that allogeneic transfusions have a dy-namic immunomodulatory effect on the recipi-ent and that leukocytes are the chief mediatorof these effects. Dr. Rosenberg reviews the sci-entific literature and his own personal experi-ence with the concept of “decreasing theamount of blood transfused to trauma pa-tients” in light of transfusion-related immuno-suppression and other risks. Dr. Sweeneyevaluates the options, risks, and potential com-plications of obtaining vascular access intrauma, illustrating the different approachesin pediatric and adult trauma patients. Theprinciples of warming IV fluid and blood arereviewed by Dr. Smith, with special emphasison the thermal stress of infusing cold or inad-equately warmed fluids, and the safety and ef-ficacy of fluid warmers and rapid infusion de-vices. Dr. Desjardins focuses on the manage-ment of exsanguinating hemorrhage (other-wise known as “massive, massive transfusion”)and reports on the washing and centrifugingof packed red blood cells prior to rapid infu-sion in order to decrease adverse metabolicconsequences such as hyperkalemia. Drs.Jernigan and D’Alessio discuss their experienceusing rapid infusion devices to deliver massivequantities of fluids, blood, and blood productsto maintain circulating blood volume. Theseauthors point out the controversies over hy-

potensive versus normotensive resuscitation,the benefits of point-of-care testing, and theuse of guidelines (in conjunction with theblood bank) for managing trauma patients whorequire “rapid infusion.”

In the final section on “New Horizons inSynthetic Blood Substitutes,” Dr. Mackenziereviews the complex issues surrounding theuse of hemoglobin solutions and hemorrhagicshock. He states that, although many of theproblems associated with oxygen-carrying so-lutions have been overcome, there is a paucityof published data concerning the use of oxy-gen-carrying solutions in humans with hem-orrhagic shock. Dr. Schubert concludes themonograph by examining the potential clini-cal uses and effectiveness of hemoglobin-basedoxygen carriers and perfluorocarbons. Thelong shelf life, long circulation half-life, andgood oxygen-carrying capacity and tissue oxy-gen delivery make these compounds particu-larly attractive in patients with high blood loss,i.e., trauma patients. In his manuscript, Dr.Schubert evaluates the different hemoglobinsolutions and the pitfalls associated with theirclinical use.

As editors and principal organizers of thisspecial ITACCS symposium, we have attemptedto provide a concise, up-to-date reference onmassive transfusion and management of hem-orrhage in the trauma patient—a reference thatintegrates both basic science and clinical prac-tice. We sincerely hope that you, the reader,will obtain essential knowledge from thismonograph that will improve your clinicalpractice when caring for trauma patients.

Massive Transfusion and Control of Hemorrhagein the Trauma Patient

Thomas M. Scalea, MDPhysician-in-ChiefR Adams Cowley Shock Trauma CenterUniversity of Maryland School of Medicine22 South Greene StreetBaltimore, MD 21201 USA

Acute blood loss is a very common prob-lem following injury. Rapid recognition andrestoration of homeostasis is the cornerstoneof the initial care of any badly injured patient.Untreated, hemorrhage robs the cardiovascu-lar system of the preload necessary to ensureadequate cardiac output and peripheral oxy-gen delivery. Inadequate perfusion, even if itis not associated with overt hypotension, canset off the neurohumoral cascade, ultimatelyleading to sequential organ failure.1 This isespecially important, as the mortality from es-

SECTION I: Etiology and Pathophysiology

Trauma, A Disease of Bleedingtablished organ failure has not changed sinceit was first described almost 25 years ago.2

Thus, it is imperative that hemorrhage is rec-ognized and treated early.

The recognition of acute hemorrhage canbe difficult. The American College of Surgeonshas developed the classification scheme forhemorrhage, stratifying blood loss from Stage1 (less than 15% of total circulating blood vol-ume) to Stage 4 (more than 40% of total circu-lating blood volume).3 Changes in variousphysiologic parameters as hemorrhagevolume increases are listed in Table 1. Unfor-tunately, many of these signs and symptomsare nonspecific. In addition, a number of otherparameters will affect the patient’s vital signsand physical findings. For instance, the rapid-ity of volume loss may be as important as thetotal volume of hemorrhage.2 Underlying car-

1


diovascular reserve also plays a role. Youngpeople with very compliant blood vessels maycompensate extremely well for large-volumeblood loss, even as much as 40% to 50% oftotal circulating blood volume.4 They then de-velop sudden cardiovascular compromisewhen compensatory mechanisms fail. Elderlypeople, on the other hand, will develop car-diovascular insufficiency and hypotension withmuch smaller blood loss.5 Prescription medi-cation and/or illicit drugs will also influencethe cardiovascular response to injury.6,7 Theamount of resuscitation, if any, the patient re-ceives in the field will affect cardiovascular re-sponse as well.4

Data from the past 10 years strongly sug-gest that normally followed vital signs are avery poor indication of the depth of hemor-rhage.8 In particular, blood pressure and pulserate, the two vital signs often used in the emer-gency department to gauge hemorrhage, aretremendously nonspecific. Central venous oxy-gen saturation and mixed venous oxygen satu-ration are far more sensitive and reliable mea-surements of acute volume loss.8,9 Degree ofmetabolic acidosis, as measured by the basedeficit from an arterial blood gas, is also ex-tremely helpful in gauging the degree ofshock.10 Base deficit has been shown to corre-late with transfusion requirements, ICU stay,and ultimate outcome.11,12 During initial resus-citation, base deficit should also correlate withserum lactate level. The ability to clear lactateto normal is one of the most important pre-dictors of survival following hemorrhage andinjury.13,14

Measures such as mixed venous oxygencontent, venous oxygen saturation, blood pres-sure, and lactate concentration are global mea-

surements. That is, flow from all vascular bedscontributes to this determination. However,some vascular beds are more sensitive thanothers to the effects of hemorrhage. Thus,shock may be detected earlier if we are able torecognize a local decrease in perfusion. Shockis defined as inadequacy of peripheral oxygendelivery. Clinically, we use indirect measure-ments to gauge hemorrhage. Target organ func-tion such as urine output or mental status areexamples of this. Unfortunately, urine outputis extremely variable and nonspecific. Althougholiguria almost certainly indicates hypov-olemia, normal urine output or polyuria is in-conclusive. Renal tubular function is affectedby as little as a 20% acute loss of blood vol-ume. The kidney develops a salt-wasting neph-ropathy, and the patient makes more urinethan is appropriate for this degree of physi-ologic insult.15 Blood flow to the gastrointesti-nal tract, however, is a relatively sensitive indi-cator of the loss of circulating blood volume.Intracellular pH, as measured in the stomach,small bowel, or colon, is a very sensitive mea-sure of hemorrhage.16 Current technology doesnot allow us to measure intracellular pH in realtime. However, that technology may be forth-coming in the not-too-distant future.

Once the clinician has made the diagno-sis of acute blood loss, several issues becomeimportant. Traditional dogma suggests thatrestoration of forward flow by crystalloid re-suscitation followed by blood is optimaltherapy. However, increases in blood pressureproduced by fluid may, in fact, increase bloodloss by displacing the hemostatic clot that wasformed at the time of hypotension.17 This is-sue will be discussed in Chapter 3. However,there are now data to suggest that sustained

hypotension produces a more injurious shockinsult than do multiple episodes of shock andresuscitation.18 Thus, the clinician must esti-mate the degree of hemorrhage, the depth ofshock, and the time to definitive hemostasiswhen making a decision.

Regardless of the resuscitation decision,patients who demonstrate ongoing bleedingrequire definitive hemostasis. Serial blood gasdeterminations and/or central venous oxygensaturation determination may be very helpfulin determining whether blood loss is continu-ing.8,9 Unfortunately, the relationship betweenblood loss and physiologic parameters may bedifferent after resuscitation than they wereduring hemorrhage. For instance, approxi-mately 12 to 16 hours following resuscitation,the relationship changes between base deficitand anion gap versus serum lactate, and an-ion gap and base deficit no longer correlatewith lactate.19 During this time, one must di-rectly measure serum lactate, as it cannot beinferred from either of the other two measure-ments. When resuscitation decisions are basedon these parameters, therapy will be inappro-priate almost 50% of the time.

Elderly patients with poor underlying car-diovascular reserve often require invasivemonitoring to precisely measure the physi-ologic deficits and to guide therapy. In fact, inhigh-risk elderly patients (Table 2), monitor-ing must be instituted extremely early, within2 to 3 hours of injury if possible. There is astatistically significant decrease in survivalwhen monitoring is delayed as long as 6 hours.5

Even young people may have inadequate car-diovascular response to substantial injuries. Asurprising percentage of young patients witheither blunt or penetrating trauma benefit frominvasive monitoring and require volume andpharmacologic therapy to support cardiovas-cular performance and clear lactate.20,21

Clearly, achieving hemostasis is the mostimportant part of resuscitating the trauma vic-tim. Resuscitation efforts will not be success-ful until blood loss is arrested. Substantial hem-orrhage usually requires operative therapy.Recently, however, other techniques haveemerged and should be considered, even inpatients with hypotension. The diagnosis ofongoing blood loss with angiography and he-mostasis with transcatheter embolization is areal alternative to standard operative therapy.22

This has been a mainstay of therapy for manyyears in patients bleeding from a blunt pelvicinjury. Retroperitoneal exploration in these

Amounts are based on the patient’s initial presentation. Assumes 70-kg male with a bloodvolume of ~70 ml/kg.

Adapted from the American College of Surgeons Committee on Trauma: Advanced TraumaLife Support Program for Physicians, Student and Instructor Manual, Chicago, American Col-lege of Surgeons, 1993.

Table 1. American College of Surgeons Classification of Acute Hemorrhage

Class I II II IV

Blood loss (ml) <750 750-1,500 1,500-2,000 ≥ 2,000

% Blood volume lost <15% 15-30% 30-40% ≥ 40%

Pulse rate <100 >100 >120 ≥ 140

Blood pressure Normal Normal Decreased Decreased

Pulse pressure Normal or Decreased Decreased Decreased(mmHg) increased

Capillary refill Normal Delayed Delayed Delayed

Respiratory rate 14-20 20-30 30-40 >35

Urine output >30 20-30 5-15 Negligible

Mental status Slightly anxious Mildly anxious Anxious, Confused,confused lethargic

Recommended fluid 0.9% saline, 3:1 0.9% saline, 3:1 0.9% saline 0.9% salinereplacement + red cells +red cells

Table 2. High-Risk Geriatric Patients

Initial systolic blood pressure <130 mmHg

Closed head injury

Multiple long-bone fractures

Metabolic acidosis

Pedestrian–motor vehicle mechanism


patients is fraught with danger, and emboliza-tion is far preferable in almost every case. Thesetechniques have been extended to other areasof the body. More recently, transcatheter em-bolization has been used for nonoperativemanagement of solid visceral injuries withinthe abdomen. Treatment algorithms usingsplenic artery embolization in patients man-aged nonoperatively have resulted in a greaterthan 90% rate of splenic salvage.23 This is farhigher than any series utilizing observationand/or operation alone. In addition,embolotherapy may be extremely helpful inpatients with vascular injuries in relatively in-accessible areas. Exposure of the carotid ar-tery in Zone 3 of the neck is extremely diffi-cult. Embolotherapy has a real role in manag-ing these injuries. Temporary hemostasis canbe achieved with percutaneous balloons usedat the time of diagnostic angiography. This tem-porary control of bleeding allows further im-aging, ongoing resuscitative efforts, and timeto plan definitive therapy. In addition to itsusefulness in Zone 3 of the neck, angiographichemostasis has great utility in injuries to thethoracic outlet and deep within the pelvis.

Embolization techniques can be com-bined with surgery, allowing the patient tobenefit from both techniques. Ideally, thisshould be done in the operating room and, insome centers, biplanar angiography is avail-able. Patients who may benefit from this tech-nology are those with a combination of intra-abdominal blood loss and pelvic blood loss.The pelvic blood loss can be embolized whileintra-abdominal blood loss is treated directlyvia surgery. Sometimes patients are too pro-foundly ill to allow definitive surgery. Damagecontrol techniques should then be employed.In these settings, major vascular injuries arerepaired and gastrointestinal contaminationcontrolled. The patient is then packed withlaparotomy pads and taken to the intensivecare unit for ongoing resuscitation and warm-ing techniques. Once patients are resuscitated,they can return to the operating room for un-packing, gastrointestinal reconstruction, andany other procedures necessary. Angiographicembolotherapy has a role in these patients aswell and can be utilized postoperatively tosupplement surgical hemostasis. Injuries deepwithin the substance of the liver, in theretroperitoneum, or in the pelvis may be moreeasily controlled via embolization than surgery.

Early recognition of hemorrhage is key tothe optimal care of trauma patients. Ongoingcontroversies exist as to the ideal resuscitationscheme. In fact, there is probably no one idealstrategy. Care must be tailored to the patient’smechanism of injury and physiology.Nonoperative homeostasis can supplement sur-gical techniques and its use should be consid-ered. Normally followed vital signs are very poorindicators of the degree of hemorrhage and theadequacy of resuscitation. Invasive monitoringis often necessary to precisely determine thephysiologic deficit and guide therapy.

References1. Dantzker D. Oxygen delivery and utilization

in sepsis. Crit Care Clin 1989; 5:81–98.2. Scalea TM, Henry SM. Inotropes in the

intensive care unit. In Advances inTrauma and Critical Care, vol. 7. St.Louis, Mosby, 1992.

3. Committee on Trauma, American Collegeof Surgeons. The Advanced Trauma LifeSupport Program, Instructors Manual.Chicago, American College of Surgeons,1988, pp 59–62.

4. Lewis FR. Prehospital intravenous fluidtherapy: a physiologic computerizedmodel. J Trauma 1986; 26:804–11.

5. Scalea TM, Simon HM, Duncan AL, et al.Geriatric blunt trauma: improved survivalwith early invasive monitoring. J Trauma1990; 30:129–36.

6. Horton JW. Ethanol impairscardiocirculatory function in treated caninehemorrhagic shock. Surgery 1986; 100:520.

7. Sloan EP, Zalenski RJ, Smith RF, et al. Toxi-cology screening in urban trauma pa-tients: drug prevalence and its relation-ship to trauma severity and management.J Trauma 1989; 29:1647.

8. Scalea TM, Holman M, Fuortes M, et al.Central venous blood oxygen saturation:an early accurate measurement of volumeduring hemorrhage. J Trauma 1988;28:725–32.

9. Scalea TM, Hartnett RW, Duncan AO, et al.Central venous oxygen saturation: a use-ful clinical tool in trauma patients. JTrauma 1990; 30:1529–44.

10. Rutherford EJ, Morris JA, Reed GW, et al.Base deficit stratifies mortality and deter-mines therapy. J Trauma 1992; 33:417.

11. Davis JW, Shackford SR, MacKersie RC, HoytDB. Base deficit as a guide to volume re-suscitation. J Trauma 1988; 28:1464–7.

12. Davis JW, Parks SN, Kaups KL, et al. Ad-

mission base deficit predicts transfusionrequirements and risk of complications. JTrauma 1996; 41:769–74.

13. Iberti TJ, Leibowitz AB, Papdakos PJ, et al.Low cardiac sensitivity of the anion gap as ascreen to detect hyperlactatemia in criticallyill patients. Crit Care Med 1990; 18:275–7.

14. Abramson D, Scalea TM, Hitchcock D, etal. Lactate clearance and survival follow-ing injury. J Trauma 1993; 35:584–9.

15. Sinert R, Baron B, Low R, et al. Is urineoutput a reliable index of blood volumein hemorrhagic shock? Acad Emerg Med1996; 3:448.

16. Baron BJ, Scalea TM. Acute blood loss.Emerg Med Clin North Am 1996; 14:35–54.

17. Shaftan GW, Chui C, Dennis C, et al. Fun-damentals of physiologic control of arte-rial hemorrhage. Surgery 1965; 58:851.

18. Sinha HA, Baron BJ, Buckley MC, et al. Fluidrestriction versus early resuscitation in hem-orrhagic shock. J Trauma 1994; 37:1015.

19. Mikulaschek A, Henry SM, Donovan R,Scalea TM. Serum lactate is not predictedby anion gap or base excess after traumaresuscitation. J Trauma 1996; 40:218–24.

20. Abou-Khalil B, Scalea TM, Trooskin SZ. He-modynamic responses to shock in youngtrauma patients: the need for invasive moni-toring. Crit Care Med 1994; 22:633–9.

21. Scalea TM, Maltz S, Yelon J, et al. Resusci-tation of multiple trauma and head inju-ries: role of crystalloid fluid and inotropes.Crit Care Med 1994; 22:1610–5.

22. Panetta T, Sclafani SJA, Goldstein AJ, et al.Percutaneous transcatheter embolizationfor massive bleeding from pelvic fractures.J Trauma 1985; 25:1021.

23. Sclafani SJA, Scalea TM, Herskowitz M, etal. Salvage of CT-diagnosed splenic inju-ries: utilization of angiography for triageand embolization for hemostasis. JTrauma 1995; 39:818–27.

Pathophysiology of Traumatic ShockRichard P. Dutton, MDDirector, Trauma AnesthesiaR Adams Cowley Shock Trauma CenterUniversity of Maryland School of MedicineBaltimore MD 21201 USAe-mail: [email protected]

Shock—a condition of decreased totalbody oxygen delivery—can be brought on bya number of mechanisms. These include fail-ure of the heart to pump blood through thebody (cardiogenic), loss of circulating fluidvolume (hemorrhagic), decreased oxygen car-rying capacity (anemic), or loss of vascular tone(neurogenic).1 “Traumatic shock”—shockbrought on by an injury in an otherwise healthypatient—is best thought of as a combinationof these factors. The initial phase is usually

hemorrhagic: the patient bleeds, and perfusiondecreases. This may be followed by an anemicphase as the patient is resuscitated with crys-talloid solutions and simultaneously mobilizesinterstitial fluid into the vasculature. A cardio-genic or neurogenic component may bepresent initially due to specific injuries to theheart or central nervous system (CNS) or maybe the secondary result of hypoperfusion andthe release of toxic factors. It is important torecognize that the traumatic shock seen clini-cally in severely injured patients may be quitedifferent from the induced shock seen in labo-ratory animals hemorrhaged under controlledconditions.

Stages of ShockTraumatic shock may be thought of as

2


occurring in four phases (Fig. 1). In compen-sated traumatic shock, an increase in heartrate and vasoconstriction of nonessential andischemia-tolerant vascular beds will allow pro-longed survival and easy recovery once coagu-lation occurs and adequate fluids and nutri-tion are provided. Decompensated traumaticshock, also known as progressive shock, is atransitory state in which the lack of perfusionto certain tissues is building up a debt of localcell damage that will produce a toxic effect onthe organism when perfusion is reestablished.Shock is still reversible at this stage. In sub-acute irreversible shock, the patient can beresuscitated hemodynamically but succumbsat a later time to multiple organ system failureas a result of the toxic effects of ischemia andreperfusion. Finally, acute irreversible shockis the condition of ongoing hemorrhage, aci-dosis, and coagulopathy that spirals steadilydownward to the patient’s demise.1,2

The patient whose hemorrhage has pro-ceeded to the point of decompensated shockrepresents a surgical and metabolic emergency.If the loss of blood (and thus the loss of oxy-gen-delivering capacity) can be reversed beforethe inflammatory cascade begins, the patientwill survive. Adequate volume resuscitationleads the patient into higher-than-normal oxy-gen consumption—a hypermetabolic state—for hours to days after the acute injury, as thebody repays the metabolic debt built up dur-ing the period of ischemia.3 Figure 1 showsthis patient following curve C and eventuallyachieving normal equilibrium.

A few minutes too late, however, and sub-acute irreversible shock will have occurred, asrepresented by curve D. Bleeding may be con-trolled and vital signs may be normal or evenhypernormal, but the damage has been doneon the cellular level. Some tissues will continueto be ischemic due to lack of reflow caused bycellular swelling and microcirculatory obstruc-tion. When flow is successfully restored on thecellular level, the process of reperfusion be-gins. This washout of toxins and inflammatoryfactors is as dangerous to the patient as thehemorrhage itself. This is the patient who de-velops adult respiratory distress syndrome thenprogresses to acute renal failure, gut dysfunc-tion, immunosuppression, cardiac failure, andeventual death due to multiple organ systemfailure. Even though the hemorrhage isstopped short of exsanguination, the body isunable to survive the ischemic insult.2,4

Curve E in Figure 1 represents acute irre-versible traumatic shock. Prolonged hypoten-sion is followed by progressive vasodilatation,loss of response to fluids and catecholamines,capillary leak, diffuse coagulopathy, cardiacdysfunction, and early death. These patientsare usually said to have exsanguinated, al-though in the presence of modern rapid infu-sion techniques and aggressive transfusion, thisis not strictly true. Rather, the patient dies fromthe acute metabolic consequences of failedperfusion, frequently in the presence of ad-

equate control of surgical bleeding and volu-minous blood product replacement.1

The Body’s Response to ShockThe stages of traumatic shock are directly

related to the body’s response to hemorrhage.The initial response is on the macrocirculatorylevel and is mediated by the neuroendocrinesystem. Decreased blood pressure leads tovasoconstriction and catecholamine release.Heart and brain blood flow is preserved, whileother regional beds are constricted. Pain, hem-orrhage, and cortical perception of traumaticinjuries lead to the release of a number of hor-mones, including renin–angiotensin, vaso-pressin, antidiuretic hormone, growth hor-mone, glucagon, cortisol, epinephrine andnorepinephrine.5 This response sets the stagefor the microcirculatory responses that willultimately determine the patient’s outcome.

On the cellular level the body respondsto hemorrhage by taking up interstitial fluid,causing cells to swell.6 This may choke off ad-jacent capillaries, resulting in the “no-reflow”phenomenon that prevents the reversal of is-chemia even in the presence of adequatemacro flow.7 Ischemic cells produce lactate andfree radicals, which are not cleared by the cir-culation. These compounds cause direct dam-age to the cell, as well as comprising the bulkof the toxic load that will be washed back tothe central circulation when perfusion is rees-tablished. The ischemic cell will also produceand release a variety of inflammatory factors:prostacyclin, thromboxane, prostaglandins,leukotrienes, endothelin, complement,interleukins, tumor necrosis factor, and oth-ers.1 These are the ingredients of acute andsubacute irreversible shock.

Organ System Responses to TraumaticShock

Specific organ systems respond to trau-matic shock in specific ways. The CNS is theprime trigger of the neuroendocrine responseto shock, which maintains perfusion to theheart and brain at the expense of other tissues.8

Regional glucose uptake in the brain changesduring shock.9 Reflex activity and cortical elec-trical activity are both depressed during hy-potension; these changes are reversible withmild hypoperfusion, but become permanentwith prolonged ischemia. Failure to recoverpreinjury neurologic function is a marker forsubacute irreversible shock, even if thepatient’s hemodynamic functions are normal.10

The kidney and adrenal glands are primeresponders to the neuroendocrine changes ofshock, producing renin, angiotensin, aldoster-one, cortisol, erythropoietin, and catechola-mines.11 The kidney itself maintains glomeru-lar filtration in the face of hypotension by se-lective vasoconstriction and concentration ofblood flow in the medulla and deep corticalarea. Prolonged hypotension leads to de-creased cellular energy and an inability to con-centrate urine, followed by patchy cell death,tubular epithelial necrosis, and renal failure.8,12

The heart is relatively preserved from is-chemia during shock because of maintenanceor even increase of nutrient blood flow, andcardiac function is generally well preserveduntil the late stages.8,11 Lactate, free radicals,and other humoral factors released by ischemiccells all act as negative inotropes, however, andin the decompensated patient may producecardiac dysfunction as the terminal event inthe shock spiral.13

The lung, which cannot itself become is-chemic, is nonetheless the downstream filter

Figure 1

Traumatic shock and its potential outcomes. A. In early shock there is only a small drop inoxygen delivery due to compensation by the cardiovascular system. B. Decompensatedshock is characterized by an accelerating defect in oxygen delivery. C. Recovery from dec-ompensated shock includes a hyperdynamic period as the body’s oxygen debt is repaid. D.In subacute irreversible shock, the macrocirculation is restored and bleeding stopped, buthypoperfusion has been severe enough that oxygen debt cannot be repaid. Lethal multipleorgan system failure develops. E. Acute irreversible shock occurs when hemodynamic con-trol is never regained. The patient exsanguinates and dies in cardiovascular collapse.


for the inflammatory byproducts of the is-chemic body. The lung is often the sentinelorgan for the development of multiple organsystem failure.4,14 Immune complex and cellu-lar factors accumulate in the capillaries of thelung, leading to neutrophil and platelet aggre-gation, increased capillary permeability, de-struction of lung architecture, and the acuterespiratory distress syndrome.15,16 The pulmo-nary response to traumatic shock is the lead-ing evidence that this disease is not just a dis-order of hemodynamics: pure hemorrhage, inthe absence of hypoperfusion, does not pro-duce pulmonary dysfunction.14,17

The intestine is one of the earliest organsaffected by hypoperfusion and may be one ofthe primary triggers of multiple organ systemfailure. Intense vasoconstriction occurs early,and frequently leads to a “no-reflow” phenom-enon even when the macrocirculation is re-stored.18 Intestinal cell death causes a break-down in the barrier function of the gut, whichresults in increased translocation of bacteriato the liver and lung.19 The impact of this onthe development of multiple organ failure iscontroversial at present.20

The liver has a complex microcirculationand has been demonstrated to sufferreperfusion injury during recovery fromshock.21 Hepatic cells are also metabolicallyactive and contribute substantially to the inflam-matory response to decompensated shock. Ir-regularities in blood glucose levels followingshock are attributable to hepatic ischemia.22

Failure of the synthetic functions of the liverfollowing shock are almost always lethal.

Skeletal muscle is not metabolically activeduring shock, and tolerates ischemia betterthan other organs. The large mass of skeletalmuscle, though, makes it important in the gen-eration of lactate and free radicals from is-chemic cells. The classic cellular response toshock of increasing intracellular sodium andfree water were first elucidated in skeletalmuscle cells.23

ConclusionTraumatic shock is a disease not just of

hemorrhage but also of tissue ischemia. Bleed-ing can be controlled surgically and oxygendelivery restored through adequate transfu-sion, and the patient can still die as a result ofthe accumulated metabolic load of prolongedhypoperfusion. Although control of bleedingand restoration of the circulating blood vol-ume must remain the cornerstones of care forthe traumatized patient, we must build on thisfoundation techniques for the management ofreperfusion injury, the inflammatory cascade,and “no reflow” if we are truly going to im-prove long-term survival.

References1. Peitzman AB, Billiar TR, Harbrecht BG,

Kelly E, Udekwu AO, Simmons RL. Hem-orrhagic Shock. Curr Probl Surg1995;929–1002.

Brian R. Plaisier, MDDepartment of SurgeryBronson Methodist Hospital252 East Lovell, Box 67Kalamazoo MI 49007 USAe-mail: [email protected]

After establishing a secure airway and en-suring adequate oxygenation and ventilation,the highest priority in the trauma patient is tocontrol hemorrhage. Because patients maybleed from multiple sites simultaneously, it isimperative that the surgeon establish a strategyto address all possible sources of bleeding andcontrol them. These sources include 1) exter-

SECTION II: Therapeutic Strategies

Surgical Perspectives to Control Bleeding in Trauma

nal – blood loss onto the “street” or the traumaroom floor, 2) left and right hemithoraces, 3)peritoneal cavity, 4) pelvis and retroperitoneum,and 5) long-bone fracture sites. Methods of de-finitive hemostatic control may be very simple,as in the application of direct pressure to a lac-eration, or very complex, such as in the patientwith a pelvic fracture who requires emboliza-tion. This article addresses surgical, pharmaco-logic, and various other nonsurgical methodsto control bleeding.

HemostasisHemostasis is the process that terminates

blood loss from an injured blood vessel. Sur-

2. Shoemaker WC, Peitzman AB, et al. Resus-citation from severe hemorrhage. CritCare Med 1996; 24:S12–23.

3. Cerra FB. Metabolic response to injury. InCerra FB, ed. Manual of Critical Care. St.Louis, CV Mosby, 1987, pp 117–45.

4. Demling R, Lalonde C, Saldinger P, KnoxJ. Multiple organ dysfunction in the sur-gical patient: pathophysiology, preven-tion, and treatment. Curr Probl Surg 1993;30:345–424.

5. Peitzman AB. Hypovolemic shock. In PinskyMR, Dhainaut JFA, eds. PathophysiologicFoundations of Critical Care. Baltimore,Williams & Wilkins, 1993, pp 161–9.

6. Shires GT, Cunningham N, Baker CRF, etal. Alterations in cellular membrane func-tion during hemorrhagic shock in pri-mates. Ann Surg 1972; 176:288–95.

7. Shires GT, Coln D, Carrico J, et al. Fluidtherapy in hemorrhagic shock. Arch Surg1964; 88:688–93.

8. Runciman WB, Sjowronski GA. Patho-physiology of haemorrhagic shock.Anaesth Intensive Care 1984; 12:193–205.

9. Bronshvag MM. Cerebral pathophysiologyin haemorrhagic shock: nuclide scan data,fluorescence microscopy, and anatomiccorrelations. Stroke 1980; 11:50–9.

10. Peterson CG, Haugen FP. Hemorrhagicshock and the nervous system. Am J Surg1963; 106:233–9.

11. Collins JA. The pathophysiology of hem-orrhagic shock. Prog Clin Biol Res 1982;108:5–29.

12. Troyer DA. Models of ischemic acute re-nal failure: do they reflect events in hu-man renal failure? J Lab Clin Med 1987;110:379–80.

13. Lefer AM, Martin J. Origin of a myocardialdepressant factor in shock. Am J Physiol1970; 218:1423–7.

14. Horovitz, JH, Carrico CJ, Shires GT. Pul-

monary response to major injury. ArchSurg 1974; 108:349–55.

15. Thorne J, Blomquist S, Elmer O. Polymor-phonuclear leukocyte sequestration in thelung and liver following soft tissue trauma:an in vivo study. J Trauma 1989; 29:451–6.

16. Martin BA, Dahlby R, Nicholls I, Hogg JC.Platelet sequestration in lungs with hem-orrhagic shock and reinfusion in dogs. JAppl Physiol 1981; 50:1306–12.

17. Fulton RL, Raynor AVS, Jones C. Analysisof factors leading to posttraumatic pulmo-nary insufficiency. Ann Thorac Surg 1978;25:500–9.

18. Reilly PM, Bulkley GB. Vasoactive media-tors and splanchnic perfusion. Crit CareMed 1993; 21:S55–68.

19. Redan JA, Rush BF, McCullogh JN, et al.Organ distribution of radiolabeled entericEscherichia coli during and after hemor-rhagic shock. Ann Surg 1990; 211:663–8.

20. Korinek AM, Laisne MJ, Nicholas NH,Raskine L, Deroin V, Sanson-Lepors MJ. Se-lective decontamination of the digestivetract in neurosurgical intensive care pa-tients: a double-blind, randomized, pla-cebo-controlled study. Crit Care Med1993; 21:1466–73.

21. Chun K, Zhang J, Biewer J, Ferguson D,Clemens MG. Microcirculatory failure deter-mines lethal hepatocyte injury in ischemic-reperfused rat livers. Shock 1994; 1:3–9.

22. Maitra SR, Geller ER, Pan W, Kennedy PR,Higgins LD. Altered cellular calcium regu-lation and hepatic glucose productionduring hemorrhagic shock. Circ Shock1992; 38:14–24.

23. Peitzman AB, Corbett WA, Shires GT III,Illner H, Shires GT, Inamder R. Cellularfunction in liver and muscle during hem-orrhagic shock in primates. Surg GynecolObstet 1985; 161:419–24.

3


geons depend greatly on normal hemostasis,often taking it for granted, so that surgery maybe conducted safely. The process is very effi-cient and utilizes circulating proteins, cellularelements, and the endothelial lining (Fig. 1).1

The first response to injury is vasoconstriction,which decreases blood flow distal to the lac-eration. The mechanism for vasoconstrictioninvolves both direct injury and reflex re-sponses. Platelets are exposed to subendothe-lial collagen and quickly adhere to each otherand the blood vessel wall. Von Willebrand’sfactor acts as a bridge between thesubendothelium and the platelet membrane,where it binds to receptor sites made availableas a result of platelet activation. Other plate-lets are then recruited from the blood, and aloose plug forms to seal the blood vessel.

If this response reaches sufficient intensity,the platelet release reaction occurs whereby thecontents of the platelet and its granules are lib-erated into the surrounding microenvironment.This is a complex reaction involving adenosinediphosphate, serotonin, platelet factor 4, plate-let-derived growth factor, thrombin, calcium,and magnesium.1,2 The result is the formationof a stable platelet plug, which, unlike the ini-tial loose plug, is no longer reversible.

Platelet reactions occur simultaneouslywith the events of the coagulation cascade. Co-agulation serves to convert prothrombin intothrombin, which, in turn, converts fibrinogento fibrin. This process utilizes circulating inac-tive proenzymes, which are converted into anactive form and then, in turn, activate the nextproenzyme in the sequence. There are two dis-tinct divisions of the coagulation process: 1) theintrinsic pathway and 2) the extrinsic pathway(Fig. 2). The intrinsic pathway is initiated by theinteraction of Factor XII and nonendothelialsurfaces, which induces a conformationalchange in Factor XII. The complicated reactionsthat follow lead to clotting, kinin formation,complement activation, and fibrinolysis.1

The extrinsic pathway is the more impor-tant pathway in hemostasis. Thromboplastin,a lipoprotein, is released from cells in responseto tissue trauma. When thromboplastin ispresent, Factor VII becomes active and the se-quence ensues.

The two pathways merge into a commonpathway with the activation of Factor X, which,in turn, converts prothrombin to thrombin. Fi-brinogen is then acted upon by thrombin, re-sulting in the formation of fibrin monomers.Polymerization of the fibrin monomers occurs,resulting in a cross-linked, stable, fibrin clot.

Fibrinolysis is the process that limits thehemostatic response to the local area of injuryand maintains vascular patency throughout theorganism. This system is initiated simultaneouslywith the clotting mechanism and is under theinfluence of numerous circulating mediators. Therelease of plasminogen activator from injuredendothelium and activation of Factor XII initiatefibrinolysis. These convert plasminogen to plas-min, which can digest fibrin and fibrinogen at

Figure 1. Complex interaction of vasoconstriction,platelet factors, and coagulation reactions.

(Reproduced with permission from Mosby-Year Book Inc.)

Figure 2. Coagulation reactions.(Reproduced with permission from Mosby-Year Book, Inc.)

the site of clotting. A complex inhibition systeminactivates any plasmin that gains access to thegeneral circulation. Other methods the body usesto limit coagulation, which are beyond the scopeof this discussion, include products of the cyclo-oxygenase enzyme pathway, protein C, and anti-thrombin III.

Abnormalities of Hemostasis Resultingfrom Injury

Injury triggers a vast array of responsesthat affect hemostasis. Patients may exhibit ei-ther a hypercoagulable or hypocoagulable statefollowing trauma. Severely injured patientshave elevated serum fibrin degradation prod-


ucts, lowered platelet counts, and activation ofthe kallikrein–kinin system.3 In survivors, thesevalues normalize within the first few days afterinjury but continue to worsen in nonsurvivors.Interestingly, a hypercoagulable state is seen inpatients with less severe injuries. This is causedby a suppression of thrombolysis.3

Coagulopathy after injury may also resultfrom the patient’s abnormal physiology(hypoperfusion or hypothermia) or the inter-ventions used to treat the patient (massive trans-fusion). It has become clear that prophylacticadministration of fresh frozen plasma or plate-lets in the absence of clinical bleeding is notwarranted.4,5 Hypothermia has been shown toadversely affect coagulation, and it is importantthat surgeons and anesthesiologists strive tomaintain normothermia during treatment. In-juries to the brain and liver and those that causeeither hypoperfusion or tissue devitalizationhave significant potential to inducecoagulopathy.4,5 The importance of the restora-tion of tissue perfusion and debridement ofdevitalized tissue cannot be overemphasized.

Priorities in the Operating RoomAt laparotomy it is absolutely necessary to

control hemorrhage and gastrointestinal con-tamination in the most rapid fashion possible.Dr. William Halsted6 considered this absolutelyessential for all types of surgery and eloquentlystated the rationale:

The confidence gradually acquiredfrom masterfulness in controllinghemorrhage gives to the surgeon thecalm which is so essential for clearthinking and orderly procedure at theoperating table.

It is only after hemorrhage is controlled thata patient’s injuries may be addressed in an or-derly fashion (Table 1). Control of gastrointes-tinal contamination is the next goal. Only afterthese goals are accomplished can a thoroughexploration of the abdomen can be conductedand all injuries addressed definitively.

The surgeon has a wide range of tools toemploy in order to control bleeding (Table 2).The most obvious method is the applicationof digital pressure. Although not definitivecontrol for large vessels, the surgeon’s fingeris the most atraumatic instrument available andwill control bleeding temporarily while theblood vessel is exposed. The offending bloodvessel must be exposed properly prior to re-pair or ligation. Occasionally, one may needto gain control of the aorta at the diaphrag-matic hiatus to allow the anesthesiologist timeto replace blood and fluids while exposure isbeing accomplished.

The patient’s condition may not allow allinjuries to be addressed fully at initial explo-ration. An abbreviated laparotomy to controlhemorrhage, followed by continued resusci-tation in the intensive care unit, is now an es-tablished concept in trauma surgery.7 If thepatient’s condition is deteriorating after con-

trol of surgical bleeding and if coagulopathy,hypothermia, and acidosis are present, laparo-tomy sponges may be placed between the ab-dominal wall and the bleeding organ to gaintamponade. The laparotomy is terminatedquickly to allow transfer to the intensive careunit so that coagulopathy, acidosis, and hypo-thermia may be corrected. A second operationis required to remove the packs once thepatient’s condition is more stable.

The most familiar means of achieving de-finitive hemostasis is the placement of surgi-cal ligatures and clips. These must be placedvery accurately so as not to endanger surround-ing structures. Small vessels may be managedwith simple ligatures; large arteries should becontrolled with a suture ligature to preventslippage of the tie. In very confined spaceswhere the placement of ties would be difficult,surgical clips may be applied.

Occasionally an organ such as the liver orspleen may be lacerated, but removal may notbe necessary. Organ-wrapping methods utilizea mesh net to envelope the liver or spleen togain tamponade. This method may be usedwhen other methods to achieve hemostasis failor where splenectomy or extensivehepatorraphy would otherwise be required.

Thermal agents such as electrocauteryproduce hemostasis by heating and denatur-ing proteins, resulting in coagulation. Both al-ternating and direct current may be employedfor this purpose. This method allows rapid ces-sation of bleeding but may result in large ar-eas of tissue necrosis if applied carelessly.

Occasionally an organ such as the liver orspleen may be lacerated, but removal may notbe necessary. Hemostasis may be accomplishedby several methods, such as direct pressure orsuture repair. The liver or spleen may also bewrapped with a mesh netting to envelop theorgan to gain tamponade. This method may beused when other methods to achieve hemosta-sis fail or where splenectomy of extensivehepatorraphy would otherwise be required butmay compromise chances for survival.

Table 1.Surgical Priorities at

Laparotomy in the Trauma Patient

• Control of exsanguinatinghemorrhage

• Stop gastrointestinal contamination• Thorough exploration of entire

abdomen• Definitive repair of all injuries

Table 2.Surgical Methods to

Control Bleeding

• Proper exposure• Digital pressure• Sutures and clips• Thermal coagulation• Topical hemostatic agents• Organ wrapping

Table 3.Comparison of Topical Hemostatic Agents

Reproduced with permission from Innovative Publishing Incorporated.


Pharmacologic agents have gained an im-portant place in the surgeon’s armamentarium.The mechanisms of action are widely varied:some act by vasoconstriction, some by supply-ing a scaffold for attracting blood elements, andstill others by promoting coagulation per se(Table 3).8 The ideal topical hemostat wouldhave several properties: 1) rapid hemostasis, 2)easily applied and manipulated, 3) holds su-tures, 4) little tissue reaction, 5) low infectiousrisk, 6) absorbable, and 7) easily removed. Eachof these agents has particular advantages anddisadvantages, which will be discussed in brief.

Topical epinephrine is used commonly inapplications such as burn surgery and exertsits action by promoting vasoconstriction. Thedrug can be used to cover wide surfaces, but itmust be applied with caution because systemiceffects may result if excess drug is used.

Oxidized cellulose (i.e., Surgicel®) acts byforming a gelatinous mass on contact withblood. This compound conforms well to ir-regular surfaces, is relatively inert, causes littletissue reaction, and is absorbed in 1 to 2 weeks.In addition, cellulose holds sutures relativelywell and is bactericidal.

Collagen sponges (Actifoam®, Helistat®,Instat®) and microfibrillar collagen (i.e.,Avitene®) have a rapid time to hemostasis andare absorbed in approximately 8 to 12 weeks.Sponges are easy to apply and they remove andhold sutures well. Microfibrillar collagen packseasily into small spaces but is difficult to re-move and sticks to gloves and instruments.

Thrombin is a protein that converts fi-brinogen to fibrin, resulting in clot formation.Thrombin may be applied as a liquid or pow-der or combined with another carrier such asGelfoam®. Hemostasis is rapid and wide sur-faces may be treated.

Denatured gelatin (i.e,. Gelfoam®) pos-sesses no clotting activity itself but provides ascaffold on which clot can form. It also helpsplug small blood vessels by virtue of its bulkwhen moistened. It may be used as a carrierfor other compounds such as thrombin. Thesponge should be pre-moistened with eithersaline or thrombin and all air should be re-moved from the interstices by compressing thesponge. Gelatin conforms well to surfaces, butit does not hold sutures.

Fibrin sealants have numerous applica-tions within the field of surgery, includingnerve anastomoses, intracranial operations,skin grafting, and cardiovascular procedures.9

Fibrin glue has also been used as a hemostaticagent in trauma surgery for lacerations of theliver and spleen. In the presence of calciumions, fibrinogen and Factor XIII are activatedby thrombin. Fibrinogen is converted to fibrinmonomers and these, in turn, are polymerizedto form a stable clot.

Fibrin glue has two components that mustbe mixed together for clotting to occur. Theprimary parts of the first component are fibrino-gen and Factor XIII. The second componentconsists of thrombin and calcium chloride. An

antifibrinolytic agent such as aprotinin may beadded to the second solution, depending onspecific requirements.9 When these two partsare combined, clotting ensues. The glue maybe applied by two methods: 1) In the “sand-wich technique,” the fibrinogen is spread ontothe surface to be sealed and the thrombin solu-tion spread over it. 2) The premixed methoduses two syringes joined by a Y-connector.

Ochsner et al used fibrin glue as the pri-mary hemostatic agent or as an adjunct to con-ventional suture repair in 26 patients with he-patic and splenic trauma.10 Seventeen patientshad liver injuries (6 blunt and 11 penetrating)and 9 had splenic injuries (7 blunt and 2 pen-etrating). Liver injuries ranged from moderateto severe and the splenic injuries were allmoderate. Fibrin glue achieved hemostasis in21 patients with the first application and withthe second in the remaining five. No patientswere re-explored for bleeding. Eight patientshad postoperative coagulopathy and thromb-ocytopenia, but the fibrin glue hemostasis re-mained effective.

A controlled in vitro review of topical he-mostatic agents was undertaken by Wagner etal.11 The tested agents included three types ofcollagen sponges (Actifoam®, Helistat®,Instat®), microfibrillar collagen (Avitene®), agelatin sponge (Gelfoam®), and oxidized re-generated cellulose (Surgicel®). Actifoam® andAvitene® caused the greatest response (bothstatistically similar) in an in vitro platelet ag-gregation test. Gelfoam® exhibited an interme-diate response, whereas Helistat®, Surgicel®,and Instat® caused a lesser degree of plateletaggregation. In a similar test using thrombin topresoak each agent, platelet aggregation oc-curred at a more rapid rate for all agents tested.

The agents were also tested in their abil-ity to induce gross blood coagulation (Lee–White clotting time). Actifoam®, Avitene®, andHelistat® responded in a manner similar tothrombin, but Instat®, Gelfoam®, andSurgicel® demonstrated no significant impacton clotting time.

Wagner et al, using the above assays as wellas tests of platelet deposition and platelet ad-enosine triphosphate secretion, constructed anoverall ranking of these hemostatic agents:Actifoam® ~ Avitene® > Helistat® >>Gelfoam® > Instat® > Surgicel®. It shouldbe noted that, although this ranking notes dif-ferences between the agents for these in vitroassays, it is certainly limited when consideringthe numerous clinical situations encounteredby surgeons in a wide variety of subspecialties.

Heat energy has a significant role intreating the hypothermic trauma patient.Hypothermia causes platelet dysfunction andprolongs clotting times.12 Laboratory assaysunderestimate the extent to which hypoth-ermia affects bleeding, since the plasma andtest reagents are heated to 37°C prior to run-ning the assay. Because of this, coagulationtest results and platelet counts may not cor-relate with nonsurgical bleeding.

Other Invasive InterventionsNumerous other tools for hemorrhage

control may be used in the field, emergencydepartment, or radiologic suite (Table 4). Al-though the surgeon does not necessarily per-form all of these procedures, he or she shouldbe responsible for combining them into a logi-cal strategy for prompt control of bleedingwhen surgical methods cannot be used.

The pneumatic antishock garment is usedto control bleeding temporarily in patients withpelvic and lower extremity fractures by actingas a splint to tamponade bleeding. It can beused for hypovolemic shock, but it is only atemporizing measure. Prolonged use may beassociated with numerous complications, suchas compartment syndrome.

The external pelvic fixator may be defini-tive in stopping bleeding from veins lining frac-tured pelvic bones. It is most effective in pa-tients with fractures associated with a diasta-sis of the pubic symphysis (“open-book” pel-vic fractures), since it draws the anterior ele-ments together. This decreases the potentialspace into which bleeding may occur. The ex-ternal fixator is not effective for fractures in-volving only the posterior elements of the pel-vic ring or in controlling bleeding from thearteries coursing through the pelvis.

For patients with bleeding from pelvicfractures in whom an external fixator is noteffective, bleeding from arteries in the pelvismust be suspected and angiography should beperformed. If an offending vessel is identified,embolization may be carried out with eitherGelfoam® or metal microcoils (Fig. 3). While

Fig. 3.Microcoils used in pelvic

arterial embolization.(Photograph courtesy of

James Newman, MD, PhD.)

Table 4.Nonsurgical Interventions

to Achieve Hemostasis

• Pneumatic antishock garment• External pelvic fixator• Angiography and embolization• Temporary balloon occlusion


not usually used for pelvic arteries, balloonocclusion may be used by the angiographer asa temporizing measure to achieve hemostasisin arteries of the chest, neck, and extremitiesbefore the causative lesions are controlled inthe operating room.

I would be remiss if I did not emphasizethe importance of the anesthesia service in themanagement of these patients. Surgeons mustfocus on control of bleeding at the surgical site.Anesthesiologists provide the necessary factorsto assist in the correction of surgical bleedingand the prevention of nonsurgical bleeding.The proper transfusion of blood componenttherapy has important implications for controlof bleeding, since platelets and coagulationfactors may be required by severely injuredpatients. Anesthesiologists must also focus onthe maintenance of normothermia to help pre-vent coagulopathy. Effective communicationbetween the surgeon and anesthesiologist isessential. The surgeon must alert the anesthe-siologist to bleeding at the surgical site, so thatcorrective methods may be undertaken.

SummaryControl of blood loss is one of the most

important priorities in the trauma patient. Wehave discussed several methods of obtaininghemostasis. These include standard surgicaltechniques such as digital pressure and su-tures. We have focussed much of our atten-

tion on pharmacologic methods, specificallytopical hemostatic agents. Each of these agentshas particular advantages and disadvantagesand must be applied to the appropriate situa-tion. There are other invasive techniques thatmay not be performed by the surgeon but thatmust be orchestrated by the surgeon into aclear strategy for hemorrhage control. The an-esthesiologist has an important role in help-ing to control hemorrhage by appropriatetransfusion therapy but, more importantly, pre-venting bleeding at the surgical site by meth-ods such as maintaining normothermia.

References1. Clagett GP. Hemostasis in surgical patients.

In Miller TA, ed. Physiologic Basis of Mod-ern Surgical Care. St. Louis, Mosby, 1988.

2. Schwartz SI, Green RM. Biology of hemo-stasis. In Schwartz SI, ed. Techniques ofHemostasis. West Berlin, New Jersey, In-novative Publishing Incorporated, 1993.

3. Rutledge R, Sheldon GF. Bleeding and co-agulation problems. In Feliciano DV,Moore EE, Mattox KL, eds. Trauma, 3rded. Stamford, Connecticut, Appleton andLange, 1996.

4. Knudson MM. Coagulation disorders. InIvatury RR, Cayten CG, eds. The Textbookof Penetrating Trauma. Baltimore, Mary-land, Williams & Wilkins, 1996.

5. Phillips GR, Rotondo MF, Schwab CW.

Transfusion therapy. In Maull KI,Rodriguez A, Wiles CE, eds. Complicationsin Trauma and Critical Care. Philadel-phia, WB Saunders, 1996.

6. Halsted WS. The Johns Hopkins HospitalReports 1920; 19:71. Cited in Schwartz SI,Green RM. Biology of hemostasis. InSchwartz SI, ed. Techniques of Hemosta-sis. West Berlin, New Jersey, InnovativePublishing Incorporated, 1993.

7. Rotondo MF, Schwab CW, McGonigal, etal. “Damage control”: An approach for im-proved survival in exsanguinating pen-etrating abdominal injury. J Trauma 1993;35:375.

8. Schwartz SI, Moore EE. Local hemostasis.In Schwartz SI, ed. Techniques of Hemo-stasis. West Berlin, New Jersey, InnovativePublishing Incorporated, 1993.

9. Lerner R, Binur NS. Current status of sur-gical adhesives. J Surg Res 1990; 48:165.

10. Ochsner MG, Maniscalco-Theberge ME,Champion HR. Fibrin glue as a hemostaticagent in hepatic and splenic trauma. JTrauma 1990; 30:884.

11. Wagner WR, Pachence JM, Ristich J, et al.Comparative in vitro analysis of topicalhemostatic agents. J Surg Res 1996; 66:100.

12. Gentilello LM. Advances in the manage-ment of hypothermia. Surg Clin North Am1995; 75:243.

Haemostatic Drugs in Trauma and Orthopaedic Practice4

Dr. David RoystonConsultant AnaesthetistRoyal Brompton and Harefield NHS TrustHarefield, Middlesex UB9 6JHUnited Kingdome-mail: [email protected]

Aprotinin is a naturally occurring serineprotease inhibitor. It is found in the mast cellsof all mammalian species as well as many lowerorders of life. Unfortunately, at this time, wedo not understand the true physiologic roleof aprotinin in nature.

What is known is that high doses of thedrug inhibit a number of the inflammatoryprocesses involved with open heart surgeryand also modify the haemostatic system to al-low reductions in bleeding and thus the needfor blood and blood products. The use ofhigh-dose aprotinin therapy followed reportsof the potential benefit of this approach intraumatically injured patients.1 Large-doseaprotinin therapy has been shown to be ex-tremely effective, and safe, in preventingblood loss and the need for blood and bloodproducts in patients undergoing open heartsurgery. The current literature contains morethan 40 reports of randomised placebo-con-trolled studies2,3 that have shown that high-

dose aprotinin therapy reduced drain losses(range, 35%–81%), the total amount of trans-fusions (range, 35%–97%), and the propor-tion of patients requiring transfusions ofblood or blood products (range, 40%–88%).Since the first description of the haemostaticactions of high-dose aprotinin therapy in pa-tients undergoing re-operation4 or high-riskcardiac procedures, this agent has been thestandard of care in this situation and is theonly product licensed for use for this indica-tion in North America.

The aim of this article is to discuss thepotential for this anti-inflammatory andhaemostatic action to benefit patients havingelective orthopaedic and trauma surgery andalso following trauma itself. The article is di-vided into three major sections dealing with

• The use of drugs to prevent bleeding dur-ing elective surgery

• The potential for aprotinin therapy in pa-tients who have sustained trauma

• The potential use of serine protease in-hibitors to prevent certain sequelae oftrauma and surgery of bones, joints, andtendons

Many forms of bone and joint surgery areassociated with a significant risk of bleedingand thus the use of blood and blood prod-ucts.5 A number of systems have been used toreduce this probability. Some of these are al-most unique to orthopaedic surgery, such ascreating a bloodless field by tourniquet ap-plication in limb surgery. In addition, in manycountries, orthopaedic and trauma surgeonshave become the principle users ofpredonated blood and blood product sys-tems. However, there is still significant scopefor the use of other techniques and methods,such as pharmacologic intervention, to inhibitbleeding and minimize the need for blood andblood product transfusions.

Nonemergency Orthopaedic SurgeryThe three most commonly used pharma-

cologic interventions in nonemergency ortho-paedic surgery are tranexamic acid,desmopressin (DDAVP), and aprotinin. Eachof these agents has a relatively unique modeof action, although there is overlap betweensome of the physiologic events produced bythese agents.

Desmopressin is a synthetic analogue ofthe natural hormone argenine vasopressin andhas been shown to increase plasma levels of


Figure 1Total numbers of transfused red cells in patients undergoingorthopaedic surgery for removal of infected bioprosthesis or

resection of tumour with and without aprotinin therapy.Individual patient data are shown and demonstrate an averagethree-fold reduction in need for blood products in patients given

aprotinin therapy. Data are drawn from Capdevila et al.17

factor VIII activity in patients withhemophilia and Von Willebrand’sdisease type I. Desmopressin hadconsiderable support for use incomplex heart surgery, but morerecent data suggest that, overall,this drug provided little, if any,benefit to the patient. Conversely,more recent data suggest that theuse of desmopressin in patientswho are currently taking aspirintherapy has significant benefitduring and after heart surgery.6,7

The results with desmopressin inorthopaedic surgery have beenuniversally poor.8 However, atpresent, there is little informationabout the use of this agent in pa-tients taking aspirin or non-steroi-dal anti-inflammatory drugs(NSAIDs) prior to surgery. This isobviously an area that needs fur-ther investigation.

With regard to tranexamicacid, there are again few data tosupport its use for surgery of cen-tral bones and joints, such as spine and hipsurgery. Some reports indicate a reduction inthe requirement for blood in patients who arehaving knee replacement surgery under tour-niquet.9,10

Aprotinin therapy has been used with ef-fect in a wide variety of surgeries, includingorthopaedic surgery. There is, however, stillonly one report from a randomised placebo-controlled study in patients undergoing pri-mary hip surgery.11 This shows a significantbenefit of high-dose aprotinin therapy to re-duce drain losses and the need for donor bloodand blood products. The dose of aprotininused in this study from Belgium was intendedto be equivalent to the high-dose regimen usedduring cardiac surgery.

A number of other studies have shown aneffect of lower doses of aprotinin on variablessuch as platelet function but without showingconsistent benefit to reduce the requirementfor transfusions.12–16 It appears that a higherdose of aprotinin is needed to ensure reducedblood transfusions than the dose that will havesignificant effects on haemostatic processes.

Similarly, there is evidence that the greaterthe surgical risk, the more benefit the high-dose regimen appears to demonstrate. For ex-ample, a recent article17 showed that aprotinintherapy produced a three-fold reduction in theneed for blood and blood products in patientsundergoing hip replacement because the jointhad become infected or invaded by tumor (Fig.1). This massive reduction in the requirementsfor blood and blood products is similar to theobservations in heart surgery, where the higherthe risk of bleeding, the more obvious is thebenefit of aprotinin therapy. In addition,aprotinin therapy has been used with benefitin patients undergoing spinal surgery.18

Soft Tissue Injury and DisseminatedIntravascular Coagulopathy (DIC)

The use of blood-sparing agents in traumasurgery has potential in soft tissue injury andin patients with intraabdominal (especiallyhepatic) trauma. Severe soft-tissue injury pre-sents a variety of challenges with problemsassociated with the initial event, the subse-quent potential for ischaemia reperfusion in-jury, and the development of a coagulopathyduring resuscitation.

Soft-tissue trauma is associated with therelease of a number of procoagulant media-tors, which can lead to a form of disseminatedintravascular coagulation and haemorrhage.The use of factors to promote haemostasis andprevent bleeding in these circumstances is stillcontroversial. The use of apureantifibrinolytics, such as a lysine analogue, ispotentially lethal in these circumstances. Thesedrugs are therefore contraindicated in the pres-ence of intravascular thrombin generation.Indeed, in animal models, the use of lysineanalogue antifibrinolytics such as tranexamicacid with excess thrombin generation leads tothe death of the animal.19–21

In contrast to the effects of these lysineanalogue agents are the effects of serine pro-tease inhibitors. A number of odious modelsof tissue injury in animals have shown signifi-cantly high early mortality. These models in-clude rotating drum experiments with rats andfracture/sepsis models in sheep. In both theseexperimental models, the use of aprotinintherapy prevented mortality and improvedoutcome.22 A number of animal models to-gether with anecdotes about humans suggestthat aprotinin therapy in addition to heparininhibits the DIC associated with trauma andsepsis.19,21,23

There are also a number of studies from

the early literature in whichaprotinin was administered topatients who had sustainedtrauma, particularly road trafficcrashes. The majority of these pa-pers are found in the German lit-erature. In one multiple centrestudy published in 1976,24 4,686patients were entered into a mul-tiple centre study to investigatethe effects of aprotinin therapyin the treatment of traumaticshock. The dosage used was rela-tively low—approximately 3 mil-lion KIU over a 2-day period—but produced an impressive ben-efit to the patient outcome whenadministered within a few hoursof the trauma. The most signifi-cant benefits of the use ofaprotinin therapy were found inpatients with injuries to the up-per extremity and soft tissues,but there were significant ben-efits following trauma to thelower limb and spine as well. The

study found no benefit of the use of aprotinintherapy in patients with either chest or headinjury (Fig. 2).

Most recently, a number of studies havefocused on aprotinin therapy following bluntliver trauma. These investigations appear to bea natural progression from studies that inves-tigated this therapy in patients having livertransplantation. In both an animal25 and a hu-man26 study, significant benefits were achievedin terms of reduction in bleeding and the needfor donor blood in liver trauma and resection.

These data suggest that aprotinin therapymay be beneficial in certain patients with soft-tissue injury and intraabdominal trauma.

Antiinflammatory ActionsIn addition to the potential haemostatic

benefits of the use of aprotinin in patientsundergoing elective surgery and in those whohave been injured, there are also a number ofother actions of the drug that may benefit thepatient. These are related to its anti-inflam-matory and anticoagulant actions.27

All serine protease inhibitors, includingaprotinin, will inhibit platelet function. Thisis achieved by a number of mechanisms re-lated to the ability to inhibit surface recep-tors and by intracellular metabolism pro-cesses. Indeed, the first use of aprotinintherapy in patients having hip surgery was asan adjunct to the use of heparin to preventdeep venous thrombosis after surgery.12,14 Pre-liminary data from these studies (involvingsmall numbers of patients) suggest a small butstatistically significant effect to reduce theincidence of venous thrombosis. This effectneeds to be investigated in larger groups ofpatients using various forms of antithromboticprophylaxis in addition to aprotinin therapyto determine if there is a significant benefit


Figure 2Percentage mortality in groups of patients with trauma

and tissue injury with (hatched bars) and without (clear bars)aprotinin therapy. Data are drawn from more than

4,000 patients.24 Mortality rates are significantly lower inaprotinin-treated patients for lower-limb and soft-tissueinjury (p <0.01) and for lower-limb trauma and spinalinjury (p <0.05). There was no significant difference

between treatment and no treatment in patientswith predominantly head injury.

in this respect togetherwith a reduction in therequirement for bloodand blood products.Other reports suggestthat the incidence andseverity of pulmonary fatembolism and the fat em-bolus syndrome followingtrauma are reduced sig-nificantly with aprotinintherapy.28,29

A further conse-quence of the use ofaprotinin and its effectson intracellular metabo-lism is inhibition of cer-tain aspects of ischaemiaand reperfusion injury. Inparticular there is consid-erable evidence to showthat aprotinin therapy isassociated with improvedmicrovascular bloodflow.27 This improvedflow together with modi-fications to the metabolicprocess may explain whythere is a significant re-duction in the amount of lactic acid producedafter ischaemia reperfusion in patients under-going hip surgery12,30 and in those undergoingvascular surgery with aorto-bifemoral replace-ment.31

One potential area for the use of aprotininas a treatment after bone surgery is to inhibitthe oedema that occurs after trauma to boneand periosteum. Oedema formation can be as-sociated with considerable discomfort. A num-ber of studies suggest that the local infiltra-tion of aprotinin significantly reduces both theoedema formation and the pain associated withbone surgery. This is especially true for pa-tients requiring maxillofacial surgery and den-tal extraction.32

Finally, there is the potential for the useof aprotinin and other protease inhibitors tobe used prophylactically and in treatment ofpatients with progressive joint destruction orfollowing joint and tendon repair. It is becom-ing increasingly recognised that many of thecells in cartilage to produce proteolytic en-zymes, which may be responsible for chronicjoint destruction.33 More modern methods ofmolecular biology suggest that one of the ma-jor participants in this process is the genera-tion of plasmin from a urokinase plasminogentype activator. This activity is inhibited byaprotinin therapy in tissue culture.33 The ra-tionale for using intra-articular aprotinintherapy is suggested by the observation thatchondrocytes produce a number of proteaseinhibitors of the proteolytic enzymes such asthe plasminogen activators. One of the majorinhibitors thus far categorised from humanchondrocytes is a 6-kD molecule that has re-markable, if not identical, amino acid sequence

homology with aprotinin.34

Preliminary data from human studies sug-gest that the chronic injection of aprotinin intothe joint space is associated with a significantinhibition of progression of disease.35 A simi-lar mechanism may also play a part in the useof aprotinin therapy to prevent adhesion for-mation and fibrosis following tendon repair.36

Summary and ConclusionThe use of aprotinin therapy in sufficiently

high doses is associated with an improvementin haemostatic function and a reduction indrain losses and blood utilisation in patientsundergoing major trauma surgery and ortho-paedic surgery. The anti-inflammatory actionsof aprotinin may also have significant benefitin reducing mortality after soft tissue traumaalone and especially in those traumas associ-ated with increased risk of embolic phenom-ena or intravascular coagulation. Althoughdrugs such as tranexamic acid have value inpatients requiring certain joint replacementsurgeries, their safety in the presence of aprothrombotic state is not proven. Therefore,at this stage, it seems inappropriate to recom-mend these drugs for patients with soft tissuetrauma. The use of drugs such as desmopressinin otherwise routine surgery has, as yet, noproven benefit, although there may be somebenefit to patients who are taking anti-inflam-matory medicines.

In addition to the benefit of reducingbleeding, protease inhibitors can improve pa-tient outcome by reducing ischaemic injuryand the oedema formation that may cause pain.At present, safety data on the use of aprotinintherapy in both open heart surgery and ortho-

paedic/trauma surgery suggest that the benefitsof this drug can be obtained without increas-ing the risk of a thrombotic episode. Whetherthe incidence of thrombosis can be reducedfurther by co-administration of a protease in-hibitor with other antithrombotic prophylaxisremains to be investigated.

References1. Clasen C, Jochum M, Mueller Esterl W.

Feasibility study of very high aprotinin dos-age in polytrauma patients. Prog Clin BiolRes 1987; 236a:175–83.

2. Davis R, Whittington R. Aprotinin: a re-view of its pharmacology and therapeu-tic efficacy in reducing blood loss associ-ated with cardiac surgery. Drugs 1995;49:954–83.

3. Royston D. High-dose aprotinin therapy:a review of the first five years’ experi-ence. J Cardiothorac Vasc Anesth 1992;6:76–100.

4. Royston D, Bidstrup BP, Taylor KM,Sapsford RN. Effect of aprotinin on needfor blood transfusion after repeat open-heart surgery. Lancet 1987; 2:1289–91.

5. Clarke AM, Dorman T, Bell MJ. Blood lossand transfusion requirements in total jointarthroplasty. Ann R Coll Surg Engl 1992;74:360–3.

6. Dilthey G, Dietrich W, Spannagl M, Rich-ter J. Influence of desmopressin acetateon homologous blood requirements incardiac surgical patients treated with as-pirin. J Cardiothorac Vasc Anesth 1993;7:425–30.

7. Laupacis A, Fergusson D. Drugs to mini-mize perioperative blood loss in cardiacsurgery: meta-analyses using perioperativeblood transfusion as the outcome. The In-ternational Study of Peri-operative Trans-fusion (ISPOT) Investigators. Anesth Analg1997; 85:1258–67.

8. Mannucci P. Hemostatic drugs. N Engl JMed 1998; 339:245–53.

9. Benoni G, Fredin H. Fibrinolytic inhibi-tion with tranexamic acid reduces bloodloss and blood transfusion after knee ar-throplasty: a prospective, randomised,double-blind study of 86 patients. J BoneJoint Surg Br 1996; 78:434–40.

10. Hiippala S, Strid L, Wennerstrand M, etal. Tranexamic acid (Cyklokapron) re-duces perioperative blood loss associatedwith total knee arthroplasty. Br J Anaesth1995; 74(5):534–7.

11. Janssens M, Joris J, David JL, Lemaire R,Lamy M. High-dose aprotinin reducesblood loss in patients undergoing total hipreplacement surgery. Anesthesiology1994; 80:23–9.

12. Haas S, Ketterl R, Stemberger A, et al. Theeffect of aprotinin on platelet function,blood coagulation and blood lactate levelin total hip replacement - a double blindclinical trial. Adv Exp Med Biol 1984;167:287–97.


John K. Stene, MD, PhDDepartment of AnesthesiaThe Milton S. Hershey Medical CenterHershey PA 17033 USAe-mail: [email protected]

Although deep venous thrombosis (DVT)and pulmonary embolism (PE) have alwaysbeen major complications of trauma, they haveonly recently become a major concern oftrauma anesthesiologists because modern ef-fective prevention of DVT affects anesthesiapractice.1 Recently developed low-molecular-weight heparins (LMWH) provide very effec-tive prevention of posttraumatic DVT with farfewer bleeding complications than intravenousunfractionated heparin; however, LMWH useis also associated with epidural hematomasfrom epidural catheters.2–8 Thus, at a time whenthe use of continuous regional anesthesia withepidural catheters is shown to reduce trauma

Antithrombotics in Trauma Care: Benefits and PitfallsOther high-risk conditions for DVT include

bed rest for longer than 72 hours; lower-extrem-ity fractures, especially pelvis and acetabularfractures; penetrating venous injuries; head in-juries inducing a low Glasgow Coma Scalescore; family history of thrombi; and a historyof DVT or PE. Also associated with appreciableDVT risk are comorbid conditions such as agegreater than 40; obesity; malignancy; pregnancy,up to 1 month postpartum; use of oral contra-ceptives; and lung operations.7

Virchow noted in the 19th century that DVTwas initiated by one or more of the followingconditions: stasis of blood flow in the deep veinsof the leg, trauma to the endothelial lining of theveins, and increased coagulability of the blood.Because trauma patients are at risk for the entireVirchow’s triad, they are at increased risk of DVTand thus PE. Aside from the obvious vasculartrauma and venous stasis caused by bed rest, theinflammatory state of trauma (e.g., cytokine re-

5

morbidity, anesthesiologists are faced withpatients who receive LMWH prophylaxis forDVT, which may preclude the use of epiduralcatheters.3,5 In this article, the risks and natu-ral history of DVT and recommendations foruse of continuous epidural anesthesia in con-junction with LMWH will be reviewed.

Venous thromboembolic disease—whichincludes both DVT and PE—is a major post-traumatic morbidity and mortality issue.1,2 Di-rect trauma to blood vessels and thrombophiliaassociated with the general inflammatory re-sponse to traumatic injury lead to an increasedincidence of DVT and subsequent pulmonaryembolism (PE).7 The overall incidence of DVTin the North American and European popula-tions is 1 in 1,000. It occurs more frequentlyin older people, obese people, and patientswith traumatic injury. Some injury patternssuch as spinal cord injury are associated witha very high incidence of DVT.

13. Freick H, Reuter HD, Piontek R. [Supple-mentary preoperative prevention ofthromboembolism through the use ofaprotinin in alloplastic hip joint prosthe-sis?] Med Welt 1983; 34:614–9.

14. Ketterl R, Haas S, Lechner F, Kienzle H,Blumel G. [Effect of aprotinin on throm-bocytic function during totalendoprosthesis surgery of the hip.] MedWelt 1980; 31:1239–43.

15. Hayes A, Murphy DB, McCarroll M. Theefficacy of single-dose aprotinin 2 millionKIU in reducing blood loss and its im-pact on the incidence of deep venousthrombosis in patients undergoing totalhip replacement surgery. J Clin Anesth1996; 8:357–60.

16. Utada K, Matayoshi Y, Sumi C, et al.[Aprotinin 2 million KIU reducesperioperative blood loss in patients un-dergoing primary total hip replacement.]Masui 1997; 46:77–82.

17. Capdevila X, Calvet Y, Biboulet P, et al.Aprotinin decreases blood loss and ho-mologous transfusions in patients under-going major orthopedic surgery. Anesthe-siology 1998; 88:50–7.

18. Llau JV, Hoyas L, Higueras J, et al.[Aprotinin reduces intraoperative bleed-ing during spinal arthrodesis interven-tions (letter).] Rev Esp Anestesiol Reanim1996; 43:118.

19. Arnljots B, Wieslander JB, Dougan P,Salemark L. Importance of fibrinolysis inlimiting thrombus formation followingsevere microarterial trauma: an experi-mental study in the rabbit. Microsurgery1991; 12:332–9.

20. Latour JG, Leger Gauthier C, DaoustFiorilli J. Vasoactive agents and produc-

tion of thrombosis during intravascularcoagulation 1: comparative effects ofnorepinephrine in thrombin and adenos-ine diphosphate (ADP) treated rabbits.Pathology 1984; 16:411–7.

21. Moriau M, Rodhain J, Noel H, et al. Com-parative effects of proteinase inhibitors,plasminogen antiactivators, heparin andacetylsalicylic acid on the experimentaldisseminated intravascular coagulationinduced by thrombin. Thromb DiathHaemorrh 1974; 32:171–88.

22. Dwenger A, Remmers D, Grotz M, et al.Aprotinin prevents the development ofthe trauma-induced multiple organ fail-ure in a chronic sheep model. Eur J ClinChem Clin Biochem 1996; 34:207–14.

23. Kolbow H, Barthels M, Oestern HJ, etal. [Early changes of the coagulation sys-tem in multiple injuries and their modi-fication with heparin and Trasylol.] ChirForum Exp Klin Forsch, April 1977, pp119–23.

24. Schneider B. [Results of a field study onthe therapeutic value of aprotinin in trau-matic shock (author’s transl).]Arzneimittelforschung 1976; 26:1606–10.

25. Thomae KR, Mason DL, Rock WA Jr. Ran-domized blinded study of aprotinin in-fusion for liver crush injuries in the pigmodel. Am Surg 1997; 63:113–20.

26. Lentschener C, Benhamou D, Mercier FJ,et al. Aprotinin reduces blood loss in pa-tients undergoing elective liver resection.Anesth Analg 1997; 84:875–81.

27. Royston D. Preventing the inflammatoryresponse to open-heart surgery: the roleof aprotinin and other protease inhibitors.Int J Cardiol 1996; 53(suppl):S11–37.

28. Morl FK, Heller W. [Fat embolism and

proteinase inhibitors.] Langenbecks ArchChir 1969; 325:369–72.

29. Weisz GM, Barzilai A. Fat embolism: phys-iopathology, diagnosis with management.Arch Orthop Unfallchir 1975; 82:217–23.

30. Wendt P, Ketterl R, Haas S, et al. [Postop-erative increase in lactate in total hipendoprosthesis operations: effect ofaprotinin. Results of a clinical double-blind study.] Med Welt 1982; 33:475–9.

31. Horl M, Sperling M, Herzog I, et al. Effectof aprotinin on metabolic changes inblood following aortofemoral bypass op-eration. Eur Surg Res 1985; 17:186–96.

32. Brennan PA, Gardiner GT, McHugh J. Adouble blind clinical trial to assess the valueof aprotinin in third molar surgery. Br J OralMaxillofac Surg 1991; 29:176–9.

33. Ronday HK, Smits HH, Quax PH, et al.Bone matrix degradation by the plasmi-nogen activation system. Possible mecha-nism of bone destruction in arthritis. BrJ Rheumatol 1997; 36:9–15.

34. Rodgers KJ, Melrose J, Ghosh P. Purifica-tion and characterisation of 6 and 58 kDaforms of the endogenous serine protein-ase inhibitory proteins of ovine articularcartilage. Biol Chem 1996; 377:837–45.

35. Capasso G, Testa V. [Infiltrations ingonarthrosis, a therapeutic turning point:the use of a proteinase inhibitor.] ArchPutti Chir Organi Mov 1990; 38:277–84.

36. Komurcu M, Akkus O, Basbozkurt M, etal. Reduction of restrictive adhesions bylocal aprotinin application and primarysheath repair in surgically traumatizedflexor tendons of the rabbit. J Hand SurgAm 1997; 22:826–32.


Dimer is released into the circulation whenintravascular clots are broken down by throm-bolysis.12 Venous duplex Doppler revealsnoncompressibility of flow in the proximaldeep veins of the leg. Venography is the goldstandard of diagnosis for DVT, allowing fillingdefects to be seen after injection of contrastdye in a peripheral limb vein. Approximately2% of DVTs occur in the upper extremities, witha risk of 12% for PE from an upper-extremityDVT.13 Upper-extremity DVT is diagnosed bydetection of obstruction to flow in the deepveins of the shoulder or upper arm.

PE is diagnosed by physical examination,radioisotopic ventilation perfusion (V/Q) scan,spiral computed tomography (CT) of the chest,or pulmonary angiography (Table 3). Physicalsigns and symptoms of PE include pleuriticchest pain, dyspnea, hemoptysis, abnormalbreath sounds, atrial dysrhythmias, hypoxia,and an increase in arterial to end tidal carbondioxide gradient. The presence of a knownDVT increases the probability that these signsand symptoms represent a PE. A perfusiondefect not matched with a simultaneous venti-lation defect demonstrates a PE on radioiso-

topic V/Q scans.11 Spiral CT of the chest en-hanced with contrast medium may reveal a fill-ing defect of a major branch of the pulmonaryartery in patients with PE.10 A pulmonary ar-tery angiogram remains the gold standard fordiagnosing PE by revealing filling defects in thepulmonary artery or its branches.

Treatment of known DVT is anticoagula-tion to prevent further propagation of thethrombus. Therapeutic heparinization is usu-ally performed with intravenous heparin ti-trated to a PTT in the range of 60 to 80 sec-onds. Because patients with one episode ofDVT are at risk for another episode, therapeu-tic heparinization is usually followed by 3months to a lifetime of warfarin or long-termsubcutaneous LMWH.14

Treatment of PE is mostly supportive ofpulmonary function along with administrationof heparin to prevent further clot buildup.Heparin dose is adjusted for PTT of 60 to 80seconds. Embolectomy of the pulmonary ar-tery has been used for “saddle emboli” ob-structing both branches of the pulmonary ar-tery, but embolectomy must be initiated almostimmediately to be effective.

Prevention of DVT relies on a combinationof mechanical methods such as stockings, se-quential compression devices, or foot pumpsand pharmacologic techniques (Table 4). Infe-rior vena caval filters are designed to prevent PEbut have no effect on the development of DVT.

Pharmacologic prevention of DVT hasbeen attempted with aspirin, dextran, heparin,warfarin, and LMWH, and thrombolytics havebeen used to lyse established DVT. Of thesepharmacologic agents, LMWH, low-dose sub-cutaneous heparin, warfarin, and intravenoushigh-dose heparin have proven effective inpreventing DVT.2,4,6,7,15 LMWH is more effica-cious than low-dose heparin, especially in pre-venting PE, and has fewer complications.4,2,15

Because of its ease of use, efficacy, and lowincidence of side effects, LMWH is the drug ofchoice for DVT prophylaxis in trauma patients.Table 5 lists the recommended doses and dos-ing intervals for available LMWH, as well as thecurrent indications for these drugs.

The complications of DVT prophylaxis in-clude bleeding, epidural hematoma, heparin-in-duced osteopenia, heparin-induced thrombocy-topenia (HIT), and warfarin-induced skin necro-sis. LMWH is much less likely to cause osteopeniathan unfractionated heparin, but both are likelyto cause HIT.7 For patients who develop HIT, DVTpropylaxis and treatment of PE can be controlledwith danaparoid or hirudin.

The use of regional anesthesia in traumapatients receiving anticoagulation therapy forDVT prophylaxis requires the compulsive fol-lowing of guidelines to prevent epidural he-matomas.3,5 Table 6 lists recommendations forepidural catheter use in patients receivingLMWH. With the use of sequential compres-sion devices during periods when LMWH can-not be administered safely, prophylaxis againstvenous thromboembolic disease as well as ex-

Table 3. Diagnosis of Pulmonary Embolism

Investigation Comments

History and physical examination Simple; easy to useDyspneaChest painHemoptysisAtrial dysrhythmiasFriction rubHypoxiaDecrease in P

ETCO

2

V/Q scan Useful only to confirm physical findingsSpiral CT of chest Good sensitivity and specificityPulmonary angiography Gold standard

PET

CO2, end-tidal carbon dioxide; V/Q, ventilation-perfusion; CT, computed tomography

Table 1. Inborn Errors of Coagulation

Factor PopulationIncidence

Activated protein C resistance 3%-4% (Factor V Leiden)Hyperhomocysteinemia 5%Protein C deficiency 0.2%-0.4%Protein S deficiency 0.1%Antithrombin deficiency 0.02%

lease) causes a generalized thrombophilia. Thisthromobophilia may cause multiple microvascu-lar thrombin formation and DIC or lead to large-scale thrombus in the deep veins of the leg.

Patients with congenital forms of throm-bophilia are at especially high risk for DVT andPE. These inborn errors of metabolism (Table1) vary in incidence from 3% to 4% in the popu-lation for activated protein C resistance (Fac-tor V Leiden) to 0.02% for antithrombin defi-ciency.7 Although protein C and protein S de-ficiency were involved in the earliest descrip-tions of thrombophilia, they are not nearly ascommon as factor V Leiden deficiency.Hyperhomocysteinemia, which is also a riskfactor for arterial thrombosis such as coronaryarterial occlusion, may be related to folic acid,vitamin B12, and vitamin B6 deficiency and itsincidence is not well defined.9

The main reason to worry about DVT inthe trauma patient is fatal PE. PE occurs symp-tomatically in 30% of patients with a DVT; as-ymptomatic PE increases the overall incidenceof PE to 50% to 60% in patients with DVT. Thedeath rate caused by PE is 50,000 per year inthe United States. Because there is no adequatetreatment for a diagnosed PE and diagnosis isdifficult, prevention of DVT is the most effec-tive measure to reduce the incidence of PE.7,8

Therefore, many trauma patients will receiveDVT prophylaxis, and the use of LMWH is themost effective prophylaxis.

DVT is diagnosed with physical examina-tion, chemical markers of coagulation such asD-dimer, venous duplex Doppler, and venog-raphy (Table 2).10–12 Symptoms and signs ofDVT include pain, a venous cord along the leg,edema distal to the occlusion, and pain in-duced by forceful dorsiflexion of the foot. D-

Table 2. Diagnosis of Deep Venous Thrombosis

Physical examination Simple and inexpensive; needs laboratory confirmationD-Dimer Indicates intravascular coagulationVenous duplex Doppler Useful screening toolVenography Gold standard


cellent analgesia from continuous epiduralanalgesia can be provided to trauma patients.

SummaryThe use of LMWH for reducing the risk of

DVT and PE has gained increasing popularityin trauma patients with pelvic fractures requir-ing operative fixation or prolonged (>5 days)bed rest, in patients with complex lower extrem-ity fractures requiring operative fixation or pro-longed bed rest, and in spinal-cord-injured pa-tients with complete or incomplete motor pa-ralysis. However, the use of LMWH in traumacan be a challenge, necessitating a fine balancebetween bleeding risk and DVT/PE risk. Thereare many unresolved issues concerning the useof antithrombotics in trauma patients, whichrequire further investigation, especially in pa-tients receiving continuous neuraxial analgesia.

5. Vandermeulen EP, Van Aken H, VermylenJ. Anticoagulants and spinal-epidural an-esthesia. Anesth Analg 1994; 79:1165–77.

6. Angelli G, Piovella F, Buoncristiani P, et al.Enoxaparon plus compression stockingscompared with compression stockingsalone in the prevention of venous throm-boembolism after elective neurosurgery.N Engl J Med 1998; 338:80–5.

7. Hyers TM. Venous thromboembolism. AmJ Respir Crit Care Med 1999; 159:1–14.

8. Goldhaber SZ. Pulmonary embolism. NEngl J Med 1998; 339:93–104.

9. Den Heijer M, Koster T, Blom HJ, et al.Hyperhomocysteinemia as a risk factor fordeep-vein thrombosis. N Engl J Med 1996;334:759–62.

10. Remy-Jardin M, Remy J, Deschildre F, etal. Diagnosis of pulmonary embolism withspiral CT: comparison with pulmonaryangiography and scintigraphy. Radiology1996; 200:699–706.

11. PIOPED Investigators. Value of the venti-lation/perfusion scan in acute pulmonaryembolism. JAMA 1990; 263:2753–9.

12. Ginsberg JS, Wells PS, Kearon C, et al. Sen-sitivity and specificity of a rapid whole-blood assay for D-dimer in the diagnosisof pulmonary embolism. Ann Intern Med1998; 129:1006–11.

13. Nemmers DW, Thorpe PE, Knibbe MA,Beard DW. Upper extremity venous throm-bosis. case report and literature review.Orthop Rev 1990; 19:164–72.

14. Kearon C, Gent M, Hirsh J, et al. A com-parison of three months of anticoagulationwith extended anticoagulation for a firstepisode of idiopathic venous thromboem-bolism. N Engl J Med 1999; 340:901–7.

15. Imperiale TF, Speroff T. A meta-analysis ofmethods to prevent venous thromboem-bolism following total hip replacement.JAMA 1994; 271:1780–5.

Table 4. Techniques to Prevent Posttraumatic DVT

Technique Effectiveness

Sequential compression device Effective; easy to use with minimal complicationsFoot pumps Effective; easy to use with minimal complicationsCompression stockings Not effective; cheap; easy to useSubcutaneous heparin Inexpensive; easy to use; high incidence of bleedingLow-molecular-weight heparins Easy to use; highly effective; low incidence

of bleedingVena caval filters Effective; highly invasive; may lead to chronic

venous obstruction

References1. Geerts WH, Code KI, Jay AM, et al. A pro-

spective study of venous thromboembo-lism after major trauma. N Engl J Med1994; 331:1601–6.

2. Geerts WH, Jay RM, Code KI, et al. A com-parison of low-dose heparin with low-mo-lecular-weight heparin as prophylaxisagainst venous thromboembolism aftermajor trauma. N Engl J Med 1996;335:701–7.

3. Horlocker TT, Heit JA. Low molecularweight heparin: biochemistry, pharmacol-ogy, perioperative prophylaxis regimens,and guidelines for regional anesthetic man-agement. Anesth Analg 1997; 85:874–85.

4. Clagett GP, Anderson FA Jr, Heit J, et al.Prevention of venous thromboembolism.Chest 1995; 108(4 suppl):312S–334S.

Table 6.Use of Neuraxial Block inAnticoagulated Patients

1. Do not mix ASA and otheranticoagulants.

2. Bloody tap requires a 24-hourdelay of anticoagulation.

3. Delay needle placement for 12hours after LMWH administration.

4. Delay catheter withdrawal for 12hours after LMWH administration.

5. Delay LMWH dosing until at least2 hours after needle placementand/or catheter removal.

6. Be extremely vigilant in patientswith epidural catheter who arereceiving LMWH.

LMWH, low-molecular-weight heparin

Table 5. Low-Molecular-Weight Heparin Preparations

DVT, deep vein thrombosis; MI, myocardial infarction; PE, pulmonary embolism.*At-risk: age > 40, obesity, general anesthesia >30 minutes, history of malignancy or DVT or

pulmonary embolism.†Danaparoid is an antithrombotic agent with an average molecular weight of ~5,500 daltons.

Drug Indications Subcutaneous Injection Dose

Ardeparin DVT prophylaxis: Knee replacement urgery 50 antiXa U/kg BID

Dalteparin DVT prophylaxis: Hip replacement 2,500-5,000 IU QDsurgery, abdominal surgery(at-risk patients*)

Danaparoid† DVT prophylaxis: Elective hip surgery 750 antiXa units BID

Enoxaparin DVT prophylaxis: Hip and knee Prophylaxis hip and knee:replacement surgery, abdominal surgery 30 mg BID or 40 mg QD(at-risk patients*) Prophylaxis abdominal:

40 mg QD

Inpatient treatment of acute DVT with Inpatient treatment:or without PE, in conjunction with 1 mg/kg BID or 1.5 mg/kg QDwarfarin

Outpatient treatment of acute DVT Outpatient Treatment:without PE, in conjunction with warfarin 1 mg/kg BID

Prevention of ischemic complications of Ischemia: 1mg/kg BIDunstable angina and non-Q wave MI(when used concurrently with aspirin)


Sherwin V. Kevy, MD*Robert Brustowicz, MD***Transfusion Service**Department of AnesthesiaChildren’s HospitalHarvard Medical SchoolBoston, Massachusetts

The experience with many trauma victimshas emphasized the need for a blood sourceother than banked blood. Cardiopulmonarybypass and vascular surgery have establishedunwashed filter autotransfusion as a safe andpractical means to supplement homologousblood usage. During major orthopedic surgi-cal procedures, autotransfusion has been dem-onstrated to reduce blood requirements.

The properties of an ideal autotransfusionsystem include 1) ease of operation, 2) rela-tively low cost, 3) in-line filtration system, 4)simplified anticoagulation, 5) high fluid aspi-ration rate and minimal hemolysis whetherevacuating a pool of blood or surface skim-ming from the operative field, and 6) the abil-ity to concentrate the aspirated blood.

The BloodStream Recovery System (BRS)(Harvest Technologies LLC, Norwell, Massa-chusetts) (Fig. 1) is a surgical suction systemthat automatically senses the pressures re-quired and adjusts from 20 to 40 mmHg dur-ing surface skimming and a maximum of 100mmHg when evacuating a pool of blood. Dur-ing trauma and cardiovascular surgery, the BRScan be utilized as a stand-alone autotransfusionsystem by transferring from the collection res-ervoir to a reinfusion bag that contains an in-tegral 40-micron filter. During orthopedic sur-gery, the BRS can be used as the front-end col-

lection system to cell-washing systems by con-necting the BRS reservoir to the intake line ofthe cell-washing machine.

MethodsThe BloodStream was compared with wall

suction (SS) at vacuum pressures of 100 to 450mmHg during blood pool evacuation and sur-face skimming. A variety of suction wands wereused with both suction systems.

Multiple red cell pools were required forthe studies. The pools are identified by dura-tion of storage, hematocrit, and pertinent con-trol values.

ResultsResults obtained during evacuation of a

pool of blood are shown in Tables 1 and 2.Flow rates obtained with the BRS are more thantwice those obtained with a Yankauer or Fraziersuction want at vacuum pressures of 100 and150 mmHg (Table 1). The latter level is greaterthan that recommended for autotranfusion orintraoperative blood salvage. When theBloodStream serves as the vacuum source, flowrates obtained with Yankauer and Frazier suc-tion wands are comparable to those obtainedwith a wall suction system at 200 mmHg (Tables1 and 2).

Atraumatic Blood Salvage and Autotransfusion in Trauma and Surgery

Fig. 1. BloodStream Recovery System(Harvest Technologies LLC,Norwell, Massachusetts).

Table 1. Comparison of the BloodStream System withWall Suction Using Yankauer (Y) and Frazier (F) Suction Wands

*Mean flow rates observed during evacuation of a pool of blood (volume, 3,000 ml; hema-tocrit, 24%)

†These pressure levels are not recommended for collection for autotransfusion.

Method/Pressure Flow rate (L/min)*SD ± ml/min

BloodStream/100 mmHg 3.74 ± 25

Wall suction/100 mmHgY50515 1.36 ± 20F3310 0.64 ± 14

Wall suction/150 mmHg†Y50515 1.75 ± 15F3310 0.84 ± 19




Table 2. Comparison of the flow rates obtained with the BloodStream, Yankauer (Y),and Fazier (F) Suction Wands when the BloodStream was the Vacuum Source

*Mean flow rates observed during evacuation of a pool of blood (volume, 3,000 ml;hematocrit, 25%).

BloodStream as the Flow Rate (L/min)*Vacuum source (100 mmHg) SD ± ml/min

BloodStream wand 3.60 ± 223

Yankauer 50515 2.00 ± 19

Frazier 3310 0.92 ± 21

6


Charles E. Smith, MD, FRCPCChair, ITACCS Special Techniques/ Equipment CommitteeDepartment of AnesthesiaMetroHealth Medical CenterProfessor of AnesthesiaCase Western Reserve UniversityCleveland OH 44109 USAe-mail: [email protected]

The acutely volume-depleted trauma pa-tient will have cool, moist, pallid, or cyanoticskin, especially at the extremities. Initial evalu-ation of the patient will include an estimate ofblood volume deficit (Table 1), rate of addi-tional blood loss, primary and secondary sur-vey according to ATLS™ principles, and anevaluation of cardiopulmonary reserve and co-existing hepatic or renal dysfunction.1 Themajor goal in resuscitation is to stop the bleed-ing and replete intravascular volume to maxi-mize tissue oxygen delivery. Cardiac output,blood pressure, and oxygenated blood flow tovital organs are important determinants of out-come.

Management priorities in the trauma pa-tient who is bleeding acutely include ventila-tion and oxygenation; measurement of bloodpressure; placement of ECG, pulse oximeter,and capnograph; and establishment or verifi-cation of adequate intravenous (IV) access forinfusion of normothermic fluids (See also

then be used for maintenance of general anes-thesia until the intravascular volume deficit hasbeen corrected and bleeding is under control.Neuromuscular relaxants, benzodiazepines,and other agents are given as clinically indi-cated.2

Fluid OptionsThere is controversy about which IV solu-

tions should be used for resuscitation. Duringhemorrhage, the interstitial space, in additionto the intravascular compartment, is dimin-ished, with a compensatory increase in reab-sorption of fluid into the capillaries. To repletethe intravascular and interstitial compartment,

crystalloid solutions such as iso-tonic 0.9% saline are given ini-tially.3 Glucose-containing solu-tions are avoided because hyper-glycemia aggravates central ner-vous system injury.4,5 LactatedRinger’s solution has an osmo-lality of 273 mOsm/L, which ishypotonic with respect toplasma. Moreover, lactatedRinger’s cannot be used to dilutepacked red blood cells. Thus,0.9% saline is preferred. Colloidsolutions such as hetastarch havebeen shown to be as effective as5% albumin for volume expan-sion. Hetastarch is used after the

Table 3 illustrates an example of the resultsobtained during surface skimming. As one in-creases the vacuum pressure, there is a tendencyfor the Frazier wand to grab onto tissue, reduc-ing the flow rate and increasing red cell dam-age. The BRS has significantly greater flow andresults in significantly less damage to red cells,as determined by plasma hemoglobin levels. Awall suction vacuum pressure of 200 mmHg isrequired to achieve the flow rate obtained withthe BRS at 40 mmHg. However, this results in atwo-fold increase in plasma hemoglobin.

The BRS system can be used forautotransusion of unwashed shed blood dur-ing cardiac and vascular surgery.

ConclusionsThe BloodStream system can rapidly

evacuate a pool of blood at more than twicethe flow rate achieved with a wall suction sys-tem at recommended pressures. TheBloodStream results in significantly less dam-age to red cells as determined by plasma he-moglobin levels compared with wall suction.The data demonstrate that the BloodStreamsystem can be used as a stand-alone mobilesurgical suction system in operating rooms,emergency departments, and trauma centers.

Table 3. Comparison of the BloodStream System with Wall Suctionat Vacuum Pressures Between 100 and 450 mmHg Using Yankauer (Y50515)

and Frazier (F3310) Suction Wands During Surface Skimming

*These pressure levels are not recommended for collection for autotransfusion.

Method/Pressure Flow rate (L/min)* Plasma Hemoglobin (mg/dl)

Blood Stream/20-40 mmHg 270 ± 17 24.3 ± 2

Wall suction/100 mmHgY50515 242 ± 21 32.7 ± 5F3310 165 ± 19 60.3 ± 3

Wall suction/150 mmHg*Y50515 250 ± 19 40.0 ± 3F3310 140 ± 18 53.0 ± 4

Wall suction/200 mmHg*Y50515 268 ± 22 59.3 ± 5F3310 153 ± 19 97.6 ± 7

Wall suction/300 mmHg*Y50515 279 ± 24 63.9 ± 7F3310 143 ± 16 117 ± 11

Wall suction/450 mmHg*Y50515 277 ± 27 109.0 ± 8F3310 138 ± 13 500 ± 27

SECTION III: Transfusion: Clinical Practice

Current Practices in Fluid and Blood Component Therapy in TraumaChapters 10 and 12). Monitoring of tempera-ture, urine output, arterial blood gases, hemo-globin, hematocrit, electrolytes, and param-eters of coagulation is routine in severely in-jured patients. An arterial catheter is usuallywarranted after basic management prioritiesare fulfilled. Consideration is given to place-ment of invasive monitors (e.g., central venouscatheter, pulmonary artery catheter),transesophageal echocardiography, and provi-sion of anesthesia as needed.

For induction of anesthesia in hemody-namically unstable patients, etomidate orketamine is useful.2 Titrated opioids and am-nestic concentrations of volatile agents can

Table 1.Estimation of Blood Volume Deficit in Trauma

Unilateral hemothorax 3,000 ml

Hemoperitoneum with abdominal distension 2,000–5,000 ml

Full-thickness soft-tissue defect 5 cm3 500 ml

Pelvic fracture 1,500–2,000 ml

Femur fracture 800–1,200 ml

Tibia fracture 350–650 ml

Smaller fracture sites 100–500 ml

7


initial phase of resuscitation, which occurs af-ter cessation of bleeding, and is characterizedby interstitial fluid sequestration and maximalweight gain. Large amounts of hetastarch(>15–20 ml/kg) are avoided because of the riskof coagulopathy.6

Delayed Fluid ResuscitationThe use of large quantities of fluids for

immediate resuscitation of victims of penetrat-ing trauma before hemorrhage is controlledby surgical means has been questioned.7 Dis-advantages of immediate fluid resuscitation arethat inserting venous cannulae and giving fluidboluses in the prehospital setting may delaytransfer and surgical intervention, may contrib-ute to secondary hemorrhage by disrupting ordecreasing resistance to flow around a partiallyformed thrombus or by increasing blood pres-sure, may dilute clotting factors, and can con-tribute to hypothermia. In a randomized, pro-spective trial of immediate versus delayed fluidresuscitation in patients with penetratingtrauma, there was increased mortality, lengthof stay, and postoperative complication rate inthe immediate versus the delayed group.7 How-ever, the study was limited to isolated torsoinjuries, and the receiving trauma center hada very rapid response time such that most pa-tients were in the operating room within 1hour of injury. Therefore, results of this studymay not be applicable to other types of inju-ries such as blunt trauma, head injury, andmultiple sites of penetrating trauma or to pa-tients in remote locations requiring long trans-port times.

Red Cell TransfusionThe lower limit of anemia is not estab-

lished in humans. Oxygen delivery is gener-ally adequate with a hemoglobin of 7 g/dl,which corresponds to an oxygen delivery of~500 ml/min in a 70-kg patient, assuming nor-mal cardiac output and hemoglobin/oxygensaturation. In otherwise healthy, normovolemicindividuals, Messmer and colleagues8 demon-strated that tissue oxygenation is maintainedwith hematocrit between 18% and 25%. Theheart and brain are often considered to be mostvulnerable to the effects of anemia. The heartbegins to produce lactic acid at hematocritsbetween 15% and 20%,9 and heart failure gen-erally occurs at hematocrit of 10%.10 Generally,hematocrits between 25% and 30% result inoptimal oxygen delivery, but therapy must beindividualized.11

Factors affecting the transfusion trigger forred cells include the rate and magnitude ofblood loss; degree of cardiopulmonary reserve;presence of atherosclerotic disease of thebrain, heart, and kidneys; and oxygen con-sumption.11 If the patient has lost largeamounts of blood and is in class III or IV shock(see table on page 4), blood administration isrequired.2 Available options are type O-nega-tive, type-specific, typed and screened, or typedand cross-matched packed red blood cells.

Whole blood is not available at the author’sinstitution. The initial choice will depend onthe degree of hemodynamic instability. Oneunit of packed red blood cells will usually in-crease the hematocrit by ~3% or the hemo-globin by 1 g/dl in a 70-kg non-bleeding adult.

Type O-negative red cells have no majorantigens and can be given reasonably safely topatients with any blood type. Unfortunately,only 8% of the population has O-negativeblood, and blood bank reserves of O-negative,low-antibody-titer blood are usually very low.For this reason, O-positive red cells are fre-quently used. This is a reasonable approach inmales but may be a problem in childbearing-aged females who are Rh negative.

If 50% to 75% of the patient’s blood vol-ume has been replaced with type O blood (e.g.,~10 units of red cells in an adult patient), oneshould continue to administer type O red cells.Otherwise, risk of a major cross-match reac-tion increases since the patient may have re-ceived enough anti-A or anti-B antibodies toprecipitate hemolysis if A, B, or AB units aresubsequently given.2

Obtaining type-specific red cells requires5 to 10 minutes in most institutions, and tem-porizing measures can sometimes be em-ployed to gain the necessary time. At our insti-tution, we use a tube system to deliver bloodsamples and products to and from the operat-ing room or trauma resuscitation suite. Thissystem significantly reduces delays and “lost”samples. The use of type-specific red cells ispreferred over O-negative and the risk of ahemolytic transfusion reaction is very low.12 Ifone can wait 15 minutes, typed and screenedblood should be available. When blood is typedand screened, the patient’s blood group isidentified and the serum is screened for majorblood group antibodies. A full cross matchgenerally requires about 45 minutes and in-volves mixing donor cells with recipient serumto rule out any antigen/antibody reactions.13

Coagulation Factors and PlateletsThe primary cause of bleeding after

trauma is surgical, while the second leadingcause is hypothermia. In the past, coagulopathyafter massive transfusion with whole blood wasusually caused by dilutional thrombocytope-nia. However, this is not necessarily the casewith red cells reconstituted in 0.9% saline.Murray et al have shown that microvascularbleeding and clinical evidence of coagulopathyoccurred in the setting of massive transfusionand was associated with decreased coagulationfactor levels, decreased fibrinogen, elevatedprothrombin times and platelet counts>100,000/µl.14,15 Moreover, administration offresh frozen plasma corrected the microvascu-lar bleeding. Two units of fresh frozen plasma(10–15 ml/kg) will achieve 30% factor activityin most adults. Coagulation factor deficienciesmay be present due to other causes such aspreexisting defects or disseminated intravas-cular coagulopathy resulting from tissue injury.

Cryoprecipitate may then be indicated to cor-rect specific factor deficiencies. It should benoted that 1 unit of fresh whole blood or 1unit of single-donor apheresis platelets also hassimilar factor levels as 1 unit of fresh frozenplasma. Similarly, 4 to 5 multiple donor plate-let units have similar factor levels as 1 unit offresh frozen plasma because the platelets aresuspended in plasma.

Dilutional thrombocytopenia in thetrauma patient also occurs. Leslie and Toy16

showed that platelet count was reduced to<50,000/µl after administration of 20 units ofred cells. Platelet transfusions are usually in-dicated in the presence of clinical bleeding anda platelet count <75,000 to 100,000/µl. Plate-let concentrates are stored at room tempera-ture and contain about 70% of the platelets ina unit of blood. One unit of platelets, equiva-lent to 50 ml, increases the platelet count inan adult by 5,000 to 10,000/µl, and is giventhrough a 170-µ filter.

Hypertonic FluidsLesser amounts of hypertonic fluids, as

opposed to isotonic fluids, can also providerapid volume expansion and improved hemo-dynamics and have the added advantage ofdecreasing tissue edema, intracranial pressure,and brain water. These hypertonic solutionsresult in an osmotic translocation of extracel-lular and intracellular water. Because the in-travascular half-life of hypertonic saline is simi-lar to that of isotonic saline, these fluids canbe combined with colloid solutions such ashetastarch or dextran to prolong their plasmavolume expansion effects. Hypertonic salinehas been associated with bleeding, hemody-namic deterioration, and increased mortalityin animal studies of uncontrolled hemorrhagicshock.17 Further, it does not improve cerebraloxygen delivery after head injury and mildhemorrhage in animals.18 Nonetheless, hyper-tonic saline combined with 6% hydroxyethylstarch has been shown to improve neurologicfunction and cerebral perfusion pressure inpatients with traumatic brain injury.18a Thishypertonic fluid solution is currently used inAustria for resuscitation of all head-injured andmajor trauma patients in the field (Mauritz W,personal communication).

Endpoints of ResuscitationBlood and fluid resuscitation is continued

until perfusion has been improved and organfunction has been restored. Manifestations ofimproved cardiac output include improvedmental status; increased pulse pressure; de-creased heart rate; increased urine output;resolution of lactic acidosis and base deficit;brisk capillary refill; and improvement in oxy-gen delivery, oxygen consumption, and cen-tral venous or pulmonary artery oxygen satu-ration (Table 2).19

Blood and Fluid WarmersFluid and blood resuscitation of the


trauma patient is best accomplished with large-gauge intravenous catheters and effective fluidwarmers with high thermal clearances.20 Be-cause alterations in red cell integrity are notapparent until 46oC,21 fluid warmers with setpoints of 42oC are now commonly used. Coun-tercurrent water and other fluid warmers us-ing 42oC set points will not damage red cells,will result in consistently warmer fluid deliv-ery, and will allow the clinician to maintainthermal neutrality with respect to fluid man-agement over a wide range of flow rates.22

Complications of Transfusion Therapy

Impaired Oxygen Release from HemoglobinThe ability of the red blood cell to store

and release oxygen is impaired after storage.The erythrocytic levels of 2,3-diphosphoglyc-eric acid decrease both with CPD and CPDA-1stored blood. Low levels of 2,3-diphosphoglyc-eric acid will shift the blood’s oxygen dissocia-tion curve to the left, and the red cell will havea higher affinity for oxygen at physiologic PO

2

and will release less oxygen at a given tissuePO2.

23 Impaired oxygen release from hemoglo-bin can be minimized by warming all bloodand by avoiding factors that shift the O

2 disso-

ciation curve to the left, e.g., hypothermia.

Dilutional CoagulopathyMost coagulation factors are stable in

stored whole blood, except factors V and VIII.13

These factors gradually decrease to 15% and50% of normal, respectively, after 21 days ofstorage. However, most centers today usepacked red blood cells and not whole bloodduring massive transfusion. Microvascularbleeding and clinical evidence of coagulopathycan occur in the setting of massive transfusionwith 1 blood volume and are associated withdecreased levels of Factor V, VIII, and fibrino-gen and increased prothrombin times.14–16 Mi-crovascular bleeding in this case can be treatedappropriately with fresh frozen plasma.Dilutional thrombocytopenia is a cause of hem-orrhagic diathesis after 1.5 to 2.0 blood vol-umes have been transfused. This correspondsto ~15 to 20 units of red cells in an adult

trauma victim.16 In the author’s opinion, theplatelet count should be monitored and main-tained at or greater than 75,000 to 100,000/µlto achieve adequate surgical hemostasis. It isadvisable that prothrombin time, activatedpartial thromboplastin time, fibrinogen, andfibrin degradation products be monitored be-cause deficiencies may be present due to dilu-tion, preexisting defects, or disseminated in-travascular coagulopathy.24 Point-of-care test-ing and rapid reporting of coagulation test re-sults should be used to guide decisions regard-ing administration of fresh frozen plasma,platelets, or cryoprecipitate.

HypothermiaThe adverse effects of hypothermia in the

trauma patient include major coagulation de-rangements, peripheral vasoconstriction, meta-bolic acidosis, compensatory increased oxygenrequirements during rewarming, and impairedimmune response.25–27 Standard coagulationtests are temperature corrected to 37oC andmay not reflect hypothermia-inducedcoagulopathy.28-30 Hypothermia impairs coagu-lation because of slowing of enzymatic ratesand reduced platelet function. Hypothermiacan cause cardiac dysrhythmias and even car-diac arrest due to electromechanical dissocia-tion, standstill, or fibrillation, especially withcore temperatures below 30oC. Hypothermiaalso impairs citrate, lactate, and drug metabo-lism; increases blood viscosity; impairs redblood cell deformability; increases intracellu-lar potassium release; and causes a leftwardshift of the oxyhemoglobin dissociation curve.A mortality of 100% has been reported intrauma patients whose body temperature fellbelow 32oC, regardless of severity of injury,degree of hypotension, or fluid replacement.31

The importance of fluid warming cannotbe underestimated in the trauma patient. Itrequires 16 kCal of energy to raise the tem-perature of 1 liter of crystalloid infused at 21oCto body temperature and 30 kCal to raise thetemperature of cold 4oC blood to 37oC. Infu-sion of 4.3 liters of crystalloid at room tem-perature to an anesthetized adult trauma pa-tient who cannot increase heat production can

result in a decrease of 1.5oC in core tempera-ture. Similarly, infusion of 2.3 liters of red cellscould result in a core temperature decrease ofbetween 1 and 1.5oC.32,33 Since the thermalstress of infusing fluids at normothermia isessentially zero, it follows that use of fluid-warming devices effective at delivering normo-thermic fluids to the patient at clinically rel-evant flow rates permits more efficient rewarm-ing of hypothermic trauma patients using othermethods such as the patient’s own metaboli-cally generated heat, or externally providedheat such as convective warming.22

Citrate Intoxication, Hyperkalemia, andAcid–Base Abnormalities

Blood is stored in citrate phosphate dex-trose with adenine or adsol at 4oC. Citrate in-toxication is caused by acutely decreased se-rum levels of ionized calcium, which occursbecause citrate binds calcium.34 Administrationof calcium is warranted during massive trans-fusion if the patient is hypotensive and mea-sured serum ionized serum calcium is low orlarge amounts of blood are infused rapidly (50to 100 ml/min). Ionized serum calcium levelswill usually return to normal when hemody-namic status is improved. The potassium levelin stored blood rises with length of storage andcan be as high as 78 mmol/L after 35 days. Thepotential for clinically important hyperkalemiastill exists in patients receiving blood adminis-tered at rates >120 ml/min35 and in patientswith severe acidosis. Monitoring the ECG forsigns of hyperkalemia is always warranted, andtreatment of hyperkalemia with calcium chlo-ride, bicarbonate, glucose, and insulin may belife saving.

The pH of bank blood decreases to about6.9 after 21 days of storage because of accu-mulation of CO

2, lactic acid, and pyruvic acid

by red blood cell metabolism. Thus, the aci-dosis seen in stored blood is partly respiratoryand partly metabolic. The respiratory compo-nent is of little consequence with adequatepatient ventilation. The metabolic componentis not usually clinically significant. It is unwiseto administer sodium bicarbonate on an em-piric basis, because there is already a pool ofbicarbonate generated from the metabolism ofcitrate, which is present in large quantities instored blood.

Hemolytic Transfusion ReactionsImmediate reactions occur from errors

involving ABO incompatibility. More than halfof these errors happen after the blood has beenissued by the blood bank, which highlights theimportance of verifying and identifying eachand every donor unit for recipient compatibil-ity. Intravascular hemolysis occurs when recipi-ent antibody coats and immediately destroysthe transfused red cells. Classic signs ofhemolytic transfusion reaction are masked bygeneral anesthesia. The only evidence may behemoglobinuria, hypotension, and a bleedingdiathesis. Treatment is supportive and involves

Table 2. Resuscitation Endpoints Within the First 24 Hours After Trauma

Parameter Value

Oxygen delivery index >600 ml/min/m2

Oxygen consumption index >150 ml/min/m2

Mixed venous oxygen tension >35 mmHg

Mixed venous oxygen saturation >65%(central venous or pulmonary artery)

Base deficit >-3 mmol/L

Lactate <2.5 mmol/L

Adapted from Ivatury RR, Simon RJ. Assessment of tissue oxygenation (evaluation of theadequacy of resuscitation). In Ivatury RR, Cayten CGC, eds. The Textbook of PenetratingTrauma. Baltimore, Williams & Wilkins, 1996, pp 927–938.


stopping the transfusion and maintaining sys-temic and renal perfusion.

MicroaggregatesMicroaggregates begin forming after ap-

proximately 2 days of blood storage. Duringthe first 7 days, microaggregates are mostlyplatelets or platelet debris. After the first week,the larger fibrin–white blood cell–platelet ag-gregates begin to accumulate.36 Whether thesemicroaggregates contribute to lung dysfunc-tion during blood transfusion and whetherthey need to be removed by micropore filtersis controversial.

InfectionHepatitis C accounts for more than 90%

of posttransfusion hepatitis. Every year, at least2,600 patients develop cirrhosis as a result ofhepatitis after blood transfusions.37 Each unitof fresh frozen plasma or platelets has the samerisk of infection as a unit of packed red cells.Recent estimates of infectious rates per unittransfused include hepatitis C, 1:103,000;hepatitis B, 1:63,000; HIV, 1:493,000; and HTLVI or II, 1:641,000.38 New screening tests usingnucleic acid/genomic amplification techniqueswill shorten the window period and reducethe risk for these viruses even further. The riskper unit for Yersinia, malaria, babesiosis, andChagas is estimated at <1:1,000,000. Othertypes of infectious diseases such as toxoplas-mosis and cytomegalovirus, Epstein-Barr virus,and bacterial infections may also be transmit-ted via transfused blood and blood products.The risk of bacterial contamination per unit ofrandom donor platelets is 1:2,500.

Transfusion-Induced Immunosuppression(See also Chapters 8 and 9)

Blood transfusion therapy is also associ-ated with allosensitization, immunosuppres-sion, and an increased incidence of postop-erative infections.39 These effects may be me-diated by reduced lymphocyte function, down-regulation of macrophage function, and alteredcytokin production and activity. Strategies toreduce the risk of immunomodulation includethe use of third-generation leukocyte filters,lower transfusion trigger, red cell salvage, andblood substitutes (Table 3).11,40–42 It is antici-pated that new devices for autotransfusion,together with the introduction of hemoglobin-based red cell substitutes, will dramatically al-ter the current approach to fluid and bloodcomponent therapy in trauma.

SummaryThe bleeding trauma patient requires

rapid evaluation and treatment to ensure ad-equate tissue perfusion and successful out-come. Resources such as thermally efficientfluid warmers, effective transfusion services,and rapid availability of coagulation tests arepractical aspects of trauma resuscitation thatdeserve priority. Preventing hypothermia andrecognizing other complications of massive

transfusion, as well as following trends in vitalsigns, urinary output, central venous pressures,and arterial and central venous blood gas analy-sis, are of vital importance to managing pa-tients with hemorrhagic shock.

References1. Stene J, Smith CE, Grande CM. Evaluation

of the trauma patient. In Longnecker DE,Tinker JH, Morgan GE, eds. Principles andPractice of Anesthesiology, 2nd ed. Phila-delphia, Mosby-Yearbook, 1997.

2. Grande CM, Smith CE, Stene J. Anesthe-sia for trauma. In Longnecker DE, TinkerJH, Morgan GE, eds. Principles and Prac-tice of Anesthesiology, 2nd ed. Philadel-phia, Mosby-Yearbook, 1997.

3. Rackow EC, Falk JL, Fein IA et al. Fluidresuscitation in circulatory shock: a com-parison of the cardiorespiratory effects ofalbumin, hetastarch, and saline solutionsin patients with hypovolemic and septic

shock. Crit Care Med 1983; 11:839.4. Lam AM, Winn HR, Cullen BF, et al. Hy-

perglycemia and neurological outcome inpatients with head injury. J Neurosurg1991; 75:545.

5. Michaud LJ, Rivara FP, Longstreth WT, etal. Elevated initial blood glucose levels andpoor outcome following severe brain inju-ries in children. J Trauma 1991; 31:1356.

6. Wilson RF. Blood replacement. In WilsonRF, Walt AJ, eds. Management of Trauma.Pitfalls and Practice, 2nd ed. Baltimore,Williams & Wilkins, 1996.

7. Bickell WH, Wall MJ, Pepe PE, et al. Imme-diate versus delayed fluid resuscitation forhypotensive patients with penetrating torsoinjuries. N Engl J Med 1994; 331:1105.

8. Messmer K, Sunder-Plassmann L, Jesch F,et al. Oxygen supply to the tissues duringlimited normovolemic hemodilution. ResExp Med 1973; 159:152.

9. Jan KM, Heldman J, Chien S. Coronary

Table 3.Clinical Strategies to Reduce Complications of Transfusion Therapy

DIC, disseminated intravascular coagulation

Complication Clinical Strategies to Reduce Complication

Impaired oxygen release from Warm all blood. Avoid alkalosis. Maintainhemoglobin normothermia (core temperature 36-37°C)

Dilutional coagulopathy Fresh frozen plasma for PT>1.5 x normal andclinically excessive bleeding. Platelets forthrombocytopenia <75,000/µl and clinicallyexcessive bleeding.

Hypothermia Warm all IV fluids and blood. Warm room>28°C. Convective warming. Humidify allinspired gases.

Decreased ionized calcium Treat with calcium chloride, 20 mg/kg, insetting of massive transfusion and hypotension

Hyperkalemia Monitor ECG and treat with calcium chloride,20 mg/kg, if hemodynamically significant.Otherwise, monitor and treat with glucose andinsulin and/or bicarbonate.

Hemolytic transfusion reaction Check and recheck every donor unit. Onceoccurred, stop transfusion and maintainsystemic perfusion and renal blood flow.Alkalinize urine. Watch for DIC. Send suspectedunit to blood bank for crossmatch.

Infection Lower transfusion trigger. Red cell salvage.Avoid indiscriminate platelet transfusions.Oxygen-carrying red blood cell substitutes.

Transfusion-induced Lower transfusion trigger. Red cell salvage,immunosuppression oxygen-carrying red blood cell substitutes.

Third-generation leukocyte filters.


hemodynamics and oxygen utilization af-ter hematocrit variations in hemorrhage.Am J Physiol 1980; 239:H326.

10. Varat MA, Adolph RJ, Fowler NO. Cardio-vascular effects of anemia. Am Heart J1972; 83:415.

11. ASA Task Force. Practice guidelines forblood component therapy. Anesthesiology1996; 84:732.

12. Gervin AS, Fischer RP. Resuscitation oftrauma patient with type-specificuncrossmatched blood. J Trauma 1984;24:327.

13. Miller RD. Transfusion therapy. In MillerRD, ed. Anesthesia, 4th ed. New York,Churchill Livingstone, 1994.

14. Murray, DJ, Pennell BJ, Weinstein SL, OlsonJD. Packed red cells in acute blood loss:dilutional coagulopathy as a cause of sur-gical bleeding. Anesth Analg 1995; 80:336.

15. Murray DJ, Olson J, Strauss R, Tinker JH.Coagulation changes during packed redcell replacement of major blood loss. An-esthesiology 1988; 69:839.

16. Leslie SD, Toy PT. Laboratory hemostaticabnormalities in massively transfused pa-tients given red blood cells and crystal-loid. Am J Clin Pathol 1991; 96:770.

17. Gross D, Landau EH, Klin B, et al. Quanti-tative measurement of bleeding followinghypertonic saline therapy in “uncon-trolled” hemorrhagic shock. J Trauma1989; 29:79.

18. Dewitt DS, Prough DS, Deal DD, et al.Hypertonic saline does not improve cere-bral oxygen delivery after head injury andmild hemorrhage in cats. Crit Care Med1996; 24:109.

18A.Hartl R, Ghajar J, Hochleuthner H, MauritzW. Treatment of refractory intracranialhypertension in severe traumatic braininjury with repetitive hypertonic/hyperoncotic infusions. Zentrabl Chir1997; 122:181–5.

19. Scalea TM, Hartnett RW, Duncan AO, et al.Central venous oxygen saturation: a use-ful clinical tool in trauma patients. JTrauma 1990; 30:1539.

20. Patel N, Smith CE, Pinchak AC. Clinicalcomparison of blood warmer perfor-mance during simulated clinical condi-tions. Can J Anaesth 1995; 42:636.

21. Uhl L, Pacini DG, Kruskall MS. The effectof heat on in vitro parameters of red cellintegrity. Transfusion 1993; 33:60S.

22. Patel N, Knapke D, Smith CE, et al. Simu-lated clinical evaluation of conventionaland newer fluid warming devices. AnesthAnalg 1996; 82:517–524.

23. Valeri CR, Collins FB. Physiologic effects of2,3-DPG-depleted red cells with high affin-ity for oxygen. J Appl Physiol 1971; 31:823.

24. Sohmer PR, Scott RL. Massive transfusion.Clin Lab Med 1982; 2:21.

25. Sessler DI. Perianesthetic thermoregula-tion and heat balance in humans. FASEB J1993; 7:638–644.

26. Smith CE, Patel N. Hypothermia in adulttrauma patients: anesthetic consider-ations. Part I: Etiology and pathophysiol-ogy. Am J Anesthesiol 1996; 23:283.

27. Smith CE, Patel N. Hypothermia in adulttrauma patients: anesthetic consider-ations. Part II: Prevention and treatment.Amer J Anesthesiol 1997; 24:29.

28. Reed RL, Johnston TD, Hudson JD, FischerRP. The disparity between hypothermiccoagulopathy and clotting studies. JTrauma 1992; 33:465.

29. Reed RL, Bracey AW, Hudson JD, et al. Hy-pothermia and blood coagulation: disso-ciation between enzyme activity and clot-ting factor levels. Circ Shock 1990; 32:141.

30. Valeri CR, MacGregor H, Cassidy G, et al.Effects of temperature on bleeding time andclotting time in normal male and femalevolunteers. Crit Care Med 1995; 23:698.

31. Jurkovich GH, Greiser WR, Luterman A etal. Hypothermia in trauma victims: anominous predictor of survival. J Trauma1987; 27:1019.

32. Gentilello LM, Moujaes S. Treatment ofhypothermia in trauma victims: thermo-dynamic considerations. J Intensive CareMed 1995; 10:5.

33. Mendlowitz M. The specific heat of humanblood. Science 1948; 107:97.

34. Kahn RC, Jasco HD, Carlon GC et al. Mas-

sive blood replacement: correlation of ion-ized calcium, citrate, and hydrogen ion con-centration. Anesth Analg 1979; 58:274.

35. Insalaco SJ. Massive transfusion. Lab Med1984; 15:325.

36. Arrington P, McNamara JJ. Mechanism ofmicroaggregate formation in stored blood.Ann Surg 1974; 179:146.

37. Zuck TF, Sherwood WC, and Bove JR. Areview of recent events related to surro-gate testing of blood to prevent non-A,non-B posttransfusion hepatitis. Transfu-sion 1987; 27:203.

38. Schreiber GB, Busch MP, Kleinman SH,Korelitz JJ. The risk of transfusion-trans-mitted viral infections. The Retrovirus Epi-demiology Donor Study. N Engl J Med1996; 334:1685–90.

39. Landers DF, Hill GE, Wong KC, Fox IJ.Blood transfusion-induced immunomo-dulation. Anesth Analg 1996; 82:187.

40. Lane TA Leukocyte reduction of cellularblood components. Arch Pathol Lab Med1994; 118:392.

41. Kevy SV et al. Evaluation of a newatraumatic surgical suction system (ab-stract). Proceedings of the 10th AnnualTrauma Anesthesia and Critical Care Sym-posium, Baltimore, 1997.

42. Cohn SM. Is blood obsolete? J Trauma1997; 42:730.

David T. Porembka, Do, FCCM, FCCPAssociate Professor of Anesthesia and SurgeryAssociate Director of Surgical Intensive CareUniversity of Cincinnati Medical CenterCincinnati, Ohio, USA

The administration of blood and its com-ponents can be life-saving, particularly dur-ing resuscitation in trauma patients whenblood loss can be severe enough to result incellular hypoxia.1 Also, during other criticalillness such as systemic inflammatory re-sponse syndrome, especially if the patient isseptic with significant acute lung injury, bloodis administered to augment oxygen deliveryto avoid cellular hypoxia and lactate produc-tion.2 Even though there are risks followingblood transfusions, the benefits appear to beinsurmountable3 (Table 1). In spite of this, therisks of infection, especially from HIV, havetaken center stage even in the lay press. Thus,the immunologic effects of transfusion havenot gained the attention deserved. Nonethe-less, in certain disciplines—hematology, criti-cal care medicine, oncology, surgery, and par-ticularly transplantation—have appreciatedthe immunologic potential from its use. Thispresentation will discuss the basics of immu-nology, concentrating on the immunologicconsequences of transfusions, the clinical and

Immunomodulatory Effects of Transfusion

Table 1. Risks of Transfusions

Reactions Frequency: unit

Febrile (FNHTR) 1–4:100Allergic 1–4:100Delayed hemolytic 1:1,500Acute hemolytic 1:12,000Fatal hemolytic 1:100,000Anaphylactic 1:150,000

Infections

Hepatitis C 1:103,000Hepatitis B 1:200,000HIV-1 1:490,000HIV-2 UnknownHTLV-I (II) 1:641,000Malaria 1:4,000,000

Miscellaneous

RBC allosensitization 1:100HLA allosensitization 1:10Graft vs. host disease Rare

From Dzieczkowski and Anderson.3

8


animal studies affecting tumor recurrence,and infection.

T-Cell Recognition and ActivationT-cell recognition of an antigen with T-cell

activation is key in the initiation of rejectionand/or tolerance of foreign tissue. Typically, Tcells require two signals for activation. The firstoccurs when an antigen is processed into pep-tides by an antigen-presenting cell (APC) andloaded into the groove of a major histocom-patibility complex (MHC) molecule. The anti-gen is then presented to the T cell, which isrecognized in the context of self-MHC (Fig. 1).The second signal occurs when the T cell re-ceives stimulation by a cytokine or by the in-teraction of the T cell with surface moleculesof an APC. Numerous cytokines (interleukins,alpha-tumor necrosis factor, and interferon)are involved in this process as well as cell sur-face receptors, adhesion molecules, and lym-phocyte functioning antigen.4–6 Other signifi-cant cell surface molecules are the CD3 com-plex and CD45. The former is associatednoncovalently with the T-cell receptor on ma-ture T cells and is a target for OKT3, whereasthe latter does not have a known ligand andallows continued activation of the T cell.7

Generally, T cells recognize antigens pre-sented as short peptides that are bound in theMHC groove. Allo-MHC molecules stimulate agreater response (in vitro mixed lymphocytesresponse and cytotoxic T-lymphocyte assay)than antigens that are not foreign.8–10 The path-ways for these alloreactivities are both indirectand direct.11–13 In the direct pathway of alloan-tigen presentation, the T cells recognize intactdonor MHC molecules on the surface of thedonor APC. This pathway may be responsiblefor early acute rejection of grafts.12 Early in thecare of these patients, radiation and other im-mune modulation strategies were used to af-fect this pathway directly by removing or de-stroying these graft leukocytes.13 The exactmechanism is not known and is, without adoubt, multifactorial. In the indirect pathway,T cells recognize processed donor allo-MHCbound to and presented in the context of self-MHC molecules on the surface of self-APC. Thispathway is normally associated with a nomi-nal antigen.

History of Donor-Specific TransfusionAs early as 1953, Billingham and associ-

ates demonstrated white blood cells as im-mune modulators when neonatal mice of onestrain injected with blood from another sub-sequently accepted skin grafts from the immu-nizing strain. This effect was long term only inthe neonatal mice, not in the adults.14 The firstsolid organ transplantation (kidney) was per-formed in 1954 between monozygotic twins.The success was probably related to matchingof the ABO blood type with compatibility ofthe (MHC) antigens, not from immune sup-pression. (The complexity of the immune sys-tem was not well understood during this era.)

However, successful transplantation of kidneysfrom HLA-mismatched donors was not possible(1963) until the advent of immunosuppressiveagents, prednisone and azathioprine.15,16 Theimmunosuppressive agents had to be contin-ued to ensure “acceptance” of the foreign tis-sue or organ. In addition, early in transplanta-tion, efforts were directed to minimize expo-sure to or sensitization from transfusions.However, in 1972 two animal studies chal-lenged that premise. Jenkins et al revealed thattransfusions administered prior to cardiac al-lografting improved survival of transplantedhearts in rats.17 Separately, Fabre and associ-ates showed that rejection of the transplantedkidney in rats can be diminished bypretransplant transfusions.18

Possibly realizing these attributes, Newtonand Anderson, in 1973, attempted to manipu-late the immune response to renal allograftsof four patients with donor-specific peripherallymphocyte buffy-coat transfusion from theirpotential living related donor over an extendedtime (22 to 66 days). Allosensitization did notensue. Critics believed that the addition of aza-thioprine contributed to the allograft’s successrather than the administration of blood.19 Sub-sequent to this new era of cadaveric donor

renal transplantation and at the same time,Opelz et al, following the success in animalmodels, provided evidence (by reviewing trans-plant data from multiple centers) in humansthat blood transfusion prior to renal transplan-tation improved renal allograft survival. Com-pared with patients who did not receive bloodtransfusions, the transfused patients (>5 trans-fusions) had a higher survival rate of the renalallografts, approaching 20%. Interestinglythough in this study, this effect appeared tohave a dose-response relationship.18,20,21 Eventhough this seminal publication was retrospec-tive, recently there appears to be more directevidence for this response.22 In 1979, Cochrumand colleagues used pretransplantation-di-rected donor-specific whole blood in patientswith renal failure. In strong mixed lymphocyteculture-responsive, one haplotype-mis-matched, and living-related donor transplants,directed transfusions improved survival up to90%. This rate is not that different from that inHLA-identical siblings.23,24 Following this suc-cess, there was equivalent survival in patientswith two haplotype-mismatched, related andunrelated donor–recipient combinations.25,27

From these studies and others, the presenceof leukocytes and one shared HLA-DR antigenwithin the transfusions are sufficient enoughfor optimal immunosuppression.20,28 Overall,there is sufficient evidence documenting thattransfusions prior to solid organ transplanta-tion improves survival and reduces the inci-dence of rejection.

Although the precise mechanisms in-volved in tolerance and sensitization are notcompletely understood, the laboratory findingshave been consistent29 (Table 2). Generally,blood transfusions induce predictable immuneresponses stimulating alloantibody productionwhen exposed to red cell, white cell, and plate-let alloantigens.30–33 Investigations have shownthe development of Fc receptor-blocking fac-tors, lymphocyte activation, lymphocyte sub-population changes, and down regulation ofAPC after transfusion34 (Table 3). These results

Table 2.Possible Mechanisms ofTransfusion-Associated

Immunomodulation

AnergyToleranceCytokines released during blood storageIron-mediated immune suppressionSuppressor cell network inhibitionAnti-idiotypic and anti-clonotypic anti-bodiesClonal deletion

From Brennan et al.29

Figure 1Representation of theantigen-presenting cell(APC) interacting with themajor histocompatibilitycomplex (MHC) molecule.In addition this interactionis presented to the T cell,which acts with the servicemolecules and APC.


are not reproduced in the neonate as com-pared with the adult immune system. Neonateswho received washed and irradiated bloodfailed to exhibit similar effects seen in adultrecipients.35

Tumor RecurrenceThere appears to be a beneficial effect of

blood transfusions on the immune systems insolid-organ transplantation, but there is adown side—the reemergence of cancer cellsin patients with neoplastic disease. However,the results are divergent, ranging from stimu-lation of tumor growth, suppression of growth,and varied to no response to the tumor cells.In 1982, Burrows and colleagues retrospec-tively reviewed 122 patients followingcolorectal surgery. They found a shorter dis-ease-free interval (6 to 12 months) in patentswho received a transfusion.36 A similar investi-gation detected contrasting results and re-vealed that 43% of transfused patients devel-oped recurrent disease or died, compared with9% who did not receive a transfusion.37 How-ever, a large multicenter, randomized, con-trolled study of colorectal patients (n=475)with cancer found no direct relationship be-tween allogeneic transfusion and prognosis(cancer-free survival rates after 4 years wereno different between the two groups), but thedata did suggest an increased in recurrenceno matter if the blood was allogeneic or au-tologous38 (Table 4). These studies did not fil-ter the white cells from the blood components.In an investigation that did filter, the resultswere confusing because a number of patientsreceived both types of blood products. Thisstudy did not reach any conclusions.39 Unfor-tunately, retrospective studies have inherentflaws and conflicting conclusions. These results

should be reviewed skeptically. To counteractor attempt to explain the association with tu-mor recurrence and blood, a meta-analysis wasperformed, reviewing retrospective studies ofcolorectal cancer patients. In this statisticalreview, this association was not confirmed.40

This study was in contradistinction to anothermeta-analysis review, in which there was ahigher recurrence rate in patients withcolorectal and head and neck cancer.30

In a prospective, nonrandomized study,315 consecutive patients with prostatic cancerwho underwent radical retropubic prostatec-tomy were analyzed. Group 1 received at leastone unit of allogeneic blood with or withoutautologous blood; group 2 received autolo-gous blood only or no blood. These patientsreceived no adjuvant hormonal therapy or ra-diotherapy. The incidence of reoccurrence wassimilar: 25% vs. 23%, respectively. In addition,mortality was not affected by the administra-tion of blood.41

In patients with high-grade soft-tissue sar-comas of the extremities and osteosarcomasof long bones, there is a suggestion that trans-fusion can alter outcome.42–44 In Rosenberg’sstudy of patients with soft-tissue tumors whounderwent various prospective, randomizedtreatment protocols, the patients without anytransfusions had a 70% actuarial 5-year disease-free survival rate while patients who received1 to 3 units of blood had a 48% rate. The over-all 5-year survival rates were 85% and 63%,respectively. As expected, tumor size correlatedinversely with outcome, but after this was takeninto consideration, the effect of transfusion stillwas a negative prognostic indicator.42 In a re-lated study that focused primarily on thecardiotoxicity of doxorubicin in patients withhigh-grade soft-tissue sarcoma, factors thatwere associated with distal metastases includedblood transfusion within 24 hours, tumors >5cm, tumors extending into the deep fascia, andother histologic subtypes.43 Similar correlationwas seen between survival and transfusions inpatients with nonmetastatic osteosarcoma oflong bones. In this retrospective study, the

survival rate was 34% with blood and 53% with-out blood. An apparent criticism (not minimiz-ing a retrospective analysis) was that 61% ofthe transfused patients had femoral tumorswhile the nontransfused group included only50%.44

In animal studies of tumor augmentation,the data are provocative but still suggestive oftransfusions as a factor. One important issueaddressed in these animal models is the re-moval of leukocytes and its timing. In athymicmice transfused with either allogeneic or syn-geneic blood or saline prior to tumor cell in-fusion, the subsequent tumor size was of equaldimensions. However, in immunocompetentmice, there were larger and heavier tumorsafter transfusion with allogeneic blood.45 Cor-respondingly, rats transfused with allogeneicor syngeneic blood stored for 1 day had higherrates of tumor growth and shorter survivaltimes than controls with saline infusion.46 Con-trary to these studies, Shirwadkar et al gavemice various doses of tumor cells with thetransfusions and concluded that theimmunomodulatory effect of transfusion issolely dependent on the dose of the inoculatedcells.47

In addressing the issue of leukocyte deple-tion, Blajchman and colleagues preempted 10days before the infusion of tumors cells eitherleukocyte-reduced or nonleukocyte-reducedblood. The pulmonary metastatic nodules wereassessed 3 weeks later. The recipients of allo-geneic transfusion had two- to five-fold in-creases in these nodules compared with theanimals receiving either leukocyte-reduced al-logeneic or syngeneic blood.48 In an acute ex-periment (tumor cells injected within 60 to 90minutes of transfusion), pulmonary metastaticnodules were greater (four- to seven-fold) inthe group with allogeneic blood. In this inves-tigation, the authors believed that the removalof allogeneic leukocytes ameliorated the tumorgrowth potential.48 Consequently, these sameinvestigators found that removal of leukocytesfollowing storage did not have similar extentof amelioration.49

Table 3.Immunologic Laboratory

Tests in Transfused Patients

Decreased lymphocyte response to mitogens/alloantigensDown-regulation of natural killer, cytotoxic T lymphocytesDecreased IL-2 productionDecreased CD4 cellsIncreased CD7 cellsDecreased natural killer cell numberIncreased B cellsMacrophage function

Decreased migration to chemotactic stimuliDecreased deposition in inflammatory foci

Increased macrophage prostaglandin E2 productionDecreased antigen-presenting cell activityDecreased delayed-type hypersensitivity

From Blumberg and Heal.34

Table 4.Multivariate Analysis of Factors Related to Disease-Free Survival in Patients

Undergoing Colorectal Curative Surgery

Factors Relative 95% Confidence pRecurrence Rate Interval

Transfusion groupAllogeneic 1 — —Autologous 1.1 0.7–1.6 0.74

Dukes’ classificationA 1 — —B 4.0 1.7–9.5 0.002C 10.8 4.7–25.1 <0.001

From Busch et al.38


Tumor Recurrence and InfectionSince there appears to be an

immunomodulatory effect of transfusions, thequestion arises, particularly in regard to pa-tients with cancer, is there a higher rate of in-fection? In reviewing the data in Heiss’s series,the postoperative infection rate was higher inthe allogeneic group (27%) compared with theautologous group (12%). Multivariant regres-sion analysis revealed that infection was relatedto transfusion, with an odds ratio of 2.84. Seg-menting the groups revealed that the infectionrate also increased with a greater certainty withallogeneic blood compared with the autolo-gous group.50 In a large prospective study ofcolorectal patients (n=871), patients were ran-domly assigned either leukocyte-filtered blood(<0.2 x 109 leukocytes) or blood without abuffy coat (0.8 x 109 leukocytes per unit). At 3-year follow-up, there was no statistical differ-ence in the infection rate. It is interesting tonote that in this study a certain number (>3)of transfusions was a marker or independentrisk factor for survival as well, similar to tumorlocation or size. This correlated with the inci-dence of infection in the curative surgery pa-tients.51 Even though statistical analysis se-lected certain factors, such as >3 units of blood(which had greater postoperative morbidityand mortality), was this associated with themore complex patient with extensive diseaseand technically difficult surgery?

InfectionSimilar controversy surrounds the associa-

tion between blood transfused and the inci-dence of postoperative infection. Animal mod-els suggest that allogeneic transfusion in-creases the appearance of infection.52 In trau-matic burn or induced peritonitis experimen-tal models, animals had shorter survival withallogeneic transfusions than the groups receiv-ing either crystalloid or syngeneic blood.30,31,33

Interestingly, contrasting this traumatic model,Brunson did not induce injury and only in-fused blood. They found that the addition oftrauma, not the dose of blood, altered the im-mune system, suggesting trauma alone ablatedthe immune response toward infection.53 Tochallenge this premise, Gianotti et al subjected

mice to burn injury and infused Escherichiacoli. They found that enhanced gut permeabil-ity and bacterial challenge responded syner-gistically in secondary infection.54 A follow-upinvestigation showed an association betweenallogeneic leukocytes and an adverse effect onhost bacterial mechanisms.55

Clinically, the results are more confusing.A retrospective analysis of orthopedic patientsrevealed that allogeneic transfusions are asso-ciated with increased frequency of postopera-tive infections, including pneumonia and uri-nary tract infection.56 Other reports (one ret-rospective and one prospective) showed anassociation.57,58 In a study of patients under-going total hip replacement, Murphy and as-sociates found a proven or suspected infectionin 32% of their select group, e.g., patients with-out prior infections, malignancy, and infusionof <3 units of blood with allogeneic infusions.This is in comparison to 3% of patients givenautologous blood. The hospital stay was con-siderably longer in patients with suspected in-fection.57 In a nonrandomized, prospective trial,Triulzi and colleagues detected infection in20.8% of the allogeneic blood recipients com-pared with 4% of the nontransfused individu-als. Apparently, the amount of units given cor-related with the infection rate.58 In another ret-rospective review in patients undergoing eitherhip, spinal, or knee surgery, the data betweenthe groups were not conclusive.59 Howard andVamvakas corroborated these inconclusive find-ings.60,61 In a different population (cardiac sur-gery patients), the results are more revealing. Amultivariate analysis demonstrated contributingfactors for postoperative infection asreoperation, blood transfusions, early chestreexploration, and sternal rewiring. The diffi-culty with this statistical review is that one wouldexpect the infection rate to be higher in emer-gency reoperations. Controlling clinical factorswould be almost impossible.62,63

ConclusionsThe weight of scientific evidence from

both basic science and clinical studies sug-gests that allogeneic transfusions have a sig-nificant but varied effect on the immune sys-tem. There is no doubt there is dynamic

immunomodulatory effect on the recipient.The leukocytes appear to be the culprit. Thereason why removing white cells prior to stor-age to minimizes complications (infections,recurrence of tumor) is not understood. Onething certain is that the extent of transfusionscorrelates with these secondary problems, butin patients who receive a greater number ofblood products, what is the predominant in-determinate factor: the underlying diseaseprocess, the patient’s co-morbidity factors, orthe aggressiveness of surgical eradication?

No large clinical trials of transfusion intrauma patients (who tend to be young andnot have complicating medical diseases) havebeen undertaken to determine the incidenceof infection when leukocytes are removed priorto storage. In the author’s opinion, one groupwill benefit, and that group includes patientstransfused with <10 units of blood. However,this investigation must be initiated upon ar-rival to the definitive area and it must beblinded and prospective. Should the standardof blood banking include prestorage filteringof all blood? Economically, it would be feasibleto filter the blood when the potential risk ofinfection and cancer is there. The precedentfor accepting increased cost without clearlydemonstrated benefit has already been set bythe much greater costs involved in the preven-tion of transmission of the AIDS virus by p24antigen testing in blood banking64 (Table 5).

References1. Gutierrez G. Cellular metabolism during

hypoxia. Crit Care Med 1991; 19:619.2. Shoemaker WC, Appel PL, Bishop MH.

Temporal patterns of blood volume, he-modynamics, and oxygen transport inpathogenesis and therapy of postopera-tive adult respiratory distress syndrome.New Horizons 1993; 1:522.

3. Dzieczkowski JS, Anderson KC. Transfusionand therapy. In Fauci AS et al, eds.Harrison’s Principles of Internal Medicine,14th ed, New York, McGraw-Hill, 1998.

4. Bretscher P, Cohn M. A theory of self-nonself discrimination. Science1979;169:1042–9.

5. Lafferty KJH, Cunningham A. A new analy-

Table 5. Allogeneic Transfusion Immunomodulation-Related PostoperativeInfection and Cancer Recurrence: Theoretic Estimates of U.S. Mortality Rates

Infection Cancer Recurrence TotalDeath Rate Death Rate Death Rate

Estimated % Deaths per Million Deaths per Million Deaths per MillionCausal Contribution per Year Transfusions per Year Transfusions per Year Transfusions

100 1,500 250 20,000 33,000 21,500 33,25050 750 125 10,000 16,667 10,750 16,89210 150 25 2,000 3,300 2,150 3,3251 15 2.5 200 330 215 332.5

From Blumberg.64


sis of allogenic interactions. Aust J Exp BiolMed Sci 1975;53:27–42.

6. Nickerson PW, Steurer W, Steiger J, StromTB. In pursuit of the “Holy Grail”: allografttolerance. Kidney Int 1994; 45(suppl44):S40-9.

7. Donovan JA, Koretzky GA.CD45 and theimmune response. J Am Soc Nephrol 1993;4:976-85.

8. Bach FH, Graupner EE, Klostermann H. Cellkinetic studies in mixed leukocyte cultures:an in vitro model of homograf reactivity. ProcNatl Acad Sci USA 1969; 63:377–84.

9. Fischer-Lindahl K, Wilson DB. Histocom-patibility antigens-activated cytotoxic Tlymphocytes. II. Estimates of the fre-quency and specificity of precursors. J ExpMed 1977; 145:508–22.

10. Widmer MB, Donald HRM. Cytolytic T lym-phocyte precursors reactive against mu-tant Kb alloantigens are as frequent asthose reactive against a whole foreign hap-lotype. J Immunol 1980; 127:48–51.

11. Lechler RI, Batchelor JR. Restoration ofimmunogenicity to passenger cell-de-pleted kidney allografts by the addition ofdonor strain dendritic cell. J Exp Med1982; 155:31–41.

12. Shoskes A, Wood KJ. Indirect presentationof MHC antigens in transplanation.Immunol Today 1994; 15:32.

13. Sayegh M, Watschinger B, Carpenter CB.Mechanisms of T cell recognition of al-loantigen: the role of peptides.Transplanation 1994; 57:1295–302.

4. Billingham RE, Brent L, Medawar PB. Ac-tively acquired tolerance of foreign cells.Nature 1953; 172:603–6.

15. Calne RY. The rejection of renalhomografts inhibition in dogs with BM6-mercaptopurine. Lancet 1960; 1:417–8.

6. Zukoski C, Lee HM, Hume DM. The pro-longation of functional survival of caninerenal homografts with BM6 mercaptopu-rine. Surg Forum 1960; 11:470–2.

17. Jenkins A. McL., Woodruff MFA. The effectof prior administration of donor strainblood or blood constituents on the sur-vival of cardia allografts in rats. Transplan-tation 1972; 12:57–60.

8. Fabre JW, Morris PJ. The effect of donorstrain blood pretreatment on renal al-lograft rejection in rats. Transplantation1972; 14:608–17.

19. Newton WT, Anderson CB. Plannedpreimmunization of renal allograft recipi-ents. Surgery 1973; 74:430–6.

20. Lagaaij EL, Hennemann PH, Ruigrok M,et al. Effect of one-HLA-DR-antigen-matched and completely HLA-DR-mis-matched blood transfusions on survival ofheart and kidney allografts. N Engl J Med1989; 321:701–5.

21. Opelz G, Sengar DP, Mickey MR, et al? Ef-fect of blood transfusions on subsequentkidney transplants. Transplant Proc 1973;5:253–9.

22. Quigley RL, Wood KJ, Morris PJ. Investi-gation of the mechanism of active en-hancement of renal allograft survival byblood transfusion. Immunology 1988;63:373–81.

23. Salvatierra O, Vincenti F, Amend W. Delib-erate donor specific blood transfusionsprior to living renal transplantation: a newapproach. Ann Surg 1980; 192:543–52.

24. Potter D, Garovoy M, Hopper S, TerasakiP, Salvatierra O Jr. Effect of donor-specifictransfusions on renal transplantation inchildren. Pediatrics 1985; 76:402–5.

25. Anderson CB, Jendrisak MD, Flye MW,Hanto DW, Anderman CK, Rodey GE,Sicard GE. Renal allograft recipientimmunomodulation by concomitant im-munosuppression and donor-specifictransfusions. Transplant Proc 1988;20:1074–8.

26. Sollinger HW, Kalayoglu M, Belzer FO. Useof the donor specific transfusion protocolin living-unrelated donor recipient combi-nations. Ann Surg 1986; 204:315–21.

27. Belzer FO, Kalayoglu M, Sollinger HW.Donor-specific transfusion in living-unre-lated renal donor-recipient combinations.Transplant Proc 1987; 19:1514–5.

28. Van Twuyver E, Mooijaart RJD, ten BergeIJM, et al. Pretransplantation blood trans-fusion revisited. N Engl J Med 1991;325:1210–3.

29. Brennan DC, Mohanakumar T, Flye MW.In-depth review donor-specific transfusionand donor bone marrow infusion in re-nal transplantation tolerance: a review ofefficacy and mechanisms. Am J Kidney Dis1995; 26:701–15.

30. Blumberg N, Heal JM. Effects of transfu-sion in immune function. Arch Pathol LabMed 1994; 118:371–9.

3. Vamvakas EC, Moore SB. Blood transfu-sion and postoperative septic complica-tions. Transfusion 1994; 34:714–27.

32. Klein HG. Wolf in wolf ’s clothing: is it timeto raise the bounty on the passenger leu-kocyte? Blood 1992; 80:1865–8.

33. Bordin JO, Heddle NM, Blajchman MA.Review: biologic effects of leukocytespresent in transfused cellular blood prod-uct. Blood 1994; 84:1703–21.

34. Blumberg N, Heal JM. Transfusion andrecipient immune function. Arch PatholLab Med 1989; 113:246–53.

35. Zinkernagel RM, Doherty PC. Restrictionof in vitro T cells mediated cytotoxicity inlymphocytic choriomeningitis within asyngeneic or semiallogeneic system. Na-ture 1974; 248:701–2.

36. Burrows L, Tartter P. Effects of blood trans-fusions on colonic malignancy recurrencerate (letter). Lancet 1982 ;2:662.

37. Blumberg N, Agarwal MM, Chuang C. Re-lation between recurrence of cancer of thecolon and blood transfusion. Br Med J1985; 290:1037–9.

38. Busch OR, Hop WC. Hoynck van

Papendrecht MA, et al. Blood transfusionsand prognosis in colorectal cancer. N EnglJ Med 1993; 328:1372–6.

39. Heiss MM, Mempel W, Delanoff C, et al.Blood transfusion-modulated tumor re-currence: First results of a randomizedstudy of autologous versus allogeneicblood transfusion in colorectal cancersurger. J Clin Oncol 1994; 12:1859–67.

40. Vamrakis E, Moore SB. Perioperative bloodtransfusion and colorectal cancer recur-rence: A qualitative statistical overview andmeta-analysis. Transfusion 1993; 33:754–65.

41. Ness PM, Walsh PC, Zahurak M, et al. Pros-tate cancer recurrence in radical surgerypatients receiving autologous or homolo-gous blood. Transfusion 1992; 32:31–6.

42. Rosenberg SA, Seipp CA, White DE, et al.Perioperative blood transfusions are asso-ciated with increased rates of recurrenceand decreased survival in patients withhigh-grade soft-tissue sarcomas of the ex-tremities. J Clin Oncol 1985; 3:698–709.

43. Casper ES, Gaynor JJ, Hajdu SI, et al. Pro-spective randomized trial of adjuvant che-motherapy with bolus versus continuousexclusion doxorubicin in patients withhigh-grade extremity soft tissue sarcomaand an analysis of prognostic factors. Can-cer 1991; 68:1221–9.

44. Chesi R, Cazzola A, Bacci G, et al. Effect ofperioperative transfusions on survival inosteosarcoma treated by mutlimodaltherapy. Cancer 1989; 64:1727–37.

45. Francis DMA, Burren CP, Clunie GJA. Ac-celeration of B16 melanoma growth inmice after blood transfusion. Surgery1987; 102:485–92.

46. Waymack JP, Chance WT. Effect of bloodtransfusions on immune function: IV. Ef-fect on tumor growth. J Surg Oncol 1988;39:159–64.

47. Shirwadkar S, Blajchman MA, Frame B, etal. Effect of blood transfusions on experi-mental pulmonary metastases in mice.Transfusion 1990; 30:188–90.

48. Blajchman MA, Bardossy L, Carmen R, etal. Allogeneic blood transfusion-inducedenhancement of tumor growth: two ani-mal models showing amelioration byleukodepletion and passive transfer usingspleen cells. Blood 1993; 81:1880–2.

49. Bordin JO, Bardossy L, Blajchman MA.Growth enhancement of established tu-mors by allogeneic blood transfusion inexperimental animals and its ameliorationby leukodepletion: the importance of thetiming of the leukodepletion. Blood 1994;84:344–8.

50. Heiss MM, Mempel W. Jauch KW, et al.Beneficial effect of autologous bloodtransfusion on infectious complicationsafter colorectal cancer surgery. Lancet1993; 342:1328–33.

51. Houbiers JGA, Brand A, van de WateringLMG, et al. Randomised controlled trialcomparing transfusion of leucocyte-de-


pleted or buffy-coat-depleted blood insurgery for colorectal cancer. Lancet 1994;344:573–8.

52. Waymack JP, Warden GD, Alexander JW, etal. Effect of blood transfusion and anes-thesia on resistance to bacterial peritoni-tis. J Surg Res 1987; 42:528–35.

53. Brunson ME, Ing R, Tchervenkov JL, et al.Variable infection risk following allogeneicblood transfusions. J Surg Res 1990;48:308–12.

54. Gianotti L, Pyles T, Alexander JW, et al. Im-pact of blood transfusion and burn injuryon microbial translocation and bacterialsurvival. Transfusion 1992; 32:312–7.

55. Gianotti L, Pyles T, Alexander JW, et al.Identification of the blood component re-sponsible for increased susceptibility togut-derived infection. Transfusion 1993;

33:458–65.56. Blumberg N, Heal JM. Transfusion and host

defenses against cancer recurrence and in-fection. Transfusion 1988; 29:236–45.

57. Murphy P, Heal JM, Blumberg N. Infectionor suspected infection after hip replace-ment surgery with autologous or homolo-gous blood transfusions. Transfusion1991; 31:212–7.

58. Triulzi DJ, Vanek K, Ryan DH, et al. A clini-cal and immunologic study of blood trans-fusion and postoperative bacterial infec-tion in spinal surgery. Transfusion 1992;32:517–24.

59. Fernandez MC, Gottlieb M, Menitove JE.Blood transfusion and postoperative in-fection in orthopedic patients. Transfu-sion 1992; 32:318–22.

60. Vamvakas EC, Moore SB, Cabanela M.

Blood transfusion and septic complica-tions after hip replacement surgery. Trans-fusion 1995; 35:15–6.

61. Howard HL, Rushambuza FG, Martlew VJ,et al. Clinical benefits of autologous bloodtransfusion: an objective assessment. ClinLab Haematol 1993;15:165–71.

62. Vamvakas EC, Carven JH, Hibberd PL.Blood transfusion and infection aftercolorectal cancer surgery. Transfusion1996; 36:1000–8.

63. Ottino G, De Paulis R, Pansini S, et al. Majorsternal wound infection after open-heartsurgery: A multivariate analysis of risk fac-tors in 2,579 consecutive operative proce-dures. Ann Thorac Surg 1987; 44:173–9.

64. Blumberg N. Allogeneic transfusion andinfection: economic and clinical implica-tions. Sem Hematol 1997; 34:34.

Andrew D. Rosenberg, MDDepartment of AnesthesiologyHospital for Joint Diseases/Orthopaedic InstituteNew York, New York, USA

The trauma patient frequently requiresmultiple blood transfusions during resuscita-tion to achieve a stable hemodynamic state.Adequate oxygen-carrying capacity necessitatestransfusion based on the patient’s pathophysi-ology after being injured, the patient’s baselinemedical condition, and actual and anticipatedblood loss.

Transfusion is often necessary, but it is notalways benign. To even consider the conceptof decreasing the amount of blood transfusedto trauma patients, we must determinewhether we can accomplish this goal withoutaffecting outcome. Obviously, many patientswould die without transfusion. Although bloodtransfusions increase oxygen-carrying capacity,massive transfusion is associated with physi-ologic alterations, immunomodulation, andpostoperative infection. Two questions havebecome important in transfusion medicine: 1)What is in the blood? and 2) What are the sys-temic effects of transfusion other than increas-ing the hematocrit?

Despite safeguards and tests to ensure thatblood is not contaminated, blood is being re-leased that is in fact contaminated. Transfusionof tainted blood can transmit the human im-munodeficiency virus (HIV), hepatitis, cytome-galovirus (CMV) and syphilis. Testing for HIVhas become increasingly accurate, so the win-dow period for possible infection has beenshortened because of earlier dectection. Thewindow period for HIV is that time in which aperson is infected with the virus but has notyet demonstrated infectivity by available test-ing methods. In a study conducted a numberof years ago, 39 patients became seropositive

Blood Transfusions from 182 “seronegative” donors. This resultedfrom the long period necessary to develop de-tectable levels of antibodies to HIV. Currentantibody testing has diminished the window to22 days. In March 1996, the U.S. Food and DrugAdministration mandated P24 antigen testing,which decreased the window period to 16 days.

In addition to HIV transmission, hepatitisB, hepatitis C, and CMV can be transmittedeasily if blood is not tested adequately. CMV isfrequently present in transfused blood, itsprevalence determined by geographic location.Special care must be taken in the immunosup-pressed patient to ensure that CMV is notpresent in transfused blood. Bacterial and para-sitic infections can also be transmitted. Othercomplications known to occur with transfu-sions include allergic reactions, hemolytictransfusion reactions, and volume overload.

Transfusions may also result in immuno-suppression or immunomodulation of the re-cipient. Studies have demonstrated that renaltransplant patients had improved allograft sur-vival times and lower allograft rejection rates ifthey received transfusions of bank blood (allo-geneic blood) prior to receiving their allograft.1

During the 1970s, some protocols requiredpatients receiving cadaveric renal transplants toreceive transfusions prior to the transplant pro-cedure. Transfusion of whole blood was a stron-ger enhancer of allograft survival than packedred blood cells. The prevalent thought was thattransfusion induced an immunosuppressiveeffect in the patient and thus, after the patientreceived the transplant, rejection did not oc-cur. Fortunately, the need for preoperative al-logeneic transfusion has been mitigated by theintroduction of cyclosporin.2

In addition to evidence that allogeneictransfusions result in immunosuppression,there is evidence that cancer patients who re-ceive these transfusions at the time of surgeryhave lower survival rates and an increased in-cidence of recurrence.2 Meta-analysis has dem-onstrated this finding to be true in patients

with colorectal cancer and head and neck can-cer.3 Osteosarcoma patients who receiveperioperative blood transfusions have an in-creased incidence of metastases and shortersurvival time.2

An altered immunologic state results fromreceiving a blood transfusion. Allogeneic bloodtransfusions have been associated with de-creases in cell-mediated immunity, macroph-age migration, and natural killer cell activity.Additionally, allogeneic transfusion affects thecells that incite B-lymphocytes to differentiateand produce antibodies. These immunosup-pressive effects are thought to be the result ofeither antigen excess, a graph-versus-host phe-nomenon, reactivation of immunosuppressiveviruses, or the white blood cells that aretransfuseed along with red blood cells.2

Allogeneic blood transfusions have alsobeen implicated in postoperative infections.Independently, Murphy and Triulzi, in sepa-rate studies on orthopaedic patients, demon-strated the effect of allogeneic blood transfu-sions in producing postoperative infection.4,5

A significant increase in postoperative infec-tion rates occurred in patients who receivedallogeneic blood transfusions during eithertotal hip or spine surgery compared with pa-tients who did not receive allogeneic blood.Of patients who received allogeneic transfu-sions, there was an infection rate of 20.8% in astudy of 102 patients undergoing 109 spinalfusions. The infection rate was only 3.5% inthose who did not receive allogeneic blood.Natural killer cell activity, an indicator of im-munologic function, decreased in the patientswho received allogeneic transfusion. A specificdose-response curve demonstrated that pa-tients who received two transfusions had ahigher infection rate than patients who re-ceived either one or no transfusion at all.4,5

Fernandez demonstrated that patients whoreceived homologous whole blood had ahigher incidence (20%) of infection comparedwith the overall (6.1%) infection rate for all

9


patients in the study.6 Some orthopaedic stud-ies do not demonstrate an association betweenallogeneic transfusion and infection. In a meta-analysis, Vamvakkas et al were unable to dem-onstrate a clear relationship between transfu-sion and infection. Their study criteria, how-ever, defined a significant relationship occur-ring between transfusion and infection as onethat would result in an infection rate more thandouble the baseline occurrence rate.3

There is significant evidence that transfu-sions are associated with immunomodulationand increased infection in trauma patients.7–11

Rosemurgy demonstrated an increased inci-dence in postoperative infection in a popula-tion of 390 uncrossmatched trauma patientswho received type O blood. In the 61% of pa-tients who survived at least 7 days, the infec-tion rate was higher in those who receivedseven or more units of packed red blood cells.7

Dellinger noted that, while wound infectionsafter open fractures of the arm or leg were af-fected by local factors, nosocomial infectionswere related to Injury Severity Score (ISS), theincidence of blood transfusion, patient age,and the mode of injury. Edna and Bjerkeset,in a Norwegian study of 484 trauma patientswho survived longer than 2 days, demonstrateda 9.5% infection rate, with a univariate rela-tionship between infection and transfusion.10

This relationship was independent of ISS, age,and surgical procedure. The risk of infectionafter colon injury is associated with bloodtransfusion, age, and the number of associatedinjuries and splenic injury. Agarwal, in a studyof 5,366 consecutive trauma patients,documentrf that blood transfusion was a pre-dictor of infection after controlling for patient’sage, sex, mechanism, or severity of injury.11

In a study of 619 geriatric patients withhip fracture, a study at the author’s institutiondocumented a significantly higher incidenceof urinary tract infections in patients who re-ceived allogeneic transfusion compared withthose who did not require any transfusion.12

Riska demonstrated a linear relationshipbetween the number of units transfused andmortality, with 21 to 39 units being associ-ated with a 25% mortality and more than 40units associated with a 52% mortality.13 Wudeldocumented 5 survivors of more than 50 unitsof blood after massive transfusion.9 Blunt andpenetrating trauma patients receiving mul-tiple transfusions had similar survival rates(59%). Shock, closed head injury, and agepredicted increased mortality but did not pre-clude survival.

Massive transfusion may be associatedwith high citrate and acid load, possible he-mostatic failure, disseminated intravascularcoagulation, large amounts of infused blooddebris, inadequate 2,3-DPG levels, and throm-bocytopenia. Thus, although multiple transfu-sion is indicated under many conditions, weneed to consider what are appropriate trans-fusion triggers. What factors are consideredimportant in determining the need for trans-

fusion? In order to tolerate low hemoglobin,patients must be able to compensate for thedecreased oxygen-carrying capacity associatedwith decreased concentrations of red bloodcells. Healthy patients can frequently compen-sate, but this ability becomes compromisedwith age and cardiac and respiratory disease.Increases in cardiac output must be sufficientto overcome existing deficits. Since oxygendelivery depends on cardiac output and arte-rial oxygen concentration, in addition to sup-plying enhanced oxygen concentration, thepatient must be able to increase stroke volumeand heart rate. The trauma patient is faced withacute decreases in hemoglobin levels and notafforded the ability to compensate, as do pa-tients with chronic anemia. Once volume sta-tus is repleted, hemoglobin (Hb) levels mustbe evaluated to determine the need for trans-fusion. Most patients require transfusion whenthe Hb is less than 6 gm/dl and few require itwhen the Hb is more than 10 gm/dl. Transfu-sion in the intermediate area requires consid-eration of physiologic status and theindividual’s ability to ensure adequate oxygen-ation to vital organs.

ConclusionMany trauma patients require blood trans-

fusions to replenish massive blood loss fromwounds. The advantages of predonation andcell salvage techniques are not present underemergency conditions or are inappropriatebased on the type of injury. Currently, this leavesbanked blood as the source of blood for trans-fusion. The advantages afforded by administer-ing allogeneic blood to enhance oxygen-carry-ing capacity must be weighed against its adverseside effects, which include immunomodulation,transmission of infectious diseases, and thepossibility of a transfusion reaction.

References1. Opelz G, Terasaki PI. Improvement of kid-

ney-graft survival with increased numbersof blood transfusions. N Engl J Med 1976;

299:798–803.2. Landers DF, Hill GE, Wong KC, Fox IJ. Blood

transfusion-induced immunomodulation.Anesth Analg 1996; 82:187–204.

3. Vamvakas EC, Moore SB, Cabanela M.Blood transfusion and septic complica-tions after hip replacement surgery. Trans-fusion 1995; 35:150–6.

4. Triulzi DJ, Vanek K, Ryan DH, BlumbergN. A clinical and immunologic study ofblood transfusions and postoperative bac-terial infection in spine surgery. Transfu-sion 1992; 32:517–24.

5. Murphy P, Heal JM, Blumberg N: Infectionor suspected infection after hip replace-ment surgery with autologous or homolo-gous blood transfusions. Transfusion1991; 31:212–7.

6. Fernandez MC, Gottlieb M, Menitove JE.Blood transfusion and postoperative in-fection in orthopedic patients. Transfu-sion 1992; 32:318.

7. Rosemurgy AS, Hart ME, Murphy CG, et al.Infection after injury associated with bloodtransfusion. Am Surg 1992; 2:104–7.

8. Phillips TF, Soulier G, Wilson. Outcomeof massive transfusion exceeding twoblood volumes in trauma and emergencysurgery. J Trauma 1987; 27:903–10.

9. Wudel JH, Morris JA, Yates K, et al. Mas-sive transfusion: outcome in blunt traumapatients. J Trauma 1991; 31:1–7.

10. Edna TH, Bjerkeset T. Association betweenblood transfusion and infection in injuredpatients. J Trauma 1992; 33:659–61.

11. Agarwal N, Murphy J, Cayten L, et al. Bloodtransfusion increases the risk of infectionafter trauma. Ann Surg 1993; 128:171–7.

12. Koval KJ, Rosenberg AD, Zuckerman JD,et al. Does blood transfusion increase therisk of infection after hip fracture? JTrauma 1997; 11:260–6.

13. Riska EB, Bostman O, von Bonsdorff H,et al. Outcome of closed injuries exceed-ing 20-unit blood transfusion need. Injury1988; 19:273–6.

Maureen Nash Sweeney, MDAttending AnesthesiologistAnesthesiology DepartmentDepartment of Veterans Affairs Medical CenterNew York, New York, USA

Vascular access in the trauma patient isessential for three reasons:• administration of intravenous fluids• administration of drugs• measurement and monitoring of cardiac

parameters

Vascular Access in Trauma: Options, Risks, Benefits, Complications

In the trauma patient presenting withmultiple serious injuries and hemorrhagicshock, vascular access is necessary to restorecirculatory volume rapidly. The urgency of theplacement and the size and number of intra-venous (IV) lines is dictated by the degree ofshock, the apparent rate of bleeding, and thetype of injury. Advanced Trauma Life Support(ATLS™) protocol recommends proceedingwith attempts at percutaneous peripheral ac-cess, followed by a surgical venous cutdownbefore resorting to central venous access. Therationale is that, in a hypovolemic patient, the

10


likelihood of success with a venous cutdownis greater than with a central line. Addition-ally, the rate of complications (e.g., pneu-mothorax and arterial puncture) is higher withcentral IV access.1 However, the most impor-tant factor in considering the procedure androute for vascular access is the individualphysician’s level of skill and expertise.

Location of the injury must be consideredwhen choosing a site for venous access. Venousaccess must never be initiated in an injured limb.In patients with injuries below the diaphragm,at least one IV line should be placed in a tribu-tary of the superior vena cava, as there may bevascular disruption of the inferior vena cava.Patients with upper thoracic and neck injuriesshould have large-bore access in the lower ex-tremity, as there may be superior vena cava dis-ruption. In patients with severe multitrauma inwhom occult thoracoabdominal damage is sus-pected, it is recommended to have one IV ac-cess site above the diaphragm and one belowthe diaphragm, thus accessing both the supe-rior vena cava and inferior vena cava, respec-tively.2,3

For rapid administration of large amountsof intravenous fluids, short large-bore cathetersshould be used. Based on Poiseulle’s law, therate of fluid flow is inversely proportional tothe length of the catheter and directly propor-tional to its internal diameter:

∏r4(∆P)

8nL

where Q=flow, r=radius of the catheter,P=driving pressure through the catheter (grav-ity or externally applied), n=viscosity of thesolution, and L=length of IV tubing. Doublingthe internal diameter of the venous cannulaincreases the flow through the catheter 16-fold.A 14-gauge, 5-cm catheter in a peripheral veinwill pass fluid twice as fast as a 16-gauge, 20-cm catheter passed centrally. Although resis-tance to flow is added by multiple stopcocksand connections, stopcocks are recommendedfor universal precautions. When using 8.5French pulmonary catheter introducers, theside port should be removed, as this increasesthe resistance roughly four-fold. For subcla-vian, internal jugular, femoral, and antecubitallines, 8.5 French introducers can be used.4

Percutaneous Intravenous InsertionATLS™ guidelines recommend rapid

placement of two large-bore (16-gauge orlarger) IV catheters in the patient with seriousinjuries and hemorrhagic shock. The firstchoice for IV insertion should be a peripheralextremity vein. The most suitable veins are atthe wrist, the dorsum of the hand, the antecu-bital fossa in the arm, and the saphenous inthe leg. These sites can be followed by the ex-ternal jugular and femoral vein.

The complication rate of properly placedintravenous catheters is low. Intravascular

placement of a large-bore IV should be veri-fied by checking for backflow. An IV site shouldinfuse easily without added pressure. Intrave-nous fluids can extravasate into soft tissueswhen pumped under pressure through an in-filtrated IV line, and a compartment syndromecan result. It is always best to have intravenoussites out where they can be examined.

Central Venous AccessRapid peripheral percutaneous IV access

may be difficult to achieve in patients with hy-povolemia and venous collapse, edema, obe-sity, scar tissue, history of IV drug abuse, orburns. Under such circumstances, central ac-cess with wide-bore catheters can be advanta-geous. An additional benefit is the ability tomonitor central venous pressure. However,subclavian and internal jugular catheterizationshould not be used routinely in trauma pa-tients, as the complications can be dangerous.

Subclavian CatheterizationSubclavian catheterization provides rapid

and safe venous access in experienced hands.The most frequent complication of subclavianvenipuncture is pneumothorax. Pneumotho-rax is more likely to occur on the left side be-cause the left pleural dome is anatomicallyhigher. Subclavian and internal jugular cath-eters should be inserted on the side of injuryin patients with chest wounds, reducing thechances of collapse of the uninjured lung. Asimple pneumothorax may result in respira-tory compromise in individuals with pulmo-nary contusions or a pneumothorax in the con-tralateral hemithorax.2 A suspected injury to thesubclavian vein is an exception to this principlebecause the infused fluid may extravasate intothe mediastinum or thoracic cavity.

A hemothorax may result from lacerationof the subclavian vein or subclavian artery. Ifthe subclavian catheter is placed inadvertentlyin the thoracic cavity, subsequent infusions ofblood or crystalloids will produce a hemotho-rax or hydrothorax. Catheter placement shouldbe ensured prior to IV infusions, whether byaspiration or by lowering the IV infusion bagbelow the patient and verifying backflow. Thesetests are suggestive of IV placement but noneis diagnostic.4 When inserting introducers overguide wires, it is important not to force theintroducer if resistance is encountered. Forc-ing the introducer could result in perforationof large veins or arteries and bleeding.

Venous air embolism is another complica-tion of central line insertion. Occlusion shouldbe maintained over the catheter lumen with agloved finger or by increasing the pressure inthe subclavian vein by Trendelenburg positionor Valsalva maneuver. Even with prompttherapy, the fatality rate with significant air em-bolism is high.5 Embolization of catheter frag-ments can occur when withdrawing a catheterwith a through-the-needle technique.

Arrhythmia may occur during line place-ment when the catheter or wire contacts the

endocardium of the atrium or ventricle. Properpositioning of the catheter in the superior venacava (SVC) usually abolishes this problem.Myocardial perforation and tamponade rarelyoccur.

Thrombosis or thrombophlebitis occurswith malpositioned or misdirected catheters.The subclavian catheter is often malpositionedinto the internal jugular vein. When the cath-eter is placed properly in the SVC, thrombosisusually does not occur because of the highflow and large caliber of the vessel. A kinkedor knotted catheter in the SVC may lead tothrombosis.

Injury to the brachial plexus or phrenicnerve may result from attempts to place a sub-clavian line. The nerves are posterior to thevein, and injury occurs when the needle haspenetrated both walls. Left-sided central lineattempts can result in thoracic duct injury.

Infectious complications associated withline placement can be prevented by usingproper sterile technique. Any lines placed dur-ing resuscitation of a trauma patient withoutstrict aseptic technique should be removed.

Internal Jugular Vein CatheterizationPercutaneous placement of internal jugu-

lar (IJ) catheters is also an excellent means ofattaining rapid large-bore catheter access. Cer-vical trauma is a contraindication for internaljugular placement. Trendelenburg positionand Valsalva maneuver help to distend the in-ternal jugular vein and improve the rate ofsuccess for venipuncture.

Carotid artery puncture is a common com-plication of IJ catheter placement. Local directpressure can prevent hematoma formation.Carotid puncture is a contraindication to at-tempting IJ catheter placement on the oppo-site side, because bilateral hemorrhages couldcompress the airway.

Other complications from IJ venipunctureare similar to those associated with subclavianvenipuncture. The incidence of pneumotho-rax is less with IJ catheter placement than withplacement of a subclavian line. The incidenceof hemothorax, mediastinal migration of thecatheter, and intrapleural catheter placementtends to be greater with left IJ placement thanright because the left IJ is more circuitous, andadvancement of a catheter can rupture thevessel. Stellate ganglion injury is a possiblecomplication.

Femoral and Basilic-CephalicCentral Lines

Femoral vein cannulation is another alter-native for line placement and is associated withfewer acute complications. Bowel perforationcan occur, especially in patients with femoralhernia. Penetration of the hip could result inseptic arthritis. Thrombophlebitis occurs moreoften with femoral than with IJ or subclaviancatheters; however, this is most likely with pro-longed use.

Basilic-cephalic catheterization may be

Q =


used for central access and central venous pres-sure monitoring with a “long-arm” catheter.Introducers can also be inserted safely. Theyare easily placed and associated with a lowcomplication rate.

Venous CutdownsVenous cutdowns can be performed when

rapid, secure, large-bore venous cannulationis desirable, such as in hemodynamic shockand in situations where percutaneous periph-eral or central access is either contraindicatedor impossible to achieve.

Most favored sites for cutdowns are thecephalic, basilic, and median antecuital veinsin the upper extremity and the greater saphe-nous in the lower. These veins can accept largecatheters, allowing rapid infusion. Strict asep-tic technique should be used. Surgical masksand caps should be worn.6

Venous cutdown has a low potential foranatomic damage. Cutaneous nerve injury is themost common problem. The infection rate isrelatively low when used acutely but increasesprecipitously over time. Therefore, it is recom-mended that venous cutdown catheters be re-moved as soon as it is possible to achieve IVaccess through standard percutaneous IV cath-eters or a central venous catheter.5

Vascular Access in Pediatric PatientsIdeally, venous access in severely injured

children should be established via a percuta-neous route. Unfortunately, this often provesto be a difficult task. ATLS™ recommends thatafter two unsuccessful percutaneous attempts,consideration should be given to intraosseousinfusion in children younger than 6 years ofage or direct venous cutdown in children over

6 years of age.1 Scalp veins should not be usedwhen rapid fluid administration may beneeded. Internal jugular and subclavian cath-eterization can be done in children but shouldbe performed only by experienced personnel.In awake children, there is a higher incidenceof pneumothorax and arterial puncture.

Intraosseous catheters can be used in allage groups but are most successful in thoseyounger than 2 because the cortical bone issofter. Fluids and drugs can be given throughthe catheter. Specially designed intraosseousneedles are available but 18- to 20-gaugeneedles, bone marrow aspiration needles, and18-gauge spinal needles can be used. Eighteen-gauge spinal needles are readily available, butthey often bend and make placement difficult.In children younger than 6 years of age, thelocations of choice are the proximal tibia andthe distal femur. When using the proximal tibialplateau, the needle should be placed 2 to 3cm distal to the level of the tibial tuberosity onthe anterior medial surface of the proximaltibia. In adults, a site 2 cm proximal to the tipof the medial malleoli is selected, with theneedle directed slightly cephalad. The distaltibia, distal femur, sternum, clavicle, andhumerous can also be used. Pressure and arotary motion should be used until there is adecrease in resistance, indicating that the med-ullary cavity has been entered. It is not alwayspossible to aspirate marrow, but IV fluid shouldrun easily without a pump.7

Complications of intraosseous infusionsinclude extavasation of fluids into surround-ing tissues, cellulitis, and osteomyelitis. Mul-tiple attempts at insertion should be avoidedsince the other holes in the bone could allowleakage of fluid into the adjacent soft tissue.

The incidence of osteomyelitis is low whencatheters are removed early. Standard periph-eral or central venous placement should beattempted when the patient is stable. Boneswith fractures and sites with open woundsshould be avoided.5

References1. Alexander RH, Proctor HJ. Advanced

Trauma Life Support Program for Physi-cians, Instructor Manual. Chicago, Ameri-can College of Surgeons, 1993.

2. Lucas CE, Ledgerwood AM. Initial evalua-tion and management of severely injuredpatients. In Wilson RF, Walt AJ, eds.. Man-agement of Trauma: Pitfalls and Practice.Baltimore, Williams & Wilkins, 1996.

3. Abrams KJ. Preanesthetic evaluation. InGrande, CM, ed. Textbook of Trauma An-esthesia and Critical Care. St. Louis,Mosby, 1993.

4. Kollef MH. Fallibility of persistent bloodreturn for confirmation of intravascularcatheter placement in patients with hem-orrhagic thoracic effusions. Chest 1994;106:1906–8.

5. Bickell W, Pepe PE, Mattox KL. Complica-tions of resuscitation. In Mattox KL, ed.Complications of Trauma. New York,Churcill Livingstone, 1994.

6. Mackersie RC. Venous and arterial cut-down. In Benumof JL, ed. Clinical Proce-dures in Anesthesia and Intensive Care.Philadelphia, J.B. Lippincott, 1992.

7. Benumof JL. Intraosseous infusion. InBenumof JL, ed. Clinical Procedures inAnesthesia and Intensive Care. Philadel-phia, J.B. Lippincott, 1992.

Charles E. Smith, MD, FRCPCMetroHealth Medical CenterCase Western Reserve UniversityCleveland, OH 44109 USAe-mail: [email protected]

[Editors’ note: Dr. Smith has received researchsupport from SIMS Level I, Augustine MedicalMallinckrodt, and Belmont Instruments.]

Hypothermia occurs frequently in traumapatients because of exposure, infusion of coldfluids and blood, opening of body cavities,decreased heat production, and impaired ther-moregulatory control.1–7 Infusion of unwarmedor inadequately warmed IV fluids and coldblood is a well known cause of hypothermiaand may contribute to the multiple adverseconsequences of hypothermia such as periph-eral vasoconstriction, metabolic acidosis,coagulopathy, wound infection, and cardiacmorbidity.1–3,8–13 This manuscript reviews the

Principles of Fluid Warming in Traumaprinciples of fluid warming as they apply tothe trauma patient.

Importance of Warming IV FluidsConclusive evidence demonstrating the

harmful effects of cold fluid infusion was pro-vided by Boyan and Howland.14 In their study,infusion of 0.5 L of cold blood reduced coretemperature of anesthetized cancer patients by0.5 to 1.0oC. When 3.0 L or more of cold bloodwas administered, esophageal temperature de-creased markedly and was associated with a highincidence of cardiac arrests (12 of 25 patients).14

When blood was warmed, the incidence of car-diac arrests in a matched group of patients withsimilar surgeries, blood loss, anesthesiologist,and surgeon was only 3 of 105 patients.15,16

The use of large quantities of unwarmedfluids for immediate resuscitation of patientswith penetrating trauma prior to emergencysurgical intervention has been discouraged.17

It is possible that the use of unwarmed fluids

may contribute to a hypothermia-induced ordilutional coagulopathy, although experimen-tal evidence suggests that hydraulic factors mayplay a more important role (e.g., disruption ofsoft clot, decreased resistance to flow arounda partially formed thrombus).18

Thermal Stress of Infusing Cold or Inad-equately Warmed Fluids and Blood

The theoretical impact of infusing fluids onbody temperature can be calculated as follows:

Change in body temperature =Thermal stress of infused fluids /(Weight x Sp heat)

where:Thermal stress = Temperature differencebetween core and infused fluids xspecific heat of infused fluid x volumeof fluid infusedWeight = weight of patient in kgSp heat = specific heat of the patient(0.83 kcal/L/oC)19,20

11


According to the specific heat of water, 1kCal of heat is required to raise the tempera-ture of 1 kg of water by 1oC. Assuming that 1 Lof crystalloid weighs 1 kg and that its specificheat is the same as water, one needs 16 kCalof energy to raise the temperature of 1 liter ofcrystalloid infused at 21oC to body tempera-ture (37oC).19-22 Similarly, infusion of 4.3 L ofcrystalloid at room temperature to an adulttrauma patient would require 71 kCal, theequivalent of 1 hour of heat production in anawake adult, or 1.5 hour of heat productionin an anesthetized adult male (heat produc-tion decreased by 33%).

The negative thermal balance of 4.3 L ofroom temperature fluids is thus equivalent toa decrease of 1oC body temperature in anawake individual and a 1.5oC temperature de-crease in an anesthetized patient. Conversely,30 kCal are required to raise the temperatureof cold 4oC blood to 37oC, such that infusionof 2 L could result in a body temperature de-crease of between 1.0 and 1.5oC.19-22

Temperature Setpoints of WarmersIn the United States, blood can be warmed

safely so as not to cause hemolysis using a tem-perature setpoint of 42oC in conjunction withan FDA-cleared blood warming device. Thissetpoint is based on observations by Uhl andcolleagues23 and is supported by a large bodyof experience with cardiac perfusion. In thestudy by Uhl et al,23 red cells were incubatedat 37, 40, 42, 44, 46, 48, and 50oC for up to 2hours in a constant-temperature waterbath.Even subtle alterations in red cell integrity suchas increased plasma hemoglobin and osmoticfragility were not apparent until 46oC.23

There has been renewed interest in deliv-ering very hot fluids in an attempt to transferheat to hypothermic patients. For example,infusion of crystalloid at 54oC will transfer ~21kCal/L to a hypothermic patient whose coretemperature is 33oC. This technique has beenshown to be relatively safe in a series of pa-tients undergoing operative burn debridementand immediate skin grafting.24 Fluids were in-fused at a rate of 110 ml/hr. In the study, therewas no evidence of intravascular hemolysis orother overt complications such as excessivebleeding or hyperkalemia.24 There is currentlynot enough safety information to recommendthis technique and there is danger that veryhot fluids may result in local vascular damageand other complications such as hemolysis.

Fluid Warming Devices (Table 1)Intravenous administration of large vol-

umes of inadequately warmed fluid can leadto significant hypothermia. Several methods towarm IV fluids are currently available. Thesemethods include immersing coiled IV tubingin a warm water bath, microwaving the bag offluid to be infused, adding heated saline toblood to be infused, passing the IV tubingthrough a heating block or through a plastictube warmed with forced air, passing the IV

Table 1. Commercially Available Warming Devices

Instrument Technology Comments

Flotem IIe Dry heat IV tubing sandwiched between heating plates

DW-100 Dry heat Plastic bag wrapped around heating cylinder

Fenwall Dry heat Plastic bag with channels sandwiched betweenheating plates

Dupaco Water bath Coiled IV tubing immersed in a bath

Level 1 H250 Countercurrent Tube in tube heat exchangewater bath

Level 1 H500 Countercurrent Tube in tube heat exchange, larger heater than H250water bath

Hotline Countercurrent Entire 254-cm patient IV line is warmed to ensurewater bath delivery of warm fluids at flow rates between 5 and

90 ml/min (300-5000 ml/hr)

Level 1 H1000 Countercurrent Tube in tube heat exchange combined with awater bath 254-cm patient IV line with Hotline characteristics

to prevent heat loss at moderate flow rates(<100 ml/min)

BairHugger Convective air Spiral IV tubing suspended in same convective2.4.1 warming hose that delivers forced air to a warming

blanket

FMS 2000 Magnetic High-speed volumetric pump with automatic airinduction detectors, line pressure sensor, and flow rate

control up to 500 mL/min; 122-cm patient line

R.I.S. Countercurrent High-efficiency pump with 3-L reservoir, three air/(Haemonetics) water bath bubble detectors, line pressure sensor, and

automatic flow rate control up to 1500 ml/min

Arrow In-line Direct microwave energy transferred in a heatingThermostat microwave chamber to coils of IV tubing wound on a900 disposable cartridge

Baxter Dry heat Disposable canister sets that fit over the heatingThermacyl unit; bubble trap to remove microbubbles

Mallinckrodt Countercurrent Metal foil cassette inserted between two heatedWarmflo metal plates; IV tubing inserts directly into metal fluidFW538 channels within the cassette

Alton Dean Countercurrent Metal foil cassette inserted between two heatedmetal plates; IV tubing inserts directly into metal fluid

channels within the cassette

Ranger Countercurrent Cartridge-style plastic disposable set insertedmetal between conductive warming plates

tubing through a conductive surface interfacedwith a counter-current heated water bath, mag-netic induction, and inline microwaving.25–29

The ideal fluid warmer should be capableof safely delivering fluids and blood productsat normothermia at both high and low flowrates. The ability of blood warmers to safelydeliver normothermic fluids over a wide rangeof flows is limited by several factors, includinglimited heat-transfer capability of materials

such as plastic, limited surface area of the heatexchange mechanism, inadequate heat trans-fer of the exchange mechanism at high flowrates, erythrocyte damage, and heat loss afterthe IV tubing exits the warmer.25,30–32 For ex-ample, adding warmed saline to blood couldhave catastrophic results unless the saline isnot heated above a certain temperature—themaximum safe temperature would be highlydependent on the relative volume of saline and


blood. The dangers of using unproven meth-ods and nonapproved approaches to bloodwarming cannot be overemphasized.

The heat-transfer capabilities of warmingdevices using dry heat exchange technology islimited by use of poorly conducting materialssuch as plastic and by limited heat transfer sur-face area. Warming devices that utilize counter-current heat exchange (Level 1 H250 [Fig. 1],H1000 [Fig. 2], and FW537) are capable ofwarming fluids even at very rapid flow rates dueto better conduction materials interposed be-tween the heating elements and the infusedfluid.32,33 Therefore, both these devices areappropriate for situations where rapid (>100ml/min) volume resuscitation is necessary.

At moderate flows (<100 ml/min), there issignificant heat loss after the IV tubing exits thewarmer. The continual countercurrent warm-ing of fluids in the tubing (Hotline [Fig. 3] andH1000) essentially eliminates the loss of heatalong the tubing distal to the warmer.33

Table 2 summarizes the implications of us-ing various fluid warmers during commonly en-countered clinical situations: pressure-driven in-fusion, and gravity-driven infusion with theroller clamp wide open.32,33 In the first situa-tion, the patient presents with severe circula-tory shock due to massive blood loss. Fluid re-suscitation via large-bore IV cannulas is requiredto prevent acidosis and irreversible shock. In

the second scenario, the fluid and blood vol-ume deficit is not as severe, although ongoingblood loss may necessitate moderately fast in-fusions with the roller clamp wide open to main-tain normovolemia and hemodynamic stability.It can be seen from the calculations in Table 1that the thermal stress of infusing cold fluidsmay result in considerable changes in meanbody temperature, especially if the patient isunable to increase heat production or preventfurther heat loss. The larger the gradient be-tween the temperature of the infused fluid and

Table 2. Implications of Using Warming Devices for CrystalloidFluid Resuscitation (5 and 10 L) in Anesthetized Adult Trauma Patients

For all devices, fluids were infused during two conditions—pressure-driven infusion andgravity-driven infusion with the roller clamp wide open. Data from references 32 and 33.

*Change in mean body temperature (MBT) was calculated as follows:

(Tfluid

- Tpatient

) Sfluid

/ Weight x Spatient

whereT

fluid = Outlet temperature of fluid delivered to the patient

Tpatient = Temperature of the patient, assumed to be 37 oCSfluid = Specific heat of infused fluid, 1 kcal/L/oCSpatient = Specific heat of the patient, 0.83 kcal/l/oC19,20

Weight of patient was assumed to be 70 kg

Device Flow Rate Outlet Decrease in Decrease in(ml/min) Temperature MBT* (5 L MBT* (10 L

infusion, °C) infusion, °C)

Flotem IIe Pressure 260 24 -1.12 -2.24 Gravity 90 27 -1.03 -2.06

Astotherm Pressure 260 25 -1.03 -2.06 Gravity 90 30 -0.60 -1.20

BairHugger2.4.1 Pressure 360 24.2 -1.10 -2.20 Gravity 80 29.6 -0.63 -1.27

Hotline Pressure 220 29.8 -0.62 -1.24 Gravity 80 34.8 -0.19 -0.38

Level 1 250 Pressure 600 33 -0.34 -0.69 Gravity 290 36 -0.09 -0.17

Level I H1000 Pressure 470 39.5 +0.22 +0.43 Gravity 150 39.4 +0.21 +0.41

FW537 Pressure 580 38.9 +0.16 +0.32 Gravity 200 39.9 +0.25 +0.49

CardioplegiaHeat Exchanger Pressure 700 35 -0.17 -0.34 Gravity 150 35 -0.17 -0.34

Figure 1.Schematic of the Level 1 250 and

500 warmer. The device consists of aheater that warms water and circulates

it through a pump and a heat-exchangesegment with a central tube for water flow(countercurrent heat exchange technology).

Fluid flows through the outer sheath,which surrounds the water core.

Note the filter and air eliminator distalto the heat exchanger.

core temperature, the greater the drop in meanbody temperature. As well, the greater the fluidrequirement relative to body weight, the greaterthe potential drop in body temperature.

Because of the marked inefficiencies ofconventional warming devices such as theFlotem IIe, Astotherm (Fig. 4), and others(Fig. 5), these devices are no longer in use atthe author’s institution and have been re-placed with the Level 1 H250 and H1000 forrapid infusion (>100 ml/min or 6 L/hr) andthe Hotline device for all other situations.


Safety of Rapid Infusion Devices withConstant Pressure

Because of the high flow rates generatedby newer warmers when used with constant-pressure devices, the limiting factor in fluidresuscitation is the time required to identifyred cell donor and recipient information, tospike and hang the fluid, and to ensure ab-sence of air from the fluid system. In the

Figure 4.Schematic of the Astotherm warmer.

This device consists of IV tubing coiledaround a circular heating element

(dry heat technology).

Figure 3a and b.Hotline warmer. This device consists of a water bath and an L-70 disposable.

The L-70 disposable heats fluid within the 254-cm patient line, which consists of acentral lumen for the IV fluid surrounded by an outer layer through which warm water

circulates down one side and then back up to the warm water reservoir in acountercurrent fashion (countercurrent heat exchange technology).

Figure 2a and b.Level 1 H1000 warmer. The device consists of a cylindrical aluminum heat exchanger

mounted on the warming unit and heated by a countercurrent water bath, similar to theLevel 1 250. After the fluid exits this first heat exchanger, it enters a 254-cm patient line

in which heat loss is prevented by surrounding the central lumen with warmed watercirculating in a countercurrent direction, similar to the Hotline device.

Figure 5.Schematic of the modified cardioplegia

heat exchange warmer. The device consistsof a water bath that circulates water

through a stainless steel cardioplegia heatexchanger in a countercurrent fashion. This

device is no longer used at the author’sinstitution because of delays in setup and

de-airing and high disposable costs.

author’s experience, it is wise to have one in-dividual solely responsible for pressurized in-fusion of fluids. This individual must utilizeextreme vigilance and caution because of thedanger of infusing air at these high flow rates.This author is aware of four cases of massiveair embolus at other institutions following useof pressurized infusions. Therefore, it is theauthor’s belief that constant pressurized infu-

sion devices not be used unless the patient isin profound hemorrhagic shock, and all air hasbeen removed from the fluid to be infused rap-idly. The automatic air eliminator incorporatedinto the design of the Haemonetic RIS andLevel 1 devices make these units somewhatsafer, but does not eliminate the risks of mas-sive air embolus.


SummaryAdverse consequences of perioperative

hypothermia include myocardial ischemia, car-diac arrhythmias, coagulopathy, shivering, in-creased oxygen consumption, alteration indrug metabolism and increased wound infec-tion. Administration of cold or inadequatelywarmed intravenous fluids contributes to hy-pothermia, whereas administration of normo-thermic fluids may reduce both the incidenceand complications of hypothermia. Therefore,infusion of adequately warmed fluids is impor-tant in order to minimize thermal stress andmaintain thermal homeostasis.

AcknowledgementThe secretarial assistance of Fran Hall is

very much appreciated.

References1. Danzl DF, Pozos RS. Accidental hypother-

mia. N Engl J Med 1994; 331:1756–60.2. Luna Gk, Maier RV, Pavlin EG, Anardi D,

Copass MK, Oreskovich MR. Incidence andeffect of hypothermia in seriously injuredpatients. J Trauma 1987; 27:1014–8.

3. Jurkovich GJ, Greiser WB, Luterman A,Curreri PW. Hypothermia in trauma vic-tims: an ominous predictor of survival. JTrauma 1987; 27:1019–24.

4. Gregory JS, Flancbaum L, Townsend MC,Cloutier CT, Jonasson O. Incidence andtiming of hypothermia in trauma patientsundergoing operations. J Trauma 1991;31:795–800.

5. Pavlin EG. Hypothermia in traumatized pa-tients. In Grande CM, ed. Textbook ofTrauma Anesthesia and Critical Care. St.Louis, Mosby-Year Book, 1993, chapter 94,pp 1131–9.

6. Little RA, Stoner HB. Body temperatureafter accidental injury. Br J Surg 1981;68:221–4.

7. Sessler DI. Mild perioperative hypother-mia. N Engl J Med 1997; 336:1730–7.

8. Smith CE, Patel N: Hypothermia in adulttrauma patients: Anesthetic consider-ations. Part 1, Etiology and Pathophysiol-ogy. Am J Anesthesiol 1996; 23:283–90.

9. Kurz A, Sessler DI, Lenhardt R.Perioperative normothermia to reduce theincidence of surgical- wound infection andshorten hospitalization. N Engl J Med1996; 334:1209–15.

10. Sessler DI. Consequences and treatmentof perioperative hypothermia. Anesth ClinNorth Am 1994; 12:425–56.

11. Watts DD, Trask A, Soeken K, et al. Hypo-thermic coagulopathy in trauma: effect ofvarying levels of hypothermia on enzymespeed, platelet function, and fibrinolyticactivity. J Trauma 1998; 44:846–54.

12. Frank SM, Higgins MS, Breslow MJ, et al.The catecholamine, cortisol, and hemody-namic responses to mild perioperative hy-pothermia. Anesthesiology 1995; 82:83–9.

13. Frank SM, Fleisher LA, Breslow MJ, et al.

Perioperative maintenance of normother-mia reduces the incidence of morbid car-diac events: a randomized clinical trial.JAMA 1997; 227:1127–34.

14. Boyan CP, Howland WS. Blood tempera-ture: a critical factor in massive transfu-sion. Anesthesiology 1961; 22:559–63.

15. Boyan CP, Howland WS. Cardiac arrest andtemperaturre of bank blood. JAMA 1963;183:58–60.

16. Boyan CP. Cold or warmed blood for massivetransfusions. Ann Surg 1964; 160:2882–6.

17. Bickell WH, Wall MJ, Pepe PE, et al. Imme-diate versus delayed fluid resuscitation forhypotensive patients with penetrating torsoinjuries. N Engl J Med 1994; 331:1105–9.

18. Martin RR, Bickell WH, Pepe PE, et al. Pro-spective evaluation of preoperative fluidresuscitation in hypotensive patients withpenetrating truncal injury. J Trauma 1992;33:354–62.

19. Mendlowitz M. The specific heat of humanblood. Science 1948; 107:97.

20. Gentilello LM, Cortes V, Moujaes S,Viamonte M, Malinin TL, Ho CH, GomezGA. Continuous arteriovenous rewarm-ing: experimental results and thermody-namic model simulation of treatment forhypothermia. J Trauma 1990; 30:1436–49.

21. Dubois EF. Basal Metabolism in Healthand Disease. Philadelphia, Lee andFebiger, 1924, p 324.

22. Uhl L, Pacini DG, Kruskall MS. The effectof heat on in vitro parameters of red cellintegrity. Transfusion 1993; 33:60S.

23. Gore DC, Beaston J. Infusion of hot crys-talloid during operative burn wound de-bridement. J Trauma 1997; 42:1112–5.

24. Uhl L, Pacini D, Kruskall MS. A compara-tive study of blood warmer performance.Anesthesiology 1992; 77:1022–8.

25. Presson RG, Bezruczko AP, Hillier SC,

McNiece WL. Evaluation of a new fluidwarmer effective at low to moderate flowrates. Anesthesiology 1993; 78:974–80.

26. Patel N, Smith CE, Pinchak AC, Hagen JF:Prospective, randomized comparison ofthe Flotem IIe and Hotline fluid warmersin anesthetized adults. J Clin Anesth 1996;8:307–16.

27. Smith CE, Holbrook C, Radesic B,Raghupathy A, Sweda S, Botero CA, PatelN, Punjabi A, Thompson L, Hagen JF,Pinchak AC. Comparison of perioperativeheating modalities in anesthetized adultpatients: a prospective randomized study.Am J Anesthesiol 1998; 25:62–8.

28. Smith CE, Desai R, Glorioso V, Cooper A,Pinchak AC, Hagen JF. Preventing hypoth-ermia: convective and intravenous fluidwarming versus convective warmingalone. J Clin Anesth 1998; 10:380–5.

29. Smith CE, Patel N: Hypothermia in adulttrauma patients: anesthetic consider-ations. Part II, Prevention and treatment.Am J Anesthesiol 1997; 24:29–36.

30. Fildes J, Fisher S, Sheaff CM, Barrett JA.Effects of short heat exposure on humanred and white blood cells. J Trauma 1998;45:479–84.

31. Herron DM, Grabowy R, Connolly R,Schwaitzberg SP. The limits ofbloodwarming: maximally heating bloodwith an inline microwave bloodwarmer. JTrauma 1997; 43:219–26.

32. Patel N, Smith CE, Pinchak AC: Compari-son of fluid warmer performance duringsimulated clinical conditions. Can JAnaesth 1995; 42:636–42.

33. Patel N, Knapke DM, Smith CE, NaporaTE, Pinchak AC, Hagen JF: Simulated clini-cal evaluation of conventional and newerfluid warming devices. Anesth Analg 1996;82:517–24.

Georges Desjardins, MD, FRCPCAttending AnesthesiologistBoca Raton, Florida, USA

Trauma is the most common cause of deathin Americans under the age of 45.1 In the UnitedStates, deaths from unintentional injuries aremost often the result of motor vehicle crashes,falls, poisoning, fires, or drowning. Althoughthe number of deaths from motor vehiclecrashes has decreased over the past few years,there has been an alarming increased in fire-arm-related deaths. If this trend continues,deaths from firearms are likely to exceed thosefrom motor vehicle crashes by the year 2003.1

Trauma anesthesiologists are faced todaywith sicker patients than in the past because of

Management of Massive Hemorrhage and Transfusion in Trauma

improvements in emergency prehospital care,initial resuscitation of trauma victims in emer-gency departments, and rapid transport to op-erating rooms. It is not uncommon to care forpatients with blunt injuries to the great vessels,penetrating injuries to the heart, severe bluntinjuries to the liver, severe open-book pelvicfractures, and penetrating injuries to the trunkand to then see these patients leave the hospi-tal to lead constructive, functional lives.

Hypotension and hypovolemia are gener-ally regarded as detrimental to the brain andother organs and are associated with worseoutcome, particularly in association with se-vere head injury. In recent reports, there isspeculation that hypovolemia and associatedhypotension are beneficial in some circum-

12


stances when hemorrhage is uncontrolled.2,3

The most frequently stated example is a lacer-ated major artery, when the administration offluid and associated increase in blood pressuremight dislodge a clot from the area of injury,increase the hemorrhage, and turn stable hy-potension into lethal recurrent hemorrhage.The evidence for the occurrence of these theo-retic effects from fluid resuscitation is stron-ger for penetrating trauma than for blunttrauma, which is more common in patientswith head injuries. Currently there is generalsupport for fluid administration as a mainstayof initial resuscitation after blunt trauma. Theinitial hemodynamic stabilization is still intra-venous access, correction of hypovolemia, andhemorrhage identification and control.

This review focuses on the managementof exsanguinating hemorrhage and massivetransfusion from blunt or penetrating traumaafter the patient’s arrival in the operating room.

DefinitionResuscitation of the severely injured patient

with fluids and blood products for hemorrhagicshock is often associated with complex meta-bolic alterations. Several definitions for massiveblood transfusions have been proposed.4,5

These range from the replacement of thepatient’s whole blood volume in 24 hours toreplacement of 50% of the volume in 3 hours.

When reviewing the physiologic conse-quences of massive transfusion, knowledge ofthese different definitions would seem essential,as they are quite different. Another definition thatwe would like to submit is the concept of mas-sive massive transfusion, which we define as thereplacement of a patient’s estimated blood vol-ume in less than 30 or 60 minutes. Certainly, themetabolic abnormalities associated with bloodreplacement and resuscitation in this type of pa-tient should be anticipated to be worse than forpatients who get 20 units of packed red bloodcells (PRBCs) in 24 hours.

Initial EvaluationAfter initial evaluation and resuscitation, the

trauma victim requiring surgery should be moni-tored during transport and be accompanied bymembers of the trauma team. The basic moni-toring devices for transport include an electro-cardiogram, automated blood pressure device,and pulse oximeter. End-tidal CO2 monitoringshould be considered if the patient is intubated.Ideally, the trauma team leader should transfercare directly to the anesthesiologist involved inthe case. Information accompanying the trans-fer should include mechanism of injury, injuriesthat have been identified, results and omissionsof investigations, medical history, and allergies.Giving this advance notice to the team in theoperating room before the actual arrival of thepatient will avoid delays and keep the focus oncontinuity and quality of care.6

After the arrival of an exsanguinating pa-tient in the operating room, the anesthesiolo-gist will have to modify his or her evaluation

and induction techniques to set new priori-ties and techniques for the resuscitation. Thismodification is the crash emergency anesthe-sia technique. It is in fact a combination of theATLS™ initial evaluation7 and the regular an-esthesia induction set-up. The first prioritieswill be evaluating and managing the airway,oxygenation, ventilation, followed by measur-ing the blood pressure; sorting out the intra-venous lines already in place; finding accessfor drug injection; attaching an ECG; infusingfluids through blood warmers; getting bloodin the room; checking the patient identity andhistory of allergy; placing an arterial catheter;drawing blood for blood gases, hematocrit, andother lab tests; titrating an anesthetic, if pos-sible; checking temperature and urine output;inserting a central venous or pulmonary arterycatheter or the TEE probe for monitoringneeds; and finally inserting a gastric tube.

The route for fluid administration intrauma is a source of controversy. There is gen-eral consensus that the first choice for cannu-lation is a vein that is visible, which most oftenmeans a peripheral vein on the upper extremi-ties. Two large-bore peripheral intravenouscatheters (16 gauge or larger) should be placedas quickly as possible for the administrationof fluids and blood. Using 14- or 16-gauge 2-inch peripheral catheters should allow a flowrate of 300 ml/min of crystalloid or 150 ml/min of blood when used in combination witha pressure bag.8 In areas with well-developedemergency medical systems, most trauma vic-tims arrive at the hospital with these intrave-nous catheters already in place.9

If peripheral intravenous access was un-successful in the field or in the resuscitationroom or if hypotension persists, additionalsites should be considered to ensure immedi-ate intravenous access. Some authors suggest,as a second choice, the cannulation of the ex-ternal jugular vein; as third choice, the use ofthe femoral vein; and as last choices, venouscutdown and catheterization through the in-ternal jugular or subclavian veins.9

If the patient does not have adequate in-travenous access, spinal precautions are still

being applied (cervical collar, backboard, triplefixation of the cervical spine), and unstable vitalsigns are present, several problems can be an-ticipated. Moving the head and neck or open-ing the cervical collar would be necessary toperform easy and timely cannulation of eitherthe external or the internal jugular vein. Re-moving some of the spinal precautions beforeclinical or radiologic clearance would not beideal and waiting for the radiologic evaluationwould be impractical. In these circumstances,the femoral vein could be an excellent secondchoice for venous access because of its largesize and easy access. However, a major con-cern with the use of the femoral vein as the“main IV” in the acute phase of resuscitationis the possibility of vascular injuries from theoriginal trauma in the pelvic and/or the abdomi-nal region, especially in patients with penetrat-ing trauma to the abdomen and in patients whohave sustained major pelvic fractures, in whomassociated vascular injuries are frequent. Rely-ing mainly on femoral access in this situationmight lead to loss of resuscitation fluid into theextravascular space. Use of a venous cutdownin the lower extremities has the same limita-tion. Although a venous cutdown in the upperextremities would avoid this problem, it is tech-nically more difficult and therefore often moretime consuming. When there is inadequate in-travenous access in the severely injured patientwith suspected intra-abdominal injuries, it is ourpractice to use the subclavian vein as our sec-ond choice for fluid administration6 (Table 1).In situations of advanced hypovolemic shockor exsanguination, where percutaneous tech-niques of IV insertion via peripheral centralveins are unsuccessful, venous cutdown at thesaphenofemoral junction may be used.10

The use of an 8.5 or 9.0 French introducerallows a flow rate higher than 500 ml/min withthe use of a pressure bag and large-caliber IVtubing.8,11 Strict aseptic technique should be usedeven in emergency situations. As a general rule,all intravenous catheters placed in the prehospitalphase and in the resuscitation room should bechanged in the first 24 hours after insertion, be-cause they may have been inserted under less-

Table 1. Intravenous Access in the Patient with Multiple Injuries

Option 1 — Peripheral IV x 2 in visible vein of the upper extremities

Option 2 — If unsuccessful, suggested second choice:

• If cervical spine injury is unlikely:External or internal jugular vein access with large-bore IV catheter

• If abdominal or pelvic injuries are unlikely:Femoral vein access with large-bore IV catheter

orVenous cutdown in the lower extremities

• If abdominal or pelvic injuries are suspected:Subclavian vein with large-bore IV catheter


than-ideal aseptic conditions.8 Our practice is toprovide the history of all IV lines to the ICU orward teams, who will then change all centralcatheters over a guidewire, culture the intracu-taneous segments and tip of the catheter with asemiquantitiative techniques, and remove theperipheral lines placed during the prehospitaland resuscitation phases of care.6 These changesshould be done only after relative hemodynamicstability has been established and/or additional“clean” intravenous access has been secured.

All intravenous fluids and blood productsshould be warmed. The H1000 infusion system(Level One Technologies, Inc., Rockland, Mas-sachusetts) is capable of infusing and heating800 ml/min of crystalloid or 500 ml/min ofblood. The Rapid Infusion System (RIS,Haemonetics Corporation, Braintree, Massachu-setts) can infuse blood products (red cells, freshfrozen plasma), crystalloids, or colloids at ratesup to 1,500 ml/min. This system is extremelyuseful in the management of exsanguinatinghemorrhage. The use of blood warming/high-volume infusion systems in addition to warm-ing the resuscitation room or operating roomto temperatures at high as 30°C is essential ifhypothermia is to be prevented effectively dur-ing resuscitation of the trauma patient.

Metabolic and Hemostatic Effects ofMassive Blood Transfusions

Since banked blood undergoes a numberof metabolic and structural changes over time,multiple severe derangements of physiologyare theoretically possible when large volumesof banked blood are given to critically ill orinjured patients. Although the volume of bloodtransfused may lead to a variety of problems(Table 2), both the depth and duration of shockappear to be more significant determinants ofphysiologic derangements than the transfusionof blood itself.12 If the patient receiving mas-sive transfusion receives adequate fluid resus-citation and maintains oxygen delivery andorgan perfusion, the sequelae of massive trans-fusion may be minimized. The volume of bloodproducts that the patient receives should notbe the primary determinant of therapeuticdecisions or prognosis.13–15

The ability to provide massive transfusionis a relatively recent medical accomplishmentresulting from a series of advances (large bloodbanks, rapid infusers of warm fluids, and bet-ter understanding of the physiology of trans-fusion). The varied definitions of massive trans-fusion, the numerous associated clinical con-ditions, and the relative lack of detailed rigor-ous studies have crated controversy in the lit-erature regarding the metabolic effects of mas-sive transfusion. The confusion is com-pounded by the use of blood of varied storagelife, nonuniform resuscitation protocols, andcomparison of patients suffering from shockof differing severity and duration.

The storage and refrigeration of pRBCsresults in progressive changes that are termedstorage lesions.16 The change in deformability

and increased hemolysis is linked to the de-creased levels of intracellular ATP. This in turnis linked to the increased levels of potassium,ammonia, and hemoglobin in the supernatantplasma or preservative solution. The changein oxygen affinity of hemoglobin for oxygenis, in large part, a consequence of decreasedlevels of intracellular 2,3-DPG. The increase invasoactive substances is a result of their releasefrom leukocytes and platelets contained in theblood or red cell concentrate. Finally, the de-velopment of microaggregates is due to theformation of small amounts of fibrin strandsduring storage and the adherence of senescentplatelets and leukocytes to them.

Massive transfusion of blood componentscontaining sodium citrate can lead to transientlydecreased levels of ionized calcium. Hypocal-cemia can cause hypotension, narrowed pulsepressure, and biventricular dysfunction. Elec-trocardiographic abnormalities such as pro-longed QT interval can occur. Adults who havenormal hepatic function, are normothermic,and are not in shock can tolerate the infusionof one unit of PRBCs every 5 minutes (20 units/hr) without developing hypocalcemia.17

Since stored blood commonly has el-evated potassium concentration, up to 30 to40 mEq/L by 3 weeks of storage, hyperkalemiais possible with massive transfusion. Hyper-kalemia may cause elevated peaked T waveson the electrocardiogram. It can significantlyalter cardiac function, especially if associatedwith hypocalcemia. The incidence of intraop-erative hyperkalemia increases with infusionrate of PRBCs above 150 ml/min. Hyperkale-mia can be treated early with intravenous cal-cium, insulin, and bicarbonate and with PRBCwashing before administration.18

Although stored PRBCs have an acid pH(about 6.3), alkalosis is the usual result of mas-sive transfusion without shock. Sodium citratecontained in the anticoagulant is converted tosodium bicarbonate in the liver. The alkalosisinitially increases the oxygen affinity of hemo-globin, resulting in less oxygen off-loading tothe tissues. The clinical significance of this al-kalosis is unknown.

Hypothermia may occur with rapid trans-fusion of large volumes of cold blood compo-nents. It remains the most under-recognizedand under-treated cause of coagulopathy intrauma patients.19 It increases the affinity of

hemoglobin for oxygen and impairs clottingfunction. Low temperature also increases thepotential for hypocalcemia because of de-creased hepatic metabolism of citrate. Preven-tion of hypothermia is essential and can beachieved by warming intravenous fluids andblood during administration, warming theoperating room to 30°C, and using convectivewarming blankets in all cases of severe trauma.

As the amount of blood replacement in-creases, the trauma patient’s own blood be-gins to take on characteristics of bank blood,with low levels of 2,3-DPG and low activitiesof Factor V and VIII, as well as dilutional throm-bocytopenia. When blood is stored at 4°C for24 to 48 hours, the platelets have only 5% to10% of normal activity. Following transfusion,these platelets are essentially nonfunctional.The massive transfusion of packed RBCs willrapidly dilute the patient’s existing plateletpool. The decrease is often less than expectedon the basis of simple dilution because of somerelease of platelets from the spleen and bonemarrow. Prompt platelet administration shouldbe considered once abnormal bleeding isnoted. In the patient who has microvascularbleeding without hypothermia, a platelet countbelow 50,000/µl or a falling count below100,000/µl indicates the need for platelet trans-fusion. Indications for fresh frozen plasma(FFP) and cryoprecipitate are not clear. Intrauma patients who receive between one andtwo blood volume replacement, dilutionalthrombocytopenia and fibrinogen levels below75 mg/dl often occur.20 Low levels of coagula-tion Factors V and VIII are usually a clinicalproblem after two blood volume replacement.Fibrinogen can be replaced with FFP or cryo-precipitate. In trauma patients, low coagula-tion factors are usually replaced with FFP.

In addition to the metabolic changes ob-served with massive transfusion, infectious andimmunologic effects can complicate the care oftrauma patients. Viral hepatitis remains themajor infectious risk of transfusion. With bet-ter donor blood screening in the United States,the estimated risks (per unit of blood trans-fused) of transmission of viral infection are asfollows21: HIV, 1:493,000; hepatitis B, 1:63,000;hepatitis C, 1:103,000; and HTLV, 1:641,000.(See Chapters 8 and 9.) Transfusion has thepotential to modify the recipients’ immune re-sponse. This is a potentially serious problem inmany survivors of massive transfusion, who gen-erally develop immune compromise and are athigh risk for sepsis and multiple organ failure.

Management of Massive TransfusionOne thing is clear: the goal of hemorrhagic

shock resuscitation is prompt restoration ofadequate perfusion and oxygen transport. Theobjective of resuscitation is to reestablish oxi-dative metabolism by providing adequate oxy-gen flow to cells, preventing reperfusion dam-age, and avoiding blood loss.

Patients in hemorrhagic shock developlow pH from the buildup of intracellular hy-

Table 2.Metabolic and Hemostatic Effects

of Massive Blood Transfusions

Decreased oxygen dissociationHypocalcemiaHyperkalemiaDerangement of acid–base balanceHypothermia (<35°C)Dilutional coagulopathy (platelets,

coagulation factors)


drogen ions, which occurs during the anaero-bic conversion of glucose to lactate. Some ofthe intracellular lactate and associated hydro-gen ions eventually leave the cell and producethe characteristic metabolic acidosis of hem-orrhagic shock. The pH, lactate level, and basedeficit are highly correlated with mortality andare thought to be an underlying cause of de-creased cardiac contractility and eventual mor-tality. However, the clinical hemodynamic con-sequences of low serum pH are unclear. Manyclinicians give bicarbonate to increase cardiaccontractility. There is some evidence that con-tractility does not decrease substantially untilthe pH is 6.9 or 6.8, unless adequate oxygen isnot available.22 The most significant determi-nants of depressed cardiac contractility inshock appear to be hypercarbia and hy-poxia.23,24 Clinically, if perfusion has been re-stored, oxygen delivery is adequate, and thepatient is well ventilated, pH correction withexogenous bicarbonate is unnecessary.

Conventional fluid warmers, such as thosein which fluid (crystalloid, colloid, or blood)is passed within plastic tubing through heat-ing blocks or those in which the tubing is sub-merged in warm water, are inefficient in deliv-ering normothermic fluids at fast flow rates(≥ 250 ml/min). With aggressive fluid resusci-tation and blood transfusions, clinicians areconfronted with five distinct problems: hypo-volemia, hypothermia, coagulopathy, hyper-kalemia, and hypocalcemia. Fluid warmers aredesigned to prevent and treat some of theseproblems. The H1000 infusion system (SimsLevel One Technologies, Inc., Rockland, Mas-sachusetts) is a very effective fluid-warmingdevice. It consists of a cylindric aluminum heatexchanger mounted on the warming unit andheated by a countercurrent water bath with aset point of 42°C. To decrease heat loss evenmore, a second device can be added— theHotline warmer (Sims Level One Technolo-gies)—on the 254-cm line between the H1000and the patient. The central lumen of the in-travenous line is warmed by water circulatingin a countercurrent direction. The countercur-rent circulation water is warmed by a heatedreservoir, with a set point of 42°C.

Countercurrent water fluid warmers us-ing 42°C set points do not damage red cells,deliver warm intravenous fluids, and allow theclinician to maintain thermal neutrality withrespect to fluid management up to 400 ml/min.With flow rates above that, the infusion fluidtemperature will decrease slightly in propor-tion to the increase in flow rate.

The H1000 infusion system is very usefulfor resuscitation of trauma victims, as it deliv-ers warm fluid at rapid rates. It takes care ofhypovolemia and prevention of hypothermiavery well. Unfortunately, it is difficult to de-liver more than 800 ml/min with this infusionsystem. To achieve infusion rates above thislevel, our practice is to use the Rapid InfusionSystem (RIS, Haemonetics). This device is ca-pable of delivering 1,500 ml/min of blood

products at normothermia. At our institution,the system is primed with a crystalloid solu-tion and blood products are added to the 3-liter reservoir as indicated during the resusci-tation. Platelets are not infused with the RISdevice. They are infused through a separateintravenous access. The usual ratio of bloodproducts used with the RIS follows the Uni-versity of Pittsburgh protocol, with 2 units ofpacked red cells (600 ml), 2 units of FFP (400ml), and 500 ml of a colloid or crystalloid so-lution. The hematocrit of this solution is 28%.All blood is filtered through a 150-micron fil-ter as it is introduced into the reservoir. It thenpasses through a 40-micron filter. The heatexchanger system also uses countercurrenttechnology. The fluid is infused with the aid ofa roller pump from a minimal rate of 10 ml/hrto a maximum of 1,500 ml/min. To our knowl-edge, at present, no other infusing system candeliver normothermic units at this rate.

The use of the RIS has introduced newproblems during resuscitation of trauma vic-tims. Although coagulopathies, hyperkalemia,and hypocalcemia have been well described inthe literature as rare phenomena, we have no-ticed a high incidence of them after massivetransfusions. As discussed previously,coagulopathies and hypocalcemia are wellknown problems associated with rapid andmassive transfusions. Hyperkalemia is a rela-tively new phenomenon. Its incidence is highwhen using flow rates of 500 to 1,000 ml/min.As described by Jameson et al,18 for preventionof transfusion-associated hyperkalemia, ourpractice is to use the Haemonetics Cellsaverblood salvage system in combination with theRIS. The Cellsaver system is used not only torecycle blood from the surgical field but also,and more importantly, to wash the blood bankPRBCs before transfusion to the trauma victim.Washing the PRBCs decreases the H+ and K+concentrations of the blood transfused and, inour experience, decreases the incidence of se-vere transfusion-associated hyperkalemia.

References1. Capan LM, Miller SM. Trauma and burns.

In Barash PG, Cullen BF, Stoelting RK, eds.Clinical Anesthesia, 3rd edition. Philadel-phia, Lippincott-Raven, 1997, pp 1173–204.

2. Bickell WH, Wall MJ, Pepe PE, et al. Imme-diate versus delayed fluid resuscitation forhypotensive patients with penetrating torsoinjuries. N Engl J Med 1994; 331: 1105–9.

3. Martin RR, Bickell WH, Pepe PE, et al. Pro-spective evaluation of preoperative fluidresuscitation in hypotensive patients withpenetrating truncal injury: a preliminaryreport. J Trauma 1992; 33:354–62.

4. Rutlege R et al. Massive transfusion. CritCare Clin 1986; 2:791–805.

5. Wudel JH et al. Massive transfusion: out-come in blunt trauma patients. J Trauma1991; 31:1–7.

6. Desjardins G, Varon AJ. Immediateintrahospital management. In Abrams KJ,Grande CM, eds. Trauma Anesthesia and

Critical Care of the Neurological Injury.Futura Publishing, 1997, pp 95–120.

7. American College of Surgeons. AdvancedTrauma Life Support Student Manual. Chi-cago, American College of Surgeons, 1993.

8. Palter MD et al. Secondary triage of thetrauma patient. In Civetta JM, Taylor RW,Kirby RR, eds. Critical Care. Philadelphia,Lippincott, 1992, pp 611–25.

9. Calcagni De et al. Resuscitation: blood,blood component and fluid therapy. InGrande CM, ed. Textbook of Trauma An-esthesia and Critical Care. St. Louis,Mosby, 1993, pp 381–416.

10. Rogers FB. Technical note: a quick andsimple method of obtaining venous accessin traumatic exsanguination. J Trauma1993; 34:142–3.

11. Milikan JS et al. Rapid volume replace-ment for hypovolemic shock: a compari-son of techniques and equipment. JTrauma 1984; 24:428.

12. Collins JA. Recent developments in thearea of massive transfusion. World J Surg1987; 11:75–81.

13. Canizaro PC, Pessa ME. Management ofmassive hemorrhage associated with ab-dominal trauma. Surg Clin North Am1990; 70:621–34.

14. Patterson A. Massive transfusion. IntAnesthesiol Clin 1987; 25:61–74.

15. Practice Guidelines for Blood ComponentTherapy. A report by the American Soci-ety of Anesthesiologists Task Force onBlood Component Therapy. Anesthesiol-ogy 1996; 84:732–47.

16. Lovric V. Alterations in Blood componentsduring storage and their clinical significance.Anaesth Intensive Care 1984; 12:246–51.

17. Denlinger JK et al. Hypocalcemia duringrapid blood transfusion in anaesthetizedman. Br J Anaesth 1976; 48:995.

18. Jameson LD et al. Hyperkalemic deathduring use of high-capacity fluid warmerfor massive transfusion. Anesthesiology1990; 73:1050–2.

19. Wilson RF et al. Electrolytes and acid-basechanges with massive blood transfusion.Am Surg 1992; 58:535–45.

20. Murray DJ et al. Coagulation changes dur-ing packed red cells replacement of majorblood loss. Anesthesiology 1988; 69:839.

21. Schreiber GB, Busch MP, Kleinman SH,Korelitz JJ. The risk of transfusion-trans-mitted viral infections. The Retrovirus Epi-demiology Donor Study. N Engl J Med1996; 334:1685–90.

22. Downing SE et al. Influences of hypox-emia and acidemia on left ventricular func-tion. Am J Physiol 1966; 210:1327–34.

23. Siegel HW, Downing SE. Contributions ofcoronary perfusion pressure, metabolicacidosis and adrenergic factors to the re-duction of myocardial contractility duringhemorrhagic shock in cats. Circ Res 1970:27:875–89.

24. Prezlosi MP et al. Metabolic acidemia withhypoxia attenuates the hemodynamic re-sponses to epinephrine during resuscitationin lambs. Crit Care Med 1993; 21:1901–7.


Jeffery R. Jernigan, MDJohn G. D’Alessio, MDElvis Presley Memorial Trauma CenterMemphis, Tennessee

For multiple trauma patients with massivehemorrhage presenting for surgery, preopera-tive efforts have been directed at stabilizing (orat least temporizing) hemodynamic status. Ob-taining adequate intravenous access and infus-ing crystalloid and packed red blood cells areimportant measures in supporting circulatingblood volume. However, anesthesiologists arestill faced with precarious situations in whicheither all the above has taken place in the faceof ongoing hemorrhage, or some stabilizationhas occurred but surgical management will, ofnecessity, entail increased blood loss. In eithercase, the clinical sequelae of hemorrhage andshock (such as acidosis, hypothermia, andcoagulopathy) will begin to present at this point,problems that become all the more difficult ifnot managed early and effectively.

In this section, we will discuss our expe-rience in the clinical management of patients’problems regarding massive transfusion. Thisdiscussion is not intended to represent a de-finitive management protocol, since muchdebate continues about such topics as appro-priate resuscitation strategies, desired clinicalend-points, and proper use of blood products.Rather, it is a “walk through” of the questions,trials, and decision-making processes that haveled us to our current use of the Rapid InfusionSystem (RIS) (Haemonetics Corporation,Braintree, Massachusetts) in conjunction withpoint-of-care chemistry-testing devices in man-aging these difficult problems.

When our trauma center opened in 1983,we were using standard pressure bags con-nected to a pneumatic pump with six outletsand infusing fluids through a separate bloodwarmer. Although this approach was adequatefor several years, we were constantly strugglingwith problems of acidosis, hypothermia, andcoagulopathy in the face of ongoing and, occa-sionally, exsanguinating hemorrhage. Therefore,we began looking for ways to improve our abil-ity to keep up with massive hemorrhage.

We initially considered the fluid-warmingpressure infusers manufactured by Level I(Level I Technologies, Rockland, Massachu-setts). This system consisted of two pressureinfusers connected to a blood warmer we hadalready been using. The advantages of this sys-tem were ease of use and portability. However,only two pressurized bags could be connectedto this system at any one time. Infusion rateswere comparable to or slightly faster than thepneumatic pumps used previously (approxi-mately 500 cc/min).

The only commercially available system

Rapid Infusion and Point-of-Care Chemistry Testing in Massive Transfusion: Avoiding Common Pitfalls

specifically developed for volume infusion >500cc/min is the RIS. This device utilizes rollerpumps that propel fluids from a 3-liter reser-voir through two limbs of high-capacity tubingat rates of up to 1,500 cc/min. The system alsodelivers 100-cc or 500-cc boluses over 1 minute.Additionally, there are three air detectors, whichautomatically stop the infusion in the event ofbubbles in the infusion path. While some plan-ning and a brief set-up period of 3 to 5 minutesare required, it was readily apparent that sig-nificantly greater volumes of fluid could be in-fused in a short time. However, this device isrelative large and expensive, and it requiresmaintenance of an adequate supply of dispos-able tubing/reservoir set-ups.

We currently employ both of these sys-tems, the Level I System 1000 being the morewidespread of the two, with units in each op-erating room (OR), shock trauma admitting,and the intensive care unit. The combinationof the two systems has proven very useful, theLevel I being used perioperatively, with theoption of large-volume infusions with the RISif need for massive transfusion arises in theoperating room.

Questions arose when we began using theRIS routinely in the OR. We found we had al-tered the dynamics of blood administration inour trauma OR, in that we were no longer the“rate-limiting step.” The blood bank raisedconcerns regarding appropriate use of bloodproducts and maintenance of an adequate sup-ply of these valuable resources. Additionally,some of our surgical colleagues expressed con-cern about striving for normotension with ag-gressive fluid administration and the effects thismay have on hemostatis. These issues were adirect result of our dramatically increased abil-ity to infuse large volumes.

Other questions arose regarding some ofthe problems well known to be associated withmassive transfusion,1-5 which are discussedelsewhere in this monograph. We noted clini-cally significant hyperkalemia on at least oneoccasion. Such related complications previ-ously thought to be infrequent were now morelikely to be encountered as infusion capabilityincreased.1

As we worked through these issues, Hamblyand Dutton concluded that using the RIS wasassociated with increased mortality. They alsoasked the question (raised by others7–11) whetherhypotensive resuscitation may be advantageousin this setting. This followed the article by Dun-ham and associates,12 which showed a positiveoutcome associated with fluid administrationthrough the RIS. These considerations led to areassessment of our use of the RIS.

Despite the problems we encountered, wefelt there were distinct advantages in using theRIS. The primary, overriding advantage is the

dramatically improved ability to maintain cir-culating blood volume. The ease and efficiencywith which these volumes are administered al-lows the anesthesia team to devote more men-tal and physical energy toward other criticalaspects of the case in progress. Further, withthe RIS there is much greater flexibility in therates of infusion. If one accepts the notion thathypotensive resuscitation is desirable, thiswould appear to be all the more reason to usethe RIS in such a scenario. In addition, the abil-ity of the RIS to arrest and reverse hypothermiato the point of warming a cold patient to nor-mothermia is significant and cannot be ignored.

Thus, it was evident that the RIS possessesseveral undeniably desirable characteristics.Indeed, when one considers the five majorproblems encountered during massive trans-fusion (hypovolemia, hypothermia,coagulopathy, hyperkalmia, and hypocalce-mia), our experience has been that the RISaddresses hypovolemia and hypothermia effec-tively which, in turn, has beneficial effects indealing with acidosis and coagulopathy.2 How-ever, we concluded the increased risks of sig-nificant hyperkalemia and hypocalcemianeeded to be addressed separately.

Since there is greater risk of physiologicderangement in this setting, we felt a need forcloser monitoring of physiologic parametersby laboratory tests. To obtain turnaround timesfaster than the hospital laboratory could pro-vide, we considered point-of-care testing de-vices. Point-of-care testing has gained favor inrecent years, one example being glucometersdeveloped for home use, which enable diabet-ics to monitor their glucose levels. Newer tech-nologies have expanded this concept intoother areas involving a variety of laboratory pa-rameters relevant to intensive care and surgi-cal settings.

After discussion with our laboratory direc-tor, we chose the i-STAT Portable Clinical Ana-lyzer (i-STAT Corp, Princeton, New Jersey). Thisdevice is hand-held and battery powered andcomes with a portable printer. It is easy to useand relatively inexpensive and provides reliableaccurate results in 2 minutes. The system em-ploys a “thin film” biosensor housed in a smallcartridge. Two to three drops of blood are placedinto the cartridge, which is inserted into the ana-lyzer. The lab values obtained depend on theparticular cartridge used. There are several typesavailable. The cartridge we use measures sodium,potassium, ionized calcium, arterial blood gases,hematocrit, and hemoglobin. There were earlyconcerns about the biosensor technology regard-ing manufacturing and failure rate. We have hadno problems in these areas. However, the i-STATdoes not provide point-of-care testing for coagu-lation studies, so we continue to send these toour trauma laboratory.

13


In addition to federally mandated qualityassurance guidelines, there are a number ofpoint-of-care testing guidelines, which varyfrom state to state. Federal guidelines were setforth in the Clinical Laboratory ImprovementAct of 1967 and amended in 1988. The cur-rent rules and regulations are referred to asthe CLIA ’88 (Clinical Laboratory ImprovementAmendments of 1988). They divide laboratorytests into three categories: 1) waived (no spe-cial qualifications to run tests); 2) moderatelycomplex (requires high school diploma); and3) highly complex (requires an associate de-gree in laboratory science). The federal gov-ernment may inspect, fine, and even close fa-cilities found not to be in compliance.13 Aninstitution that performs laboratory tests isresponsible for compliance regardless of wherewithin the facility that testing is done. In addi-tion, four states (California, Florida, New York,and Tennessee) require that anyone not a cer-tified medical technologist (including MDs andCRNAs) must be granted a waiver in order torun lab tests. Thus, in order to avoid thesetypes of problems, we recommend consultingthe lab director of your institution if one ofthese devices is being considered.

Another option available in avoiding com-plications of massive transfusion is washing redblood cells (RBCs) prior to infusion. Storageof packed red blood cells (PRBCs) results inaccumulation of potassium over time.14 Wash-ing RBCs prior to administration removesmuch of this potassium as well as a significantproportion of existing citrate, which, in somecases, can result in hypocalcemia and cardio-vascular depression.15 The removal of theseagents can preempt some of the problems as-sociated with massive transfusion. This optionhas been employed successfully in a variety ofclinical settings.1,16,17 We do not perform thisroutinely, except when treating patients witha history of renal insufficiency.

In using the RIS in conjunction with the i-STAT, we employ the following strategy whenmassively transfusing a patient:

* When the decision is made to use the RIS,we notify the blood bank than an RIS caseis starting. The blood bank then sets upwhat are termed “RIS units” consisting of10 units PRBCs, 4 units FFP, and 4 units ofplatelets (not to be infused through theRIS). The blood bank continues to holdone of these units until informed by usthat we are no longer in a massive trans-fusion mode.

* Baseline labs are drawn, consisting of ar-terial blood gasses, complete blood count,PT/PTT, fibrinogen, potassium, and ion-ized calcium.

* In filling the pump reservoir, PRBCs arediluted with 500 cc normal saline per unit.

* We aim for a hematocrit in the low tomid-20s.

* FFP are infused through the RIS in addi-tion to the NS.

* Platelets are infused separately.

* With each five units of packed cells givenin 15 minutes or less, 1 gram of CaCl

2 is

given.

* Labs are repeated after each 10 unitsPRBCs.

* Hyperkalemia (>6.0) is treated with 10units regular insulin with D5W.

* Acidosis is treated with volume infusionand sodium bicarbonate as deemed ap-propriate.

* Cryoprecitipate is given based on fibrino-gen levels.

* We continue to strive to maintain a rela-tively normotensive state in this setting.Communication with the surgical team,monitoring of urine output, and consid-eration of cerebral perfusion help guidedecisions regarding target pressures.

As to the controversies concerning hypoten-sive resuscitation, use of the RIS in this scenario,and possible increased mortality associated withits use, close examination of the pertinent litera-ture led us to the following analysis.

The conclusions reached in the study byHambly and Dutton are clouded by two prob-lems. First, selection bias may have played asignificant role, as noted by the authors. Sec-ond, their findings are predicated on a com-parison of expected versus observed mortalitybetween the study groups. They defined ex-pected mortality in this population based on alogistic regression equation published by Dun-ham et al from their institution in 1986.18 Thisequation was written as a statistical descriptorof observed events at that institution, not as apredictor of mortality. They state, “To ensurevalidity of the equation used to determine theprobability of death, a prospective assessmentneeds to be performed on another popula-tion.” A search of the literature and conversa-tions with the author have not revealed such astudy. Therefore, the applicability of this equa-tion in predicting mortality in this populationmust be questioned. Such an equation or simi-lar predictive tool remains elusive.

Regarding the question of hypotensiveresuscitation, it should be noted that the studyby Bickell et al deals with penetrating trauma,whereas Hambly and Dutton raise this issue intheir study on blunt trauma patients. Further,Bickell found increased survival with minimalresuscitation prior to, not in, the operatingroom, and full resuscitation once surgical con-

trol of blood loss was obtained. This would notappear to be applicable to intraoperative use ofthe RIS as studied by Hambly and Dutton, sincetheir patients were, presumably, resuscitated inthe usual fashion prior to and after arrival atthe Shock Trauma Center. Before conclusionscan be drawn regarding the appropriatenessand timing of use of the RIS, more uniformitybetween Bickell’s and Dutton’s patients wouldhave to be demonstrated.

Unanswered questions remain, along withthe need for further controlled, well-focusedstudies. Whatever strategy is employed duringfluid resuscitation of the trauma patient andmassive transfusion, it is important to remem-ber to treat each patient individually, globally,and according to clinical judgment rather thanby strict protocol. Use of the RIS together withpoint-of-care testing and improved communi-cation with blood bank personnel, laboratorypersonnel, and surgeons improves our abilityto manage trauma patients requiring massivetransfusion.

References1. Jameson LC et al. Hyperkalemic death

during use of a high-capacity warmer formassive transfusion. Anesthesiology 1990;73:1050–2.

2. Ferrara A et al. Hypothermia and acidosisworsen coagulopathy in the patient requir-ing massive transfusion. Am J Surg 1990;160:515–8.

3. Phillips GR et al. Massive blood loss intrauma patients: the benefits and dangersof transfusion therapy. Post-GraduateMedicine: Transfusion Therapy 1994;95(4):61–70.

4. Hamilton SM. The use of blood in resus-citation of the trauma patient. Can J Surg1993; 36(1):21–7.

5. Wilson RF et al. Electrolyte and acid-basechanges with massive blood transfusions.Am Surg 1992; 58(9):535–45.

6. Hambly PR, Dutton RP. Excess mortalityassociated with the use of a rapid infusionsystem at a level I trauma center. Resusci-tation 1996; 31:127–33.

7. Bickell WH et al. Immediate versus de-layed resuscitation of hypotensive patientswith penetrating torso injuries. N Engl JMed 1994; 331:1105–9.

8. Bickell WH. Are victims of injury some-times victimized by attempts at resuscita-tion? Ann Emerg Med 1993; 22:225–6.

9. Bickell WH et al. Intravenous fluid admin-istration and uncontrolled haemorrhage.J Trauma 1989; 38:227–33.

10. Stem A et al. Effect of blood pressure ofhaemorrhagic volume in a near-fatalhaemorrhage model incorporating a vas-cular injury. Ann Emerg Med 1993;22:155–163.

11. Capone A et al. Treatment of uncontrolledhaemorrhagic shock: improved outcomewith fluid restriction. J Trauma 1993;35:984.


12. Dunham CM et al. The Rapid InfusionSystem: a superior method for the resus-citation of hypovolaemic trauma patients.Resuscitation 1991; 21:207–27.

13. Passey RB. CLIA ’88 penalities and how toavoid them. In Coping with CLIA: An 11-Part series. Medical Laboratory Observer.Medical Economics Publishing, June 1993.

14. Estrin JA et al. A new approach to massive

blood transfusion during pediatric liverresection. Surgery 1986; 99(5):664–9.

15. Westphal RG. Special topics. In WestphalRG, ed. Handbook of Transfusion Medi-cine, 3rd ed. The American Red Cross,1996; p 106.

16. Kang YG. Hemodynamic instability duringliver transplantation. TransProc 1989;21(3):3489–92.

17. Ramsay AE, Swygert TH. Anesthesia forhepatic trauma, hepatic resection and livertransplantation. Balliere’s Clinical Anes-thesiology 1992; 6:863-94.

18. Dunham, CM, Cowley R, Gens DR, et al.Methodologic approach for a large func-tional trauma registry. Md Med J 1989;38:227–33.

SECTION IV: New Horizons in Synthetic Blood Substitutes

Hemoglobin-Based Oxygen-Carrying Solutions & Hemorrhagic ShockColin F. Mackenzie, MB, ChB, FRCA, FCCMDirector, National Study Center for Traumaand Emergency Medical SystemsUniversity of Maryland School of MedicineBaltimore, MD 21201 USAe-mail: [email protected]

[Editors’ note: Dr. Mackenzie receives grantsupport from Biopure Corporation andAnjinomoto Corporation.]

There has been only one reported use, in1949, of a hemoglobin solution for resuscita-tion of a human in hemorrhagic shock.1 Awoman suffering from postpartum hemorrhagewas given 2.3 liters of 9% hemoglobin solutionin saline after all available compatible blood hadbeen given. Consciousness returned, her bloodpressure rose, and her heart rate fell. However,the patient died 9 days later from renal failure.

Attempts to develop blood substitutes goback many hundreds of years2 (Table 1). In1916, hemoglobin solutions were given insmall quantities to 33 subjects to determinethe renal threshold for hemoglobin withoutadverse effects. Many studies, however, usinglarger quantities of hemoglobin solutions, hadadverse effects, including hypertension, brady-cardia, oliguria, and anaphylaxis.3 In 1957,Chang encapsulated hemoglobin,2 and sincethen development of liposome-encapsulatedhemoglobin has continued. The problems as-sociated with disposal of the encapsulated

hemoglobin and stimulation of the reticuloen-dothelial system and macrophages have notbeen resolved. Leland Clark demonstrated thata mouse could survive while breathing liquidperflurocarbons saturated with oxygen.5 In1972, Benesch discovered reagents that couldbind the 2,3-DPG binding site so that theycould reduce hemoglobin affinity for oxygen.The most widely used agent is pyridoxal, 5 PO4

(so-called pyridoxalation), which is used toreduce oxygen affinity.6 The normal P5O (thepartial pressure of oxygen when Hb is 50%saturated) of blood is 26.7 mmHg. P5O is in-creased by pyridoxalation. Human stroma-freehemoglobin has a P5O of 12 to 15 mmHg andtherefore has a high oxygen affinity and tendsto hold onto the oxygen rather than give oxy-gen up at the tissue level.

General PropertiesRed cells can be stored in liquid form with

citrate phosphate dextrose adenine (CPDA)anticoagulant for 35 days and in AS-1 for 42days. They can also be frozen after addition ofglycerol to prevent lysis or they can be instantlyfreeze-dried or lyophilized. Oxygen-carryingsolutions (Table 2) may consist of free hemo-globin from which the stroma or cell wall hasbeen removed, or liposome-encapsulated he-moglobin containing hemoglobin with a syn-thetic membrane. Perfluorocarbons are organicsolutions with high oxygen solidity.

Toxicities of free hemoglobin solutions

(Table 3) include vasoactivity, with binding ofnitric oxide by free hemoglobin being the mainsuspect causing vasoconstriction.7 Nephrotox-icity from stromal remnants is probably of onlyhistorical interest, because better purificationtechniques have resulted in lack of renal tox-icity with newer hemoglobin-based oxygencarriers.8 In human volunteers given recombi-

Table 1. History of Transfusion and Oxygen-Carrying Solution Use

1667 First human blood transfusion (Denis), causing death and moratorium

1863 Gum-saline transfusion (Ludwig)1916 Hemoglobin infusion in humans (Sellards and Minot)

1941–45 Albumin and hemoglobin infusion

1957 Encapsulated hemoglobin described (Chang)

1966 Perfluorocarbon “bloodless mouse” (Clark and Gollman)1969 Amberson’s report of hemoglobin infusion in human hemorrhagic shock

1972 Pyridoxalation to reduce hemoglobin affinity (Benesch et al)1978 Human safety trial, unmodified hemoglobin (Santsky et al)

1980–97 Human trials with human and bovine hemoglobin-based solutions

Human trials with second-generation perfluorocarbons

Table 2.Currently Available Products

That Can Be Used As Oxygen-CarryingSolutions in Humans

Whole blood

Liquid red cellsFrozen red cellsLyophilized red cellsFree hemoglobin

Encapsulated hemoglobinPerfluorocarbons

Table 3.Toxicities and Interferences of

Hemoglobin-Based Oxygen-CarryingSolutions

Vaso-activityInterference with mononuclear

phagocyte systemAntigenicityOxidation to methemoglobinActivation of complement, kinin,

and coagulationThrombocytopeniaRed cell and platelet aggregationHistamine releaseFever, chills, gastrointestinal upset,

headache, backacheIron depositionBinding nitric oxideColorimetric interference, pulse

and fiberoptic oximetryInterference with liver function,

blood compatibility, andchemical testing

14


nant hemoglobin, 0.23 g/kg, there was no evi-dence of nephrotoxicity. Immunologic effectsof hemoglobin-based oxygen-carrying solu-tions remain somewhat of an unknown. In fact,the immunologic effects of blood transfusionhave been extensively explored only recently.Interferences occur with free hemoglobin so-lutions (Table 3). Use of hemoglobin-basedoxygen-carrying solutions interferes withfiberoptic oximetry because of the red color.9

Mixed venous oxygen saturation is overesti-mated at low levels of 60% to 70%—a danger-ous situation that may cause patients to beunderresuscitated. The interference is nonlin-ear, as it overestimates oxygen saturation athigh venous oxygen tension. Pulse oximetryinterference occurs because of methemoglo-bin.10 Hemoglobin-based oxygen-carrying so-lutions make it impossible to carry out someliver function tests such as alkaline phosphatemeasurement11 and coagulation tests such aspartial thromboplastin time.12 They can alsointerfere with cross-matching, but this can beovercome with dilution.

Methods to Prevent Complications ofOxygen-Carrying Solutions

For many reasons, including avoidance ofhuman disease transmission, sources otherthan outdated human blood have been usedto produce hemoglobin solutions. Transgenicpigs and mice have been bred to produce hu-man hemoglobin, and recombinant hemoglo-bins can be produced from bacteria and yeastby modifications that incorporate globin genes.For example, it is possible to express bothhuman a and b globin chains in Escherichiacoli; however, the yields are still very low.About 750 liters of cell culture would beneeded to produce 1 unit of blood. Endotoxincontamination may also occur.2

Sources of hemoglobin other than humaninclude bovine hemoglobin. In addition, any ofthese hemoglobins can be modified to optimizetheir characteristics such as retention time; oxy-gen affinity, reduction of dimer conversion intotetramers, and prevention of oxidation to meth-emoglobin. Bovine hemoglobin has a high P5Owithout modification and is therefore of inter-est since it is also in plentiful supply.13

Because of osmotic effects, most hemoglo-bin-based oxygen-carrying solutions are in con-centrations no greater than 7 to 8 g/dl.Perfluorocarbons have a linear oxygen dissocia-tion curve, and a relatively high oxygen contentof 50% or more is required for them to carryequivalent amounts of oxygen to hemoglobin.The second-generation perfluorocarbons(Perflubron) have more efficient oxygen carriage,even breathing 50% oxygen, whereas the first-generation (Fluosol) required 100% oxygenbreathing to achieve even one-fourth the oxy-gen carriage of blood.2

Hemoglobin may be modified by polymer-ization14 (Table 4). Polymerized hemoglobin isproduced by addition of reactive groups to thesurface of hemoglobin. These reactive groups

prolong intravascular retention time but alsomake the hemoglobin more rigid. The polymer-ization reaction is very difficult to control, sothere is some batch-to-batch variability. An al-ternative to polymerization is conjugation to alarger molecule, and this also prolongs reten-tion time. Some solutions can be polymerizedand conjugated. Intravascular retention timecan be prolonged from 7 hours in the unmodi-fied form to about 36 hours after modification.

The hemoglobin can be incorporated intoan artificial cell, and liposome encapsulationis currently under study. However, the lipo-somes cause substantial drops in plateletcounts, and during excretion, they block thereticuloendothelial system.4 A hemoglobin so-lution that has been studied in hemorrhagicshock is a pyridoxalated hemoglobinpolyoxyethylene conjugate made from stroma-free hemoglobin by conjugation withpolyoxethylene to increase its half-life from 7to 36 hours and by pyridoxalation to increaseP5O from 15 to 20 mmHg. Maltose is added toprevent oxidation to methemoglobin.15

The stimulus for all recent activity in de-velopment of hemoglobin-based oxygen-car-rying solutions is reduction of disease trans-mission, particularly of human immunodefi-ciency virus (HIV) and hepatitis virus. From theperspective of the manufacturers of oxygen-carrying solutions, there is much interest be-cause it is estimated to be a potential $12 bil-lion a year industry. Their use in hemorrhagicshock is important because huge quantities ofblood are currently used for this. In 1993, atthe Shock Trauma Center at the University ofMaryland, 1,300 patients were given 8,500units of blood, an average of 6.5 units per pa-tient, or about 50% to 60% of blood volumereplacement. The potential for replacing someof this blood use with an alternative is veryenticing for the manufactures of oxygen-carry-ing solutions and is also of interest to the RedCross, which goes to great efforts to maintainthis vital supply.

Vascular and Other Physiologic Effectsof Hemoglobin-BasedOxygenSubstitutes

How do we judge whether hemoglobin-based oxygen-carrying solutions are efficacious

in hemorrhagic shock? The objectives of suc-cessful resuscitation from hemorrhagic shockinclude 1) restoration of intravascular pres-sures, 2) increase in cardiac output, and 3)reversal of the increased oxygen extraction thatoccurs in hemorrhagic shock. When studiesusing red cell substitutes to achieve the firsttwo of these objectives are examined, difficul-ties in interpretation occur. The protocol andanimal model can influence the judgment ofefficacy. In one study in which a hemoglobinsolution was tested, the protocol specified thatfluid resuscitation should be given to restorecardiac filling pressures to baseline values.16 Ifa vasoconstrictor response occurred with in-fusion of the hemoglobin-based oxygen-carry-ing solution, it would appear very efficaciousat restoring vascular pressures. In addition, theevidence for a vasoconstrictor response wouldbe minimized. Furthermore, if an awake dehy-drated pig model had been used instead of adog, as other studies have shown, the animalmay have died as a result of the hemoglobin-based oxygen-carrying solution causing pro-found vasoconstriction and reduced cardiacoutput.17

If cardiac output and arterial blood pres-sure changes during resuscitation with oxygen-carrying solutions are examined, confoundingdata are also obtained. In two studies, cardiacoutput or blood pressure was less with hemo-globin solution infusion than with autologousblood transfusion.18,19 In these four studies,cardiac output and arterial pressure changeswere no different with hemoglobin solutionand blood resuscitation.20–23 Only one studyshowed that the hemoglobin solution sus-tained oxygen transport at higher levels thandid non-oxygen-carrying solution volume ex-panders such as albumin or lactated Ringer’ssolution.16 Transient cardiac output and bloodpressure increases were greater half an hourafter resuscitation began with hemoglobin so-lution resuscitation compared with autologousblood reinfusion in another study.15 So thereare no clear-cut data showing what effects he-moglobin solutions in general have on arte-rial pressure and cardiac output, nor is theremuch information showing they are conclu-sively more beneficial than non-oxygen-carry-ing volume expanders. In some studies, oxy-gen transport was significantly impaired com-pared with whole blood because of a fall inhematocrit,15 whereas in other studies oxygentransport is no different than with autologousblood resuscitation.20–23

How can oxygen-carrying solutions haveadded value over blood as a means of deliver-ing oxygen to tissues? There are several po-tential ways, some of which have been con-firmed by experiments in animals. Becauseoxygen-carrying solutions are acellular, theyare less viscous than blood and flow more eas-ily through narrow vessels and the microcir-culation. It is therefore possible that oxygen-carrying solutions may be useful in hemor-rhagic shock. There is experimental evidence

Table 4. Characteristics of SomeClinically Used Hemoglobin-Based

Oxygen-Carrying Solutions

1. Stroma-free hemoglobin (SFH):simple removal of the cell wallHuman and bovine SFH

2. Modifications includea. Cross-links

(alpha-alpha and beta-beta)b. Polymerizationc. Configurationd. Encapsulation


that hemoglobin-based oxygen-carrying solu-tions can enhance oxgyen diffusion from thevascular to the intracellular space.20 In addi-tion, when compared with whole autologousblood, a hemoglobin-based oxygen carrier pre-served exercise capacity in humans. Diffusionof carbon monoxide across the alveolar-capil-lary membrane (DLCO) and blood lactate lev-els were measured during exercise in hu-mans.25 There was a greater oxygen uptake andfor CO

2 production and normal lactate levels

were maintained in those given hemoglobin-based oxygen carriers in comparison with sub-jects given autologous blood. Infusion of 1gram of hemoglobin-based oxygen carriers in-creased DLCO as much as 3 grams of autolo-gous blood.

In addition to the short-term benefits ofenhanced diffusion, hemoglobin-based oxy-gen carriers may have longer lasting effects,as it has been shown that serum iron, fer-ritin, and erythropoietin increase in parallelwith plasma levels of hemoglobin. The ironinfusion adds the equivalent of one unit ofblood transfusion within 1 week of hemo-globin infusion.25

Newly discovered allosteric and electricproperties of hemoglobin appear to controlblood pressure and may facilitate tissue oxy-genation. S-Nitrous-hemoglobin (SNO-Hb) isfree of vasoactivity and may be a route for effi-cient delivery of nitric oxide to the mitochon-dria. Nitric oxide controls mitochondrial res-piration. Cell-free SNO-Hb may be a good he-moglobin-based oxygen-carrying solution.7.

Other properties besides oxygen transportaffect assessment of efficacy of red cell substi-tutes. Profound increases in pulmonary pres-sures can cause fatalities in some animals andprevented any benefit from being realized dueto hemoglobin solution infusion.17 Whenchanges in pulmonary artery pressure are com-pared after resuscitation from hemorrhagicshock with oxygen-carrying solutions,interspecies differences as well as protocolsand models become confounding variables.The rise in pulmonary artery pressure afterresuscitation in the swine model greatly ex-ceeds that seen in the dog.17 In some studies,fluid resuscitation is given with the objectiveof returning filling pressures to baseline val-ues, so that if a vasoconstrictor response oc-curred with infusion, the protocol used wouldprevent this difference from becoming appar-ent.20 Several investigators have also notedthrombocytopenia after infusion of red cellsubstitutes, and clearly it is critical that red cellsubstitutes for use in the management of hem-orrhagic shock should not interfere with resi-dent blood cells or the coagulation system, asthese toxicities would preclude their use in themanagement of patients with trauma or thoseundergoing surgery.

Hemostatic EffectsThe effects of free hemoglobin solutions

on coagulation and blood cellular components

were examined with resuscitation from severehemorrhagic shock in dogs.24 The solutionsused were 8% pyridoxalated hemoglobinpolyoxyethylene conjugate and 8% maltose,known as PHP88, a 4% solution of the samesolution made by diluting PHP88 with equalvolume of Plasmalyte A (PHP44), and stroma-free hemoglobin (SFH), a simple non-conju-gated hemoglobin solution. Both hemoglobinsolutions were highly purified and endotoxinfree. Use of these three hemoglobin solutionswas compared with re-infusion of autologousblood. The volume of blood removed to pro-duce 2 hours of shock was 63% of the estimatedblood volume. Resuscitation began with flu-ids infused at 20 ml/min by infusion pump; infour dogs, no resuscitation was given. Samplesfor coagulation and hematology profiles and ablood smear were taken one-half hour afterresuscitation began, when all the hemoglobin-based oxygen-carrying solutions were infusedor, in the case of non-resuscitated dogs, noadditional fluids were given. Measurementswere repeated at 2, 4, and 6 hours after resus-citation, and then daily for 7 days after awak-ening from anesthesia.

All dogs not resuscitated died within 2hours. All autologous blood and PHP44 dogssurvived 8 days, while mortality among PHP88dogs was 63% and among SFH dogs, 14%.Clinical coagulopathy occurred in all dogsgiven PHP88 and in four of the six dogs givenSFH, and there was evidence of hematoma for-mation around cannulation sites in all six dogsgiven PHP44 and five of the six dogs given au-tologous blood when autopsy was performed.Clinical coagulopathy with spontaneous devel-opment of oozing from percutaneously placedcannulae, spontaneous development of he-matomas in the femoral areas where catheterswere placed, and in some dogs receiving bothPHP and SFH petechiae were visible subcuta-neously all over the body, and submucosallyin the mouth. In the dogs that died, exsan-guination was the major cause of mortalitysecondary to thrombocytopenia. Death oc-curred between 7 and 254 hours after infusionof the hemoglobin solution.

There was a fall in hematocrit (Hct) in allanimals resuscitated with these cellular fluids.However, the fall in animals given PHP88 wassignificantly greater than in those receiving theother solutions, with an average Hct of 3% af-ter resuscitation.15 Since 63% of the estimatedblood volume was removed and because he-matocrit was, on average, about 40% beforeresuscitation, it was expected on the basis ofhemodilution alone, that Hct would be about25%. The finding that Hct was between 9%and 11% with PHP44 and SFH suggests thatthese hemoglobin solutions also had someadverse effect on red cells. The possibility ofhemolysis occurring was explored by hemo-globin electrophoresis of the plasma samples—the PHP and SFH were both derived from hu-man hemoglobin. If hemolysis had occurred,canine hemoglobin would be found in the

plasma. None was identified by hemoglobinelectrophoresis (which can clearly distinguishthe two types of hemoglobin). In addition,measurements of plasma hemoglobin gavevalues consistent with the quantities of hemo-globin-based oxygen-carrying solutions in-fused. Red cell counts show the same pictureas Hct. These data strongly suggest that cellswere being removed from the circulation. Itwas postulated that they may be sequesteredin circulatory beds as a result of endothelial orother interactions.

Why thrombocytopenia occurred andwhy there was a reduction in all other cellu-lar components remains the cause of muchspeculation and investigation. Many factorsare known to give rise to platelet adhesive-ness and rouleaux formation, including re-lease of thromboxane, and endothelial reac-tions, including binding of nitric oxide by freehemoglobin; the presence of free heme alsoinduces platelet aggregation. Hemodilutionis another important factor causing thromb-ocytopenia, as this was a severe hemorrhagicshock model in which 63% of the circulatingblood volume, and therefore cellular compo-nents of the blood, were removed from thecirculation. In addition, the high colloidoncotic pressure of these hemoglobin solu-tions may have further accentuated the circu-lating volume increase and dilution of cellu-lar components.15 It was speculated that, as aresult of platelet aggregation and rouleauxformation, platelets and red cells are trappedin the microcirculation and this sequestrationprevents their subsequent employment incoagulation and oxygen transport. A factorthat may additionally or singularly be thecause of the problem is the polyoxyethylenemoiety attached to hemoglobin in PHP. It isused to increase molecular size and prolongvascular retention time. However, it may alsocause electrostatic charges that increase thelikelihood of platelet aggregation and red cellrouleaux formations. Mediators released dur-ing hemorrhagic shock are probably an im-portant determinant of the cell aggregationseen with infusion of PHP. No coagulopathyor mortality was found in dogs undergoingexchange transfusions with PHP of 80% ofblood volume or in nonvolemic dogs given20 ml/kg of PHP. Changes that occur duringhemorrhagic shock exacerbate the effects oflarge doses of PHP.

The whole issue is very complex. An ob-vious possible mechanism of platelet aggre-gation, namely, binding of nitric oxide by freehemoglobin, has not been excluded. Freeheme can cause platelet aggregation, as canthromboxane release secondary to hemor-rhagic shock. The coagulopathy that occurredwith PHP may be a combination of some orall of these mechanisms.

ConclusionThe studies discussed illustrate some very

important facts about the data that are avail-


able on oxygen-carrying solutions. First, thereis virtually no published data on use of any ofthese products in humans for resuscitationfrom hemorrhage shock. Second, much ofthe evidence for the oxygen-carrying solu-tions currently under study in humans un-dergoing Phase 2 and Phase 3 trials to ob-tain FDA approval is proprietary. As a result,the data in the literature may not be scien-tifically valid, as adverse effects may be mini-mized. Third, there are interspecies varia-tions, so that toxicities or benefit seen witha product in animal studies may not trans-late into reality in human studies. Fourth,endothelial interactions are still not com-pletely understood and could result in ad-verse reactions to oxygen-carrying solutionsthat preclude their use in shock. Fifth, me-diators released during reperfusion or con-ditions that develop in hemorrhagic shockmay accentuate the toxicities of hemoglobin-based oxygen-carrying solutions, since ad-ministration of similar quantities of one he-moglobin-based oxygen-carrying solution(PHP) to animals not in shock does not re-sult in coagulopathy or mortality. Sixth, mi-nor changes among several potential modi-fications can significantly alter the toxicitiesof oxygen-carrying solutions.

What then is the future of hemoglobin andperfluorocarbon-based oxygen-carrying solu-tion? From news through the proprietary grape-vine, it appears that an equivalent of a two-unittransfusion of oxygen-carrying solution is welltolerated by the majority of individuals whengiven in elective surgical circumstances. The sideeffects are relatively minor, including gas-trointestinal upset, musculoskeletal aches, andheadache. One worrisome side effect that isrumored has been the development of pancre-atitis or signs of pancreatic changes seen in avery few individuals receiving some hemoglo-bin-based oxygen-carrying solutions. Anotherworrisome issue was the indefinite postpone-ment of a Phase 3 trial of a hemoglobin-basedoxygen solution in patients with trauma andhemorrhagic shock, presumably because ofincreased mortality in the study group (WallStreet Journal, February 6, 1998). The datathat led to the action have not been madepublic to date.

Other potential future uses include man-agement of ischemic disease and angioplasty.The acellular oxygen-carrying fluids have veryfavorable rheologic properties and enhancemitochondrial oxygenation. In some tumors,radiosensitivity is increased by means of in-creased oxygen levels. A further potential useof oxygen-carrying solutions is as an adjunctto radiation therapy for certain tumors. Insickle cell crisis, perfusion and oxygenationmay be improved with oxygen-carrying solu-tions and hematopoietic stimulation may bea result of infusion of a hemoglobin-basedoxygen-carrying solution. Because of the abil-ity to carry oxygen, these solutions may alsobe useful for organ preservation, extracorpo-

real organ perfusion, and cardioplegia. Poten-tial future uses also include transfusion alter-native in patients with red cell incompatibili-ties. For Jehovah’s Witnesses, perfluorocarbon,but not hemoglobin-based oxygen-carryingsolutions, are an acceptable alternative toblood transfusion.

References1. Amberson WR, Jennings JJ, Rhode CM.

Clinical experience with hemoglobin-sa-line solutions. J Appl Physiol 1949;1:469–89.

2. Winslow RM. Hemoglobin-Based Red CellSubstitutes. Baltimore, Johns HopkinsUniversity Press, 1992, pp 1–16.

3. Sellards AW, Minot GR. Injection of he-moglobin in man and its relation to blooddistribution, with especial reference tothe anemias. J Med Res 1916; 34:469–94.

4. Rabinovici R, Rudolph AS, Vernick J, etal. A new salutary resuscitative fluid: li-posome encapsulated hemoglobin/hy-pertonic saline solution. J Trauma 1993;35:121–7.

5. Clark LC, Gollman F. Survival of mammalsbreathing organic liquids equilibratedwith oxygen at atmospheric pressure. Sci-ence 1996; 152:1755–6.

6. Benesch RE, Benesch R, Renthal RD,Maeda N. Affinity labeling of thepolyphosphate binding site of hemoglo-bin. Biochemistry 1972; 11:3576–82.

7. Jai L et al. S. Nifroso Hb. A dynamic activ-ity of blood involved in vascular control.Nature 1996, 380:221–6.

8. Viele MK, Weiskopf RB, Fisher D. Recom-binant human hemoglobin does not af-fect renal function in humans. Anesthe-siology 1997; 86:848–58.

9. Kang LS, Ryder IG, Kahn R, et al. In vitrooxyhemoglobin saturation measure-ments in hemoglobin solutions usingfiberoptic pulmonary artery catheters. BrJ Anaesth 1995; 74:201–8.

10. Barker SJ, Tremper KK, Hyatt J. Effects ofmethemoglobinemia on pulse oximetryand mixed venous oximetry. Anesthesi-ology 1989; 70:112–7.

11. Bucci E, Fronticelli C, Razynska A,Urbaitis B. Overview of chemically ob-tained oxygen carriers from hemoglobin:pseudo cross-linked tetramers. BiomaterArtif Cells Artif Organs 1989; 17:637–9.

12. Eldridge J, Russell R, Christianson R, etal. Liver function and monorphology fol-lowing resuscitation from severe hemor-rhagic shock with hemoglobin solutionsor autologous blood. Crit Care Med 1996;24:663–71.

13. Alonsozana GL, Elfarth MD, MackenzieCF, et al. In vitro interference of the redcell substitute pyridoxalated hemoglobin-polyoxethylene with blood compatibility,coagulation, and clinical chemistry test-ing. J Cardiothoracic Vasc Anesth 1997;11:845–50.

14. Winslow RM. Hemoglobin-Based Red CellSubstitutes. Baltimore, Johns HopkinsUniversity Press, 1992, pp 72–95.

15. Sprung J, Mackenzie CF, Barnas GM, etal. Oxygen transport and cardiovasculareffects of resuscitation from severe hem-orrhagic shock using hemoglobin solu-tion. Crit Care Med 1995; 23:1540–53.

16. Harringer W, Hodakowski GT, Svizzero T,et al. Acute effects of massive transfusionof a bovine hemoglobin blood substitutein a canine model of hemorrhagic shock.Eur J Cardiothorac Surg 1992; 6:649–53.

17. Hess JR, MacDonald VW, Brinkley WW.Systemic and pulmonary hypertensionafter resuscitation with cell free hemo-globin. J Appl Physiol 1993; 74:1769–78.

18. Nho K, Glower D, Bredehoeft S, et al.PEG-bovine hemoglobin: safety in a ca-nine dehydrated hypovolemic-hemor-rhagic shock model. Biomat Art CellsImmobilization Biotechnol 1992;20:511–24.

19. Ning J, Anderson PJ, Biro GP. Resuscita-tion of bled dogs with pyridoxalated-po-lymerized hemoglobin solution. BiomatArt Cells Immobilization Biotechnol1992; 20:525–30.

20. Teicher RA et al. Oxygenation of tumorsby a hemoglobin solution. J Cancer ResClin Oncol 1993; 120:85–90.

21. Marks DH, Lynett JE, Letscher RM, et al.Pyriodoxalated polymerized stroma-freehemoglobin solution (SFHS-PP) as anoxygen-carrying fluid replacement forhemorrhagic shock in dogs. Milit Med1987;152:265–71.

22. Nees JE, Hauser CJ, Shippy C, et al. Com-parison of cardiorespiratory effects ofcrystalline hemoglobin, whole blood, al-bumin, and Ringer’s lactate in the resus-citation of hemorrhage shock in dogs.Surgery 1978; 83:639–47.

23. Greenberg AG, Schooley M, Ginsburg KA,Peskin GW. Pyridoxalated stroma-free he-moglobin in resuscitation of hemorrhagicshock. Surg Forum 1978; 29:44–6.

24. Mackenzie CR, Parr M, Christenson R, etal. The effect of free hemoglobin solu-tions on coagulation and hematology af-ter resuscitation from severe hemor-rhagic shock in dogs. Anesthesiology1993; 79(suppl):A269.

25. Hughes GS Jr, Yancey EP, Albrecht R, etal. Hemoglobin-based oxygen carrier pre-serves submaximal exercise capacity inhumans. Clin Pharmacol Ther 1995;58:434–43.

26. Hughes GS Jr, Francome SF, Antal EJ, etal. Hematologic effects of a novel hemo-globin-based oxygen carrier in normalmale and female subjects. J Lab Clin Med1995; 126:444–51.


Armin Schubert, MD, MBAChairman, Department of General AnesthesiaCleveland Clinic FoundationCleveland, Ohio, USA

[Editors’ note: Dr. Schubert is a consultant toBiopure (Hemosol).]

DefinitionsTechnically, a blood substitute is a sub-

stance that can effectively replace most func-tions of human blood. However, oxygen-carry-ing modified hemoglobin solutions andperfluorocarbons have been referred to as“blood substitutes.” Since these recently devel-oped solutions can only carry out selected func-tions of blood, they are more accurately referredto as “oxygen-carrying volume expanders.”

Hemoglobin-based oxygen carriers(HBOCs) are modified hemoglobin solutionsor hemoglobin packaged into liposomes,which are able to deliver oxygen to tissues. Ahemoglobin therapeutic is a hemoglobin so-lution optimized through chemical modifica-tion to bring about certain pharmacologic andtherapeutic effects. Hemoglobin therapeuticsmay possess a combination of therapeuticallyactive properties such as oxygen-carrying ca-pacity, favorable rheologic properties, andpressor action.

Need for Blood SubstitutesAlthough blood transfusions represent a

life-saving measure for many medical and sur-gical patients, there are still problems with ho-mologous blood transfusions in the United

Hemoglobin Therapeutics, Blood Substitutes, and High Volume Blood Loss

States. Oxygen-carrying volume expanders maybe particularly helpful in situations whereblood is not available (remote areas; difficultcross match; rare blood type, etc.). Further-more, a national blood shortage is predictedwith the aging of America, since the over-65age group has a high demand for blood. Thisage group represents 12.5% of the populationbut receives 50% of all blood transfusions. Therisk of infection from blood has decreased dra-matically, but potentially could be eliminatedwith blood substitutes (although it is recog-nized that prions and other agents appear toresist sterilization). Allogeneic blood also isassociated with a higher surgical infection rate,presumably related to the immunosuppressiveeffects of white blood cells contained in non-leuko-depleted blood.

Desirable “blood substitutes” have a longshelf life, a long circulation half-life, good oxy-gen carrying capacity and tissue oxygen deliv-ery, few side effects, and reasonable cost. Fur-thermore, their use should not interfere withdiagnostic tests or the clinical diagnosis of se-rious disease processes.

Hemoglobin-Based Oxygen CarriersStructure and Design

Free, unmodified human tetrameric he-moglobin rapidly dissociates into dimers andmonomers when removed from its normalenvironment inside the erythrocyte. Dissocia-tion into hemoglobin fragments leads to renaltoxicity and greatly increased oxygen affinity,precluding effective tissue oxygen delivery.

Manufacturers of HBOCs therefore have

undertaken a variety of strategies to modify thenative hemoglobin molecule in order to stabi-lize it, extend intravascular residence time, andreturn its oxygen-unloading properties into therange of erythrocyte-based hemoglobin. Onesuch method is intramolecular cross-linkingbetween alpha and beta chains. Other meth-ods involve polymerization, pyridoxylation, orconjugation to larger molecules, includingpolyethylene glycol (PEG). Encapsulation ofhemoglobin into a liposome or polymer struc-ture has also been pursued. There is a dilemmain the trade-off between desirable properties:Larger hemoglobins and liposomes may havelonger half-lives and are less active in scaveng-ing nitric oxide (NO) from the endothelium(which limits their hypertensive properties).Unfortunately, they also undergo acceleratedauto-oxidation, hemoglobin peroxidation, andheme loss.1 On the other hand, smaller spe-cies are less antigenic but can be filtered bythe kidneys, are more oncotically active, andhave shorter vascular residence times.

Such “designer” modifications stabilize themolecule’s tetrameric structure and affect mo-lecular size, renal filtration, P50 (defined as theoxygen tension at which hemoglobin oxygensaturation is 50%), affinity to NO binding, cir-culation half life, and more. The raw materialfor hemoglobin solutions can be human redblood cells, bovine red blood cells, or recombi-nant Escherichia coli bacteria. To date, no he-moglobin solution has been approved for hu-man use, although several are being investigatedfor safety and efficacy (Table 1).

Table 1. Hemoglobin Solutions Undergoing Clinical Testing

HBOC Raw Material Structure for Size (kD) T1/2 (hr) Oncotic Viscosity P50*Stabilization Pressure vs. blood (mmHg)

(mmHg)

Hemassist (DCLHb™; Human RBC Alpha-alpha cross-linked 64 4-16‡ 42 50% 32Baxter)†

Optro (Baxter-Somatogen) Recombinant Alpha-alpha cross-linked; 64 12-24 <20 50% 33E coli Hb Presbyterian mutation

Hemopure (HBOC-201; Bovine RBC Glutaraldehyde- >150 8-17§ 17 30% 34Biopure) polymerized

Hemosol (Fresenius) Human RBC o-Raffinose cross-linked >150 10-11 24 25% 34polymerized

Polyheme (Northfield) Human RBC Polymerized Hb; >150 24 20 30%-40% 28-30pyridoxylated 2,3-DPG site

*Normal human P50 = 28 mmHg†Baxter has discontinued the DCLHb program in favor of developing a second-generation hemoglobin.‡Varies directly with dose (0.1–1.0 g/kg)§Dose = 0.2–0.6 g/kg

15


PropertiesAlthough there are product-specific varia-

tions, the P-50s of HBOC solutions are gener-ally similar to those of fresh blood but higherthan those of stored blood. Circulation half-lives are measured in hours (4–24 hours, of-ten dose dependent) rather than days, aswould be the case for red blood cells.

All currently investigated hemoglobinselevate systemic and pulmonary vascular re-sistance, resulting in a mild reduction in car-diac index. For example, diaspirin cross-linkedhemoglobin (DCLHb™), an alpha-alpha cross-linked tetramer, produces a predictable, rapid,and sustained rise in mean arterial pressure(MAP) and in systemic and pulmonary vascu-lar resistance.2 At the microcirculatory level,functional capillary density is reduced.3 Thepressor response is dose dependent and phar-macologically reversible and exhibits a “ceil-ing effect.”4,5 In human volunteers, 100 mg/kgDCLHb™ raised median systolic BP maximallyby no more than 10 mmHg and diastolic BP byno more than about 15 mmHg.6 Biopure’sHBOC-201 raised MAP by about 10 mmHgwhen a dose of 0.6 g/kg was administered tohealthy volunteers,7 but it had no significanteffect on blood pressure when given to surgi-cal patients.8 In the author’s clinical investiga-tive experience with 1g/kg DCLHb™ adminis-tered to patients undergoing major orthope-dic and urologic surgery, MAP was elevated byan average of about 20 mmHg, the hyperten-sive effect persisting for 24 to 30 hours afteradministration.9 Although HBOC-associatedhypertension has not been associated withadverse cardiac events, selected patients arelikely to require treatment of hemoglobin-in-duced systemic and pulmonary hypertension.

The mechanisms thought to account forthis pressor effect are the scavenging of NOfrom vascular endothelium, facilitation ofendothelin production and, possibly, a sym-pathomimetic effect. The smaller the hemoglo-bin molecule the more effectively it interactswith the endothelium, penetrating it and scav-enging endothelial NO to form met-hemoglo-bin and NO-hemoglobin.10

In the operative setting, several factorsmay blunt HBOC-associated hypertensive ten-dencies. The hypotensive action of surgicalhemorrhage,11 as well as volume depletion,12,13

may diminish hypertension. Furthermore, ha-lothane and propofol, but not isoflurane, havebeen shown to decrease the hypertensive ac-tion of DCLHb™ on pulmonary vein rings.14

Hemoglobin solutions have colloidalproperties (Table 1), are highly purified, gen-erally do not affect coagulation, and are onlyweakly antigenic. Modified molecular hemo-globin undergoes oxidation to methemoglo-bin and leaves the circulation primarilythrough the reticuloendothelial system. Pre-clinical and clinical studies indicate that modi-fied hemoglobins can mildly increase the con-centrations of plasma CPK (but not MB frac-tion), hepatic enzymes, reticulocyte count, bi-

lirubin, and amylase.15–17 In a study of patientsundergoing high-blood-loss (approximatelyhalf of an adult’s blood volume) surgical pro-cedures, 1 g/kg DCLHb™ was associated withtransient elevations in serum LDH, AST, totalbilirubin, CK, BUN, and amylase; a high inci-dence of yellow skin discoloration; and asymp-tomatic hemoglobinuria.9

Gastrointestinal side effects include flatu-lence, nausea, vomiting,6,7 and possibly pan-creatitis. However, pancreatitis occurs fre-quently after major abdominal surgery18-20 evenin the absence of HBOC administration. Thegastrointestinal side effects of DCLHb™ maybe related to its ability to interfere with NOproduction and signaling,21 thus possibly af-fecting gastrointestinal and biliary motility.Judging from preclinical studies22,23 of intesti-nal and portal system blood flow after admin-istration of DCLHb™, gastrointestinal side ef-fects are unlikely the result of tissue ischemia.

ToxicityToxicity of hemoglobin solutions has his-

torically been related to impurities such as RBCmembrane residues, endotoxin, free dimers,and monomers. With vastly improved purifi-cation procedures, concern over toxicity fromimpurities is waning. In particular, the issueof renal toxicity appears to have been over-come. In rats, 0.4 g/kg DCLHb™ did not affectrenal blood flow.23 Creatinine clearance wasneither decreased by 0.1 g/kg DCLHb™4 norby 0.32 g recombinant hemoglobin24 in humanvolunteers. It was similarly unaffected by upto 0.7 g/kg in critically ill patients with sepsissyndrome25 by 750 ml DCLHb™ in cardiac pa-tients,26 and by 1.0 g/kg DCLHb™ in patientsundergoing high-blood-loss surgery, despitethe occurrence of hemoglobinuria at the higherdoses.9 Neither was renal toxicity observedwith polymerized hemoglobin.27

Free hemoglobin, then directly applied tocentral nervous system tissue, is neurotoxic.It stimulates leukocyte migration and vascularadherence. Hemoglobin also activates plate-lets, promoting aggregation.28 Circulating fer-rous hemoglobin, even when highly purified,undergoes a number of reactions that maycontribute to toxicity.29 Ferrous hemoglobinbinds NO about 3,000 times more tightly thancarbon monoxide and therefore effectively re-moves any NO in its vicinity, accounting forthe vasoactive properties. Free hemoglobin isconverted to methemoglobin at a rate as fastas 4% per hour; this reaction can lead to thegeneration of free radicals. Hemoglobin alsohas a number of “pseudo-enzymatic” proper-ties, which could lead to oxygenation, lipidperoxidation, and cytotoxicity. Further possi-bilities for toxicity arise from the degradationproducts of hemoglobin’s heme moiety suchas hemin. Red blood cells contain antioxidantenzymes such as catalase and superoxidedismutase, which may help limit ischemiareperfusion injury. It has been speculated thatthe administration of pure hemoglobin (i.e.,

without antioxidants) may lead to a potentiallyhigher risk of reperfusion injury.30

Although many early trials indicate thatsome HBOCs have not been associated withsevere toxicity, more study in a wide variety ofclinical situations is required before their sideeffects are fully known. Investigation into theeffects of HBOCs on the gastrointestinal sys-tem, pulmonary vasculature, and organ func-tion during hemorrhagic and other stress isparticularly needed. Furthermore, the charac-teristics of HBOC-assisted oxygen delivery andtissue oxygen availability during supply-depen-dent conditions need additional investigation.

Perfluorocarbons (PFCs)Perfluorocarbons are inert aromatic or

aliphatic chemicals that can dissolve oxygenand carry it in solution throughout the body.They typically carry 4 to 50 vol% at a PaO

2 of

160 mmHg; their ability to carry oxygen is di-rectly proportional to their concentration inblood and, importantly, to the partial pressureof oxygen. The first fluorocarbon to be ap-proved for clinical use (during percutaneoustransluminal coronary angioplasty) was fluosolDA-20, which contains 20% emulsified fluoro-carbon. When used as an oxygen-carrying vol-ume expander, fluosol DA was associated witha number of limitations, including low oxygen-carrying capacity, short shelf life, temperatureinstability, and serious side effects. Second-generation perfluorocarbons, such asperfluoro-octylbromide (PFOB; Alliance Phar-maceuticals), are being investigated and showpromise because of a much higher oxygen-car-rying capacity, a 2- to 4-year refrigerated shelflife, low viscosity, and less interference withnormal pulmonary surfactant mechanisms.31

Since PFCs are not metabolized, but ex-creted unchanged via the lungs, their poten-tial for cytotoxicity is thought to be limited.There is no antigenicity. However, since PFCsare taken up avidly by the reticuloendothelialsystem, they increase liver enzymes and resultin hepatosplenomegaly. Because of the exten-sive uptake in the reticuloendothelial systemand impairment of neutrophil function, theymay interfere with host defense mechanisms.Monocyte and macrophage activation may leadto release of prostaglandins, endoperoxides,and cytokines, which probably accounts for thesymptoms of flushing, backache, fever, chills,headaches, and nausea observed in clinical tri-als. Platelet count decreases by as much as 40%due to increased platelet clearance from PFC-induced modification of platelet surfaces.32

PFCs also may prolong the effects of certaindrugs, including barbiturates.

Potential Clinical Uses and Effective-ness of HBOCs Major Surgical Bleedingand Hemorrhagic Shock

Fluid therapy for the acutely bleeding pa-tient can be accomplished initially with eithercrystalloid or colloid solutions. Blood transfu-sion is begun when, despite volume resuscita-


tion with non-oxygen carrying solutions, thereis evidence of tissue ischemia and resultantorgan dysfunction. Accumulation of base defi-cit and serum lactate and low central venousoxygen concentration are all indices of tissueischemia, which should be taken into consid-eration in the transfusion decision.33 Alterna-tively, blood is transfused when organ ischemiacan be anticipated, given the extent and rapid-ity of ongoing bleeding.

Hemoglobin solutions are as effective aswhole blood in restoring MAP in animals34-36 andhumans.12 In contrast to typical catecholamineeffects, the pressor response of DCLHb™ is as-sociated with an increase in perfusion (as indi-cated by organ flow measurements) in both top-load and hemorrhagic, hypovolemic, animalmodels.36–40 DCLHb™, compared with non-oxy-gen-containing crystalloid or colloid solutions,resulted in substantially better survival fromexperimental hemorrhagic shock.37,38 This salu-tary effect may be related to DCLHb™’s effecton tissue perfusion and peripheral oxygenation.For example, tissue oxygenation, measured di-rectly by a fluorescence-quenching optode, wasrestored more effectively in a rat hemorrhagicshock model treated with DCLHb™ comparedwith lactated Ringer’s solution and albumin.41

Despite increased total peripheral vascular re-sistance, rat coronary blood flow23 was aug-mented after DCLHb™ and human cerebralblood flow was unchanged after infusion ofpolymerized hemoglobin.42

Despite their vasoconstrictive properties,HBOCs may counteract tissue hypoperfusionwith added blood oxygen-carrying capacity andbetter rheologic properties. In spontaneouslyhypertensive rats subjected to middle cerebralartery occlusion, hemodilution with DCLHb™(to hematocrits of 30, 16, or 9%) resulted in asignificant dose-dependent reduction in theextent of brain injury and cerebral edema.43

The most effective reductions in ischemic in-jury occurred in those animals in which theinherent hypertensive response to DCLHb™was not inhibited.

The effect of HBOCs on cardiac index ismore controversial, with some studies report-ing a slight decrease,44 others no change.25 Cal-culated oxygen delivery generally follows car-diac output, thus accounting for the slight de-creases reported. However, increased tissue-dffusing capacity has been shown.7 Further-more, the equivalent or enhanced oxygen-un-loading capacity of HBOCs compared withblood should allow favorable tissue oxygendelivery, or at least counteract vasoconstrictiveeffects of free hemoglobins. Therefore, their usein trauma patients and in those with substan-tial surgical bleeding would seem reasonable.Table 2 suggests potential uses of HBOCs fortherapy of patients suffering large blood losses.

However, because of their short half-lives,current hemoglobin solutions are likely to beused essentially as a “bridge to transfusion.”For example, the half-life of DCLHb™ admin-istered to patients undergoing high-blood-loss

surgery was approximately 10 hours.45 Admin-istration of DCLHb™ after bypass spared nearly20% of cardiac patients from allogeneic trans-fusion.46 Nevertheless, there is also concernthat the administration of modified hemoglo-bins merely delays blood transfusion ratherthan truly substituting for it.

Preoperative acute normovolemic hemodi-lution (ANH) is likely to become more attrac-tive with the use of modified hemoglobins as adiluent. The short half-life of HBOCs does notpresent a significant liability for this clinical ap-plication. Patients with low preoperative hema-tocrit might receive an infusion of HBOC to “tide

them over” a limited period of intraoperativeor postoperative bleeding, after which autolo-gous or allogeneic blood would be administeredif still needed. Furthermore, volume replace-ment with HBOCs (compared with crystalloidor colloid) during ANH for autologous collec-tion would likely result in a greater yield ofpheresed blood components.

Caveats Regarding the Use of HBOCsfor Major Blood Loss (Table 3)

The safety of large-scale and rapid transfu-sion of HBOCs in human traumatic injury re-mains to be demonstrated. While the author’ssmall series of patients undergoing high-blood-loss elective surgery tolerated up to 1g/kgDCLHb™ relatively well,9 a phase III trial ofDCLHb™ for resuscitation of traumatically in-jured patients was halted among concerns aboutincreased mortality in the study group.

Because HBOCs are associated with sys-temic hypertension, concern has been raisedover a potential for increased blood loss inhemorrhage. This concern could not be cor-roborated in preclinical40 or clinical studies9

conducted in a setting of hemorrhage, but theissue has yet to be clarified in the setting ofpenetrating trauma.

The clinical use of HBOCs with relativelyshort half-lives must take into account theirtendency for transvascular migration and theirrapid clearance through the reticuloendothe-lial system. Although the initial effect of trans-fusion may be an immediate increase in vascu-lar volume (enhanced by some HBOCs’ col-loidal properties) and blood pressure (medi-ated by the HBOCs’ NO-scavenging effect),rapid dissipation of the HBOC requires care-

Table 2.Use of Hemoglobin-Based Oxygen

Carriers in Patients with High Blood Loss

Emergency administration• Trauma, especially penetrating• Unexpected surgical bleeding• Unexpected bleeding from

disease (e.g., gastrointestinal tract)• Difficult cross-match

Elective administration• Acute normovolemic hemodilution*• Acute hypervolemic hemodilution*• Replacing blood transfusion during

expected active surgical bleeding• Replacing blood transfusion

postoperatively

*Especially in patients presenting with low initial hematocrit

Table 3. Caveats and Potential Remedies in the Clinical Use of HBOCs

NO, nitric oxide; Hb, hemoglobin

Caveats Remedies

Hypertensive tendency; Co-administration of nitroglycerin,cardiac afterload stress other vasodilator

Pulmonary hypertension, Co-administration of pulmonaryright ventricular dysfunction vasodilator

Short intravascular residence time; More frequent assessment andrecurrence of hypovolemia adjustment of intravascular volume

Interference with diagnostic Avoidance of photospectrometric methods;blood tests removal of free hemoglobin from

specimens; other correction algorithms

Hemoglobinurla interfering with Special pre-arranged testing protocoldiagnosis of transfusion reaction

Immune depression as larger Hb species Unknown overwhelm the reticuloendothelial system

Possible NO-related gastrointestinal or Co-supply NO donor or precursor;other organ injury redesign molecule


ful and frequent monitoring of circulatory ad-equacy, since hypovolemia may re-manifestrather quickly.28

At least within the first 24 to 36 hours ofadministration, free hemoglobins can interferewith the photospectrometric methods used ina variety of clinical laboratory tests. Interfer-ence with laboratory testing constitutes animportant limitation for the potential clinicaluse of artificial hemoglobin species.

Other Uses: Hemoglobin Therapeutics?Since NO plays a part in the pathogenesis

of septic shock, modified hemoglobins maybecome useful in the treatment or preventionof severe septic shock. Artificial oxygen carri-ers also may be used for oxygen delivery toischemic tissues (as in stroke or intestinal is-chemia) and tumor cells to improve their sus-ceptibility to radiation and chemotherapy. Be-cause iron is one of the breakdown productsof hemoglobin metabolism, hemoglobintherapy may stimulate erythropoiesis undercertain circumstances. These and other poten-tially salutary effects of HBOCs, which tran-scend basic oxygen-carrying properties, haveled to an emerging interest in the area of “he-moglobin therapeutics.”

References1. Sanders KE, Ackers G, Sligar S. Engineer-

ing and design of blood substitutes. CurrOp Struc Biol 1996; 6:534–40.

2. Hess JR, Macdonald VW, Brinkley WW. Sys-temic and pulmonary hypertension afterresuscitation with cell-free hemoglobin. JAppl Physiol 1993; 74:1769–78.

3. Tsai AG, Kerger H, Intaglietta M. Microcir-culatory consequences of blood substitu-tion with alpha-alpha-hemoglobin. InWinslow RM, Vandegraff KD, Intaglietta M,eds. Current Issues in Blood Substitute Re-search—1995. Boston, Birkhauser, 1995,pp 155–74.

4. Malcolm DS, Hamilton IN, Schultz SC, etal. Characterization of hemodynamic re-sponse to intravenous diaspirincrosslinked hemoglobin solution in rats.Artif Cells Immobil Biotech 1994; 22:91–107.

5. Hamilton I, Schultz SC, Cole F, Burhop K,Malcolm D. Characterization of diaspirincross-linked hemoglobin’s pressor re-sponse. Crit Care Med 1992; 20:S106.

6. Przybelski RJ, Daily EK, Kisicki JC, Mattia-Goldberg C, Bounds MJ, Colburn WA.Phase I study of the safety and pharmaco-logic effects of diaspirin crosslinked he-moglobin solution. Crit Care Med 1996;24:1993–2000.

7. Hughes GS, Antal EJ, Locker PK, FrancomSF, Adams WJ, Jacobs EE Jr. Physiology andpharmacokinetics of a novel hemoglobin-based oxygen carrier in humans. Crit CareMed 1996; 24:756–64.

8. Monk TG, Goodnough LT, Peruzzi WT,Byers PM, Jacobs Jr EE, Silverman MH. A

dose-escalation study to evaluate the ki-netics and hemodynamic effects of hemo-globin-based oxygen carrier-201. Anesthe-siology 1997; 87:A214.

9. Schubert A, Bedocs N, O’Hara JF Jr, TetzlaffJE, Marks KE, Novick AC. Effect ofperioperative administration of diaspirincross-linked hemoglobin on indices of or-gan function. Anesthesiology 1997;87:A220.

10. Scott MG, Kucik DF, Goodnough LT, MonkTG. Blood substitutes: evolution and fu-ture applications. Clin Chem 1997;43:1724–31.

11. Malcolm D, Kissinger D, Garrioch M.Diaspirin crosslinked hemoglobin solu-tion as a resuscitative fluid following se-vere hemorrhage in the rat. Biomat, ArtifCells, Immmob Biotech 1992; 20:495–7.

12. Swan SK, Halstenson CE, Collins AJ,Colburn WA, Blue J, Przybelski RJ. Phar-macologic profile of diaspirin crosslinkedhemoglobin in hemodialysis patients. AmJ Kidney Dis 1995; 26:918–23.

13. Malcolm D, Kissinger D, Garrioch M.Diaspirin crosslinked hemoglobin solu-tion as a resuscitative fluid following se-vere hemorrhage in the rat. Biomat, ArtifCells, Immmob Biotech 1992; 20:495–7.

14. Jing M, Ledvina MA, Bina S, Hart JL,Muldoon SM. Effects of halogenated andnon-halogenated anesthetics on diaspirincrosslinked hemoglobin induced contrac-tions of porcine pulmonary veins. ArtifCells Blood Subs Immobil Biotech 1995;23:487–94.

15. Leone BJ, Chuey C, Gleason D, Steele SM,et al. Can recombinant human hemoglo-bin make ANH more effective? An initialfeasibility and safety study. Artif CellsBlood Subs Immobil Biotech 1996;24:A379.

16. Lessen R, Williams M, Seltzer J, Lessin J,et al. A safety study of recombinant hu-man hemoglobin for intraoperative trans-fusion study. Artif Cells Blood SubsImmobil Biotech 1996; 24:A380.

17. Gould SA, Moore EE, Moore FA, HaenelJB, et al. The clinical utility of human po-lymerized hemoglobin as a blood substi-tute after acute trauma and urgent surgery.J Trauma 1995; 39:157.

18. Jurkovich GJ, Carrico CJ. Pancreatic trauma.Surg Clin North Am 1990; 70:575–93.

19. Miyagawa S, Makuuchi M, Kawasaki S,Kakazu T. Changes serum amylase levelfollowing hepatic resection in chronic liverdisease. Surgery 1994; 129:634–8.

20. Akagi Y, Yamashita Y, Kurohiji T, et al. Inves-tigation of hyperamylasemia after hepatec-tomy. J Jpn Soc Clin Surg 1991; 52:314–8.

21. Sharma AC, Sinh G, Gulati A. Role of NOmechanism in cardiovascular effects ofdiaspirin crosslinked hemoglobin in anes-thetized rats. Am J Physiol 1995; 269(4 Pt2):H1379–88.

22. Sharma AC, Rebello S, Gulati A. Regional

circulatory and systemic hemodynamiceffects of diaspirin crosslinked hemoglo-bin in the rat. Artif Cells Blood SubsImmobil Biotech 1994; 22:593–602.

23. Sharma AC, Gulati A. Effect of diaspirincrosslinked hemoglobin and norepineph-rine on systematic hemodynamics andregional circulation in rats. J Lab Clin Med1994; 123:299–308.

24. Viele MK, Weiskopf RB, Fisher D. Recom-binant human hemoglobin does not affectrenal function in humans: analysis ofsafety and pharmacokinetics. Anesthesiol-ogy 1997; 86:848–58.

25. Rhea G, Bodenham A, Mallick A, PrzybelskiR, Daily E. Initial evaluation of diaspirincrosslinked hemoglobin (DCLHb) as avasopressor in critically ill patients. CritCare Med 1997; 25:1480–8.

26. Baron JF, Berridge J, Brichant JF,Demeyere R, Lamy M, et al. The use ofdiaspirin crosslinked hemoglobin(DCLHb) as an alternative to blood trans-fusion in cardiac surgery patients follow-ing cardiopulmonary bypass: a pivotal ef-ficacy trial. Anesthesiology 1997; 87:A217.

27. Gonzalez P, Hackney AC, Jones S.Strayhorn D, Hoffman EB, Hughes G,Jacobs EE, Oringer EP. A phase I/II studyof polymerized bovine hemoglobin inadult patients with sickle cell disease notin crisis at the time of study. J Inv Med1997; 45:258–64.

28. Hess JR, Reiss RF. Resuscitation and thelimited utility of the present generationof blood substitutes. Transfusion Med Rev1996; 10:276–85.

29. Everse J, Hsia N. The toxicities of nativeand modified hemoglobins. Free RadicalBiology & Medicine 1997; 22:1075–99.

30. Chang TM. Recent and future develop-ments in modified hemoglobin and mi-croencapsulated hemoglobin as red bloodcell substitutes. Artif Cells Blood SubsImmobil Biotech 1997; 25:1–24.

31. Spiess BD. Perfluorocarbon emulsions:one approach to intravenous artificial res-piratory gas transport. Inter Anesth Clin1995; 33:103–13.

32. Smith DJ, Lane TA. Effect of a high con-centration perfluorocarbon emulsion onplatelet function. Biomater Artif CellsImmobil Biotech 1991; 19:383–5.

33. Baron BJ, Scalea TM. Acute blood loss.Emerg Med Clin North Am 1996; 14:35–55.

34. Barve A, Sen AP, Saxena PR, Gulati A. Doseresponse effect of diaspirin crosslinkedhemoglobin (DCLHb) on systemic hemo-dynamics and regional blood circulationin rats. Artif Cells Blood Substit ImmobilBiotechnol 1997; 25(1-2):75–84.

35. Chang TMS, Varma R. Effect of a singlereplacement of one of Ringer lactate, hy-pertonic saline/dextran, 7g% albumin,stroma-free hemoglobin, o-raffinosepolyhemoglobin or whole blood on thelong term survival of unanesthetized rats


with lethal shock after 67% acute bloodloss. Biomater Artif Cells Immobil Biotech1992; 20:503–10.

36. Przybelski RJ, Malcolm DS, Burris DG, etal. Cross-linked hemoglobin solution as aresuscitative fluid after hemorrhage in therat. J Lab Clin Med 1991; 117:143–7.

37. McKenzie JE, Scandling DM, Ahle NW.Diaspirin crosslinked hemoglobin(DCLHb™): Improvement in regionalblood flow following administration in hy-povolemic shock in the swine (abstract).5th International Symposium Blood Sub-stitutes, San Diego, 1993.

38. Schultz SC, Powell CC, Bernard E, MalcolmD. Diaspirin crosslinked hemoglobin(DCLHb™) attenuates bacterial transloca-tion in rats. Artif Cells Blood Subs Immobil

Biotech 1994; 22:A260.39. McKenzie JE, Scandling DM, Rohrer MJ,

Jacot JL. Diaspirin crosslinked hemoglobin(DCLHb™) during hypovolemic shock inswine. ISBS 1993 Program and Abstracts.

40. Schultz SC, Powell CC, Burris DG, et al. Theefficacy of diaspirin crosslinked hemoglobinsolution resuscitation in a model of uncon-trolled hemorrhage. J Trauma 1994; 37:408.

41. Powell C, Schultz SC, Burris DG, DruckerWR, Malcolm DS. Subcutaneous oxygentension: a useful adjunct in assessment ofperfusion status. Crit Care Med, in press.

42. Brauer P, Standl T, Wilhelm S, BurmeisterMA, Schulte am Esch J. Transcranial dopplersonography mean flow velocity during in-fusion of ultrapurified bovine hemoglobin.J Neurosurg Anesth 1998; 10:146–52.

43. Cole DJ, Schell RM, Drummond JC,Przybelski RJ, Marcantonio S. Focal cere-bral ischemia in rats: effect of hemodilu-tion with (-( crosslinked hemoglobin onbrain injury and edema. Can J Neurol Sci1993; 20:30–6.

44. Lamy M. Personal communication, 1997.45. O’Hara JF, Tetzlaff JE, Udayashankar SV,

Connors DF, Bedocs NM, Schubert A. Theeffect of diaspirin cross-linked hemoglobinon coagulation in surgical patients. Anes-thesiology 1997; 87:A230.

46. Baron JF, Berridge J, Brichant JF, et al. Theuse of diaspirin crosslinked hemoglobin(DCLHb) as an alternate to blood transfu-sion in cardiac surgery following cardiop-ulmonary bypass. Anesthesiology 1997;87:A217.

1. The most sensitive measure of acuteblood loss isa. blood pressure.b. urine output.c. heart rate.d. mixed venous oxygen saturation.

2. High-risk elderly patients require invasivehemodynamic monitoringa. only if they have a history of coronary

artery disease.b. as early as possible, following emer-

gency department admission.c. once they are evaluated for injuries.d. if they have evidence of hypotension

and tachycardia.

3. The relationship between serum lactateand base deficit remains constant for howlong following resuscitation?a. 12 hoursb. 24 hoursc. 36 hoursd. 48 hours

4. Which of the following is the most accu-rate predictor of survival following injury?a. serial blood pressure determinationb. adequacy of urine outputc. ability to clear lactate to normald. resolution of tachycardia

5. Interventional radiologic techniques canbe useful in which body area?a. Zone 3 of the neckb. Zone 2, the thoracic outletc. deep in the pelvisd. all of the above

6. Stages of traumatic shock include all of thefollowing excepta. subacute irreversible shock.b. compensated shock.c. decompensated shock.d. cardiogenic shock.e. acute irreversible shock.

CME QuestionsThis monograph can be used to earn 15 AMA category 1 credit hours.

The International Trauma Anesthesia and Critical Care Society (ITACCS) is accredited by the Accreditation Council for Continuing MedicalEducation (ACCME) for physicians. This CME activity was planned and produced in accordance with the ACCME Essentials. ITACCS designates thisCME activity for 15 credit hours in Category 1 of the Physicians Recognition Award of the American Medical Association.

Educational ObjectivesThis activity is designed to provide trauma

care professionals interested in the treatmentof critically ill trauma patients with a regularoverview and critical analysis of the most cur-rent, clinically useful information available, cov-ering strategies and advances in the diagnosisof traumatic injuries and the treatment oftrauma patients. Controversies, advantages, anddisadvantages of diagnosis and treatment plansare emphasized. There are no prerequisites forparticipation in this activity.

After reading this document, participantsshould have a working familiarity with the mostsignificant information and perspectives pre-sented and be able to apply what they havelearned promptly in clinical practice.

Accreditation StatementThis activity is planned and produced in

accordance with the Essential Areas and Poli-cies of the Accreditation Council for Continu-ing Medical Education (ACCME) through thesponsorship of the International Trauma Anes-

thesia and Critical Care Society (ITACCS).ITACCS is accredited by the ACCME to sponsorcontinuing medical education (CME) for phy-sicians and takes responsibility for the content,quality, and scientific integrity of this CME ac-tivity.

Credit Designation StatementITACCS designates this educational activ-

ity for a maximum of 15 hours in category 1credit toward the AMA Physicians RecognitionAward.

Faculty Disclosure StatementIt is the policy of ITACCS that faculty mem-

bers disclose real or apparent conflict of inter-est relating to the optics of this educationalactivity and also disclose discussions of unla-beled/unapproved uses of drugs or devices intheir presentations. Sincere effort was made tocontact the contributors to this publication. Anyresponses that could possibly suggest conflictof interest are published on the opening pagesof the individual articles. The authors’ com-

pleted disclosure forms are on file in the man-aging editor’s office.

INSTRUCTIONS• Print the answer form and the evaluation

form on page 51.• Complete both forms. On the answer form,

circle only one response next to each num-ber.

• Sign and date the certification statementon page 51.

• Write a check for $150 (or $75 accompa-nied by verification of current ITACCSmembership), payable to the InternationalTrauma Anesthesia and Critical Care Soci-ety.

• Mail the forms and your check (and mem-bership verification, if applicable) toITACCS, Department of CME Credit, POBox 4826, Baltimore, MD 21211.

• The completed text will be accepted forgrading if received by January 31, 2004.

• Please allow 4 to 6 weeks for processing.


7. Of the following organ systems, the onemost directly affected by decreased bloodflow in traumatic shock isa. cardiac.b. intestinal.c. pulmonary.d. central nervous.e. skeletal muscle.

8. Which of the following statements is correct?a. In compensated shock, the body is not

developing an oxygen debt.b. In subacute irreversible shock, normal

hemodynamics are never achieved.c. In neurogenic shock, ischemia is caused

by decreased oxygen-carrying capacity.d. Traumatic shock is the same as hem-

orrhagic shock.e. Decompensated shock is a stable clini-

cal state that can persist for many days.

9. Which of the following is not an inflamma-tory mediator produced by ischemic cells?a. Prostacyclinb. Tumor necrosis factorc. Complementd. Thromboxanee. Angiotensin II

10. Acute irreversible shock includes all of thefollowing clinical signs excepta. hyperthermia.b. coagulopathy.c. hypotension not responsive to fluids.d. hypotension not responsive to

inotropes.e. diffuse edema.

11. The first response of the body to obtainhemostasis isa. initiation of the coagulation cascade.b. platelet aggregation.c. initiation of fibrinolysis.d. platelet release reaction.e. vasoconstriction.

12. Concerning platelets and hemostasis:a. Exposure of platelets to

subendothelial collagen leads toadherence between platelets andthe blood vessel wall.

b. Platelet release reaction refers tothe liberation of plateletssequestered in the spleen.

c. The intrinsic coagulation pathwayis the predominant pathway in thecoagulation cascade.

d. The intrinsic and extrinsic pathwaysmerge with the activation of Factor IX.

e. Fibrinolysis occurs only after theclotting mechanism is completed.

13. The ideal topical hemostatic agent pos-sesses which of the following properties:a. Rapid time to hemostasisb. Easily applied and manipulatedc. Holds suturesd. Low infectious risk and minimal

tissue reactione. All of the above

14. Concerning topical hemostatic agents:a. Collagen sponges are unique in that

they are bactericidal.b. Denatured gelatin (Gelfoam®)

possesses clotting activity similarto collagen preparations.

c. Thrombin is effective only ifcombined with a carriersuch as Gelfoam®.

d. Fibrin glue has been shown to beeffective as either the primaryhemostatic agent or as an adjunct toconventional suture repair in patientswith hepatic or splenic trauma.

e. In vitro testing reveals that oxidizedregenerated cellulose (Surgicel®) ismore effective than collagenpreparations for inducing plateletaggregation and clotting.

15. Severely injured patientsa. have been shown to have elevated se-

rum fibrin degradation products (FDP).b. may exhibit thrombocytopenia.c. may progress to death if FDP and

platelet assays trend in an abnormalmanner.

d. benefit from prophylactic transfusion offresh frozen plasma and platelets evenin the absence of pathologic bleeding.

e. a, b, c

16. Risk factors for DVT in trauma patients in-cludea. spinal cord injury.b. prolonged bed rest.c. hypercoagulability.d. lower extremity fractures.e. all of the above.

17. The most common inborn metabolic er-ror that causes thrombophilia isa. activated protein C resistance.b. protein C deficiency.c. protein S deficiency.d. hypohomocysteinemia.e. serum porciline deficiency.

18. Physical examination is the most accuratemethod of diagnosing DVT.a. True b. False

19. The primary reason to provide prophylac-tic treatment to prevent DVT in traumapatients is toa. prevent leg swelling.b. enhance fracture healing.c. prevent fatal pulmonary embolism.d. increase billable services.e. decrease length of hospitalization.

20. Epidural analgesia in the patientreceiving LMWHa. is absolutely contraindicated.b. is associated with epidural abscesses.c. is no problem.d. may be performed at least 12 hours

after the last dose.e. is not associated with problems of

catheter removal.

21. Management priorities in the acutelybleeding trauma patient include all of thefollowing except:a. Measurement of BPb. Securing the airway and verifying

adequacy of ventilation andoxygenation

c. Insertion of a pulmonary artery

catheterd. Placement of the ECGe. Obtaining large-bore venous access

22. Appropriate intraoperative fluid manage-ment for a 70-kg multiple blunt traumapatient in class 4 hemorrhagic shock in-cludes which of the following:a. Hetastarch, 2.5-L bolusb. Lactated Ringer’s, 5 Lc. 7.5% saline, 1.5 Ld. Two units type-specific

uncrossmatched red blood cells and3 L normal saline

e. Four units of fresh frozen plasma

23. Resuscitation endpoints after major traumainclude which of the following:a. Resolution of lactic acidosis and

base deficitb. Mixed venous oxygen saturation 45%c. Normalization of ventilation

–perfusion mismatchd. All of the abovee. a and c

24. The differential diagnosis of hypotensionin the setting of massive transfusion aftermajor blunt trauma includes all of the fol-lowing except:a. Hypocalcemiab. Transfusion reactionc. Hypovolemiad. Tension pneumothoraxe. All of the above

25. Which of the following products carries thehighest risk of infection?a. 5 units packed red blood cellsb. 2 units fresh frozen plasmac. 6 units plateletsd. 3 units whole bloode. 2 L 0.9% saline

26. Is there an exact transfusion trigger HCTat which all patients should be transfused:a. Yes b. No

27. Can hepatitis C be transmitted via bloodtransfusion?a. Yes b. No

28. Does transfusion result in immunosup-pression of the recipient?a. Yes b. No

29. In general, will patients with histories ofimpaired cardiac function or cardiac is-chemia require transfusion at higher orlower HCT levels?a. Higher b. Lower

30. Is there a relationship between the num-ber of units of blood transfused and infec-tion in trauma patients?a. Yes b. No

31. Complications of subclavian and internaljugular catheterization includea. air embolismb. hemothoraxc. pneumothoraxd. sepsise. all of the above


32. In a patient with multiple stab wounds tothe abdomen, which of the followingwould provide adequate venous access?a. a large-bore femoral catheterb. two upper extremity 14-gauge

IV cathetersc. a saphenous cutdownd. a right internal jugular triple lumene. none of the above

33. Choose the incorrect statement regardingvenous access in the trauma patient:a. Venous cutdowns provide rapid, se-

cure, large-bore venous access.b. Two large-bore percutaneous cath-

eters should be placed immediately.c. A central line should be inserted in

all trauma patients.d. Main complications of venous cut-

down are nerve injury and infection.e. The major complications of internal

jugular cannulation are pneumotho-rax and carotid puncture.

34. Choose the incorrect statement:a. Long, large-bore IV catheters should

be used for rapid IV fluid infusion.b. Thrombosis of femoral catheters oc-

curs more often than with subclaviancatheters.

c. Subclavian catheterization should beattempted in the side of injury in apatient with a chest wound.

d. Strict aseptic technique should alwaysbe used in central line placement.

e. Venous air embolism is often a fatalcomplication.

35. Intraosseous cathetersa. are recommended only in children.b. should be considered after two un-

successful percutaneous IV attemptsin the pediatric trauma patient.

c. do not provide adequate venous ac-cess for fluid administration.

d. should be used only as a last resortin a trauma patient.

e. none of the above.

36. Regarding the impact of rapidly infusingunwarmed IV fluids in a 70-kg anesthe-tized patient:a. 4.5 L of 21°C crystalloid will result in

~1.0-1.5°C decrease in mean bodytemperature.

b. 4 units of 4°C red cells diluted in 0.9%saline will result in ~ 1.0-1.5°C de-crease in mean body temperature.

c. Red cells may be warmed safely to amaximum temperature of 42°C.

d. Gas embolism may occur, especiallywith the use of constant pressurizedinfusion devices.

e. All of the above

For questions 37-40, match the fluid/bloodwarmer with the one best answer.Answers may be used only once.

37. Hotline38. Flotem IIe39. Level 1- H100040. Alton Dean/Mallinkrodt FW537 or FW538

a. Coiled IV tubing immersed in a wa-ter bath

b. Aluminum tube in tube countercur-

rent heat exchanger combined withheated patient line

c. Metal foil countercurrent heat ex-change

d. IV tubing sandwiched between alu-minum heating plates in a serpentinefashion—dry heat technology

e. Countercurrent heated patient lineto insure delivery of 37°C fluid atflow rates of 5–80 ml/min (300–5,000 ml/hr)

41. When a critically injured patient enters theoperating room for emergency surgicaltherapy, which of the following should bethe anesthesiologist’s #1 priority?a. TEE probe insertionb. Pulmonary artery catheter insertionc. ECG monitoringd. Blood pressure measuremente. Evaluation and management of

the airway, oxygenation, andadministration

42. Which of the following is not ideal as aroute for fluid administration in trauma?a. Use of two peripheral intravenous

catheters in the upper extremitiesb. Use of the internal jugular vein with

a short, large-bore IV catheterc. Use of the femoral vein with a large-

bore IV catheter in a patient with agunshot wound in the neck

d. Use of the femoral vein with a large-bore IV catheter in a patient with sus-pected cervical spine, abdominal, andpelvic injuries

43. Which of the following is a known stor-age lesion for PRBCs?a. decreased pHb. hemolysisc. increased concentration of potassiumd. decreased 2,3-DPGe. all of the above

44. The following are true regarding sodiumcitrate excepta. Calcium chloride should always be

given when more than 2 units ofblood are transfused to an adulttrauma patient.

b. Sodium citrate transiently decreasesionized calcium.

c. Hypocalcemia can cause hypotension.d. Hypocalcemia can cause a prolonged

QT interval.e. Hypocalcemia can cause biventricular

cardiac dysfunction.

45. Which of the following infection is themost frequently associated with bloodtransfusion in the United States?a. HIVb. Hepatitis Bc. Hepatitis Cd. HTLV 1e. HTLV 2

46. True statements regarding the Rapid In-fusion System (Haemonetics Corporation)include all of the following excepta. It features a roller pump mechanism.b. Fluids are pumped from a 3-liter hard

shell reservoir.

c. Infusion rates of up to 1,500 cc/mincan be achieved.

d. 100-cc and/or 500-cc boluses over 1minute can be infused periodically.

e. All forms of blood components maybe infused through it.

47. True statements regarding the i-STAT Por-table Clinical Analyzer (i-STAT Corp.,Princeton, NJ) include all of the follow-ing except:a. It is a hand-held unit.b. It utilizes a “thin film” biosensor re-

quiring 2 to 3 drops of blood in or-der to give results over a variety oflaboratory parameters.

c. Coagulation studies available includePT, PTT, and fibrinogen levels.

d. Blood chemistry results are obtainedwithin 2 minutes.

e. Various laboratory results can be ob-tained, depending on the particularcartridge inserted into the unit.

48. Current guidelines regarding quality con-trol in laboratory testing are mandatedthrougha. The National Committee for Clinical

Laboratory Standardsb. The Health Care Financing Adminis-

trationc. The clinical director of an individual

laboratory facilityd. The 1988 Amendment to the Clinical

Laboratory Improvement Law of 1967e. The Department of Health and Hu-

mane Services

49. In the massive transfusion scenario, truestatements regarding banked red bloodcells include all of the following excepta. Pre-washing RBCs removes a signifi-

cant proportion of citrate that may bepresent in the infused blood.

b. May be indicated in patients with ahistory of renal insufficiency.

c. Pre-washing RBCs decreases K+ con-centration of blood administered tothe patient.d. K+ concentration is unrelated tothe length of time a unit of blood hasbeen stored.

e. The risk of untoward effects of mas-sive transfusion of banked red bloodcells increases with rapidity of trans-fusion.

50. Key points of the rapid infusion strategyemployed by anesthesia personnel at theElvis Presley Memorial Trauma Center in-clude all of the following excepta. Transfusion of blood products through

the Rapid Infusion System in units of10 units PRBCS, 4 units fresh frozenplasma, and 7 units pooled platelets.

b. Dilution of each unit of red bloodcells with 500 cc of normal saline.

c. Maintenance of relative normotension.d. Communication with surgeons, the

blood bank, and lab personnel re-garding use of the RIS.

e. Infusion of fluids through the RIS at1,500 cc/min until hemostasis isachieved.


Answer Form: Please circle the one best answer for each question.Massive Transfusion Monograph — 1999/2002

Name: _______________________________________________________________________________________________________

Address: _______________________________________________________________________________________________________

_______________________________________________________________________________________________________

City: _______________________________________________________________________________________________________

State: _______________________________________________________________________________________________________

Phone: _______________________________________________________________________________________________________

1. a b c d2. a b c d3. a b c d4. a b c d5. a b c d6. a b c d e7. a b c d e8. a b c d e9. a b c d e10. a b c d e11. a b c d e12. a b c d e13. a b c d e14. a b c d e15. a b c d e16 a b c d e17. a b c d e

CME ANSWER FORM

18. a b19. a b c d e20. a b c d e21. a b c d e22. a b c d e23. a b c d e24. a b c d e25. a b c d e26. a b27. a b28. a b29. a b30. a b31. a b c d e32. a b c d e33. a b c d e34. a b c d e

Evaluation Form: Please rate this self-study activity by marking one response for each statement.

Did the articles meet the monograph’s stated objectives? ___Yes ___No

How do you rank the quality of this educational activity? ___5 (high) ____4 ___3 ___2 ___1 (low)Comments: __________________________________________________________________________________________________________

Did you perceive any evidence of bias for or against any commercial products? ___Yes ___No If yes, please explain.Comments: __________________________________________________________________________________________________________

How do you rank the effectiveness of this activity as it pertains to your practice? ___5 (high) ____4 ___3 ___2 ___1 (low)

Did this material stimulate your intellectual curiosity? ___5 (high) ____4 ___3 ___2 ___1 (low)

Additional comments about this activity: ______________________________________________________________________________________________________________________________________________________________________________________________________

35. a b c d e36. a b c d e37. a b c d e38. a b c d e39. a b c d e40. a b c d e41. a b c d e42. a b c d43. a b c d e44. a b c d e45. a b c d e46. a b c d e47. a b c d e48. a b c d e49. a b c d e50. a b c d e

Mail answer form and check ($75, members; $150, nonmembers) to ITACCS Department of CME Credit, P.O. Box 4826, Baltimore, MD 21211.Allow 4 to 6 weeks for processing. Credit for this activity is offered until January 31, 2004.

I certify that I have completed the “Massive Transfusion Monograph” activity as designed and claim 15 credit hours in Category 1 of the Physicians RecognitionAward of the American Medical Association.________________________________________________________________________________________________________________ ________________________________________________________Signature Date

Copyright © 2003 International Trauma Anesthesia and Critical Care Society

ITACCS, P.O. Box 4826, Baltimore, MD 21211 USA

http://www.itaccs.com • Fax: 410/235.8084

Orginally Published October 1999Reviewed and “reprinted” on the web January 2003

THIS EDUCATIONAL PROGRAM SUPPORTED,IN PART, THROUGH UNRESTRICTED

EDUCATIONAL GRANTS FROM

Massive Transfusion and Control of Hemorrhage in the...

Documents

Transcript of Massive Transfusion and Control of Hemorrhage in the...