transfusion 2

Review article

Intraoperative pediatric blood transfusion therapy:a review of common issues. Part II: transfusiontherapy, special considerations, and reduction ofallogenic blood transfusions

SANDRA L. BARCELONA M DM D*, ALEXIS A. THOMPSON

M DM D†§ AND CHARLES J. COTE M DM D*†‡

Departments of *Anesthesiology and †Pediatrics, The Feinberg School of Medicine atNorthwestern University, Chicago, IL, USA and Departments of ‡Pediatric Anesthesiologyand §Hematology-Oncology-Stem Cell Transplantation, Children’s Memorial Hospital,Chicago, IL, USA

Keywords: blood transfusion: massive; coagulopathy; thrombocyto-

penia; recombinant factor seven (rFVIIa)

Introduction

Part two of this paper reviews transfusion therapy

with a focus upon massive blood transfusion, new

strategies for the treatment of massive transfusion

therapy, specific situations where transfusion may be

anticipated in pediatric patients, and a brief overview

for strategies to reduce allogenic blood transfusions.

III: Transfusion therapy

A. Massive blood transfusion andcoagulopathy

Massive blood transfusion is defined as the loss of

one or more circulating blood volumes. It is, there-

fore, important to calculate the patient’s estimated

blood volume (EBV) and to relate this to the volume

of blood products and other fluids administered

(Table 1). In addition the anesthesiologist should

estimate how much blood will be allowed to be lost

before the initiation of packed RBC transfusion. The

patient’s EBV is generally related in part to the

patient’s age as well as body habitus. Younger

patients have a higher fraction of their weight as

blood while more obese patients have a lower

fraction. Once the patient’s circulating blood volume

has been estimated then a further simple calculation

allows estimation of the maximal allowable blood

loss (MABL). A simple mathematical calculation is:

MABL ¼ ðstarting hematocrit � target hematocritÞstarting hematocrit

� EBV

For example, if the child weighs 25 kg there will

be a circulating blood volume of 70 ml · 25 kg ¼�1750 ml. If that child also has a starting hematocrit

of 36% and the target hematocrit is 21% then the

following calculation applies:

MABL ¼ ð36 � 21Þ36

� 1750 ¼ � 730ml

This shed blood can be replaced with a balanced

salt solution such as lactated Ringer’s solution in a

volume of 3 : 1, i.e. �2200 ml or with 5% albumen

1 : 1 or hetastarch 1 : 1, i.e. 730 ml. Once the estima-

ted blood loss reaches this target value then RBC cell

transfusion should be initiated. Obviously preterm

and term infants, children with cyanotic congenital

heart disease, large ventilation/perfusion mismatch,

high metabolic demand, and children with respira-

tory failure, may benefit from having the hematocrit

Correspondence to: Sandra L. Barcelona, Department of PediatricAnesthesiology (Box 19), Children’s Memorial Hospital, 2300Children’s Plaza, Chicago, IL 60614, USA (email: [email protected]).

Pediatric Anesthesia 2005 15: 814–830 doi:10.1111/j.1460-9592.2004.01549.x

814 � 2005 Blackwell Publishing Ltd

maintained at a higher target value. Conversely,

children in countries with poor blood product

screening may benefit from avoiding transfusion

until a much lower target hematocrit, e.g. �5 gÆdl)1.

Since the hematocrit in packed RBCs is approxi-

mately 70%, each 100 ml of packed RBCs transfused

will provide 70 ml of RBCs. In the example given

above, if the blood loss exceeded the MABL by

150 ml and assuming that the target hematocrit was

30%, then replacement would be made according to

the following formula:

(i) volume of blood to replace (150 ml) · desired

hematocrit (30%) ¼ 45 ml of 100% RBCs;

(ii) the approximate hematocrit in packed RBCs is

70%;

(iii) therefore, 45 ml ‚ 0.70 ¼ approximately 65 ml

of packed RBCs;

(iv) 65 ml of packed RBCs (�hematocrit 70%) is

equivalent to 150 ml of (whole) blood with a

hematocrit of 30%.

This can usually be simplified by transfusing

approximately 0.5 ml packed RBCs for each millilit-

er of blood loss beyond the MABL; this will result in

a slightly higher hematocrit than the target 30% but

since all of these calculations are estimates the end

result is usually close to the desired value.

The coagulopathy of massive blood transfusion is

caused by both a reduction in platelets (dilutional

thrombocytopenia) and a reduction in clotting fac-

tors. In general, a good rule of thumb is that a patient

will lose approximately 40% of the starting platelet

count with the first blood volume lost, another 20% of

the starting platelet count with the second blood

volume lost, and approximately 10% of the starting

platelet count with the third blood volume lost

(Figure 1) (1). Thus at three blood volumes lost with

no platelet replacement, the platelet count would be

expected to be reduced by approximately 70% from

baseline values (1). This emphasizes why it is so vital

to have a baseline platelet count when massive blood

loss is anticipated since patients who begin with a low

platelet count are at risk for pathological bleeding

even after only one blood volume loss whereas those

who start with an abnormally high platelet count

may not require transfusion despite blood loss equiv-

alent to four or five circulating blood volumes (1).

The coagulopathy secondary to dilution of clotting

factors depends upon the type and volume of

transfused blood product. Whole blood contains all

clotting factors including fibrinogen at normal val-

ues except for the labile factors (FV and FVIII); even

these factors are present in 20–50% of their normal

values. Therefore with transfusion of whole blood,

despite reduced amounts of the labile clotting factors

(FV and FVIII), pathological coagulation generally

does not occur until three or more blood volumes

have been lost (2–4). However, such large quantities

of whole blood are rarely used with modern blood

banking, instead the blood loss is replaced with

albumen, starch, crystalloid, and packed RBCs. With

this type of replacement multiple clotting factor

deficiencies will appear much earlier since approxi-

mately 80% of the coagulation factors have been

separated into the plasma fraction. In this situation,

clotting factor deficiency should be anticipated once

blood loss exceeds one blood volume (Table 2).

The evaluation of coagulopathy depends in part on

the availability of specific laboratory tests that can be

rapidly returned to the anesthesiologist so as to allow

the intraoperative evaluation of specific abnormal-

ities of coagulation. The simplest approach, when

Blood volumes lost

0.5 1.0 1.5 2.0 2.5 3.0 3.5Per

cen

t ch

ang

e in

pla

tele

t co

un

t fr

om

bas

elin

e

–100

–80

–60

–40

–20

0

Figure 1

Serial changes in platelet counts as a percent of baseline(mean ± SD) in 26 children who lost one to three blood volumesduring surgery and did not receive exogenous platelets. Patientswhose platelet counts were ‡50 · 109 l)1 did not exhibit clinicallyimportant bleeding [modified from (1)].

Table 1

Estimated circulating blood volume versus age or weight

Age Estimated blood volume (mlÆkg)1)

Premature infant 90–100Term infant – 3 months 80–90Children older than 3 months 70Very obese children 65

INTRAOPERATIVE BLOOD TRANSFUSION THERAPY 815

� 2005 Blackwell Publishing Ltd, Pediatric Anesthesia, 15, 814–830

laboratory tests are slow to return or unavailable, is

simply drawing a sample of blood into a blood tube

without anticoagulant and observing the rapidity of

clot formation and the presence or absence of clot

lysis. A clot should form in 4–8 min and there should

be no clot lysis if all coagulation parameters are

within normal limits and there is no coagulopathy. A

platelet count will provide a reference number to

indicate the need to increase platelets via transfusion

therapy. Generally, in a normal patient a platelet

count >50 000 mm)3 does not require treatment

unless further blood loss is expected. Obviously

cardiac bypass is a different situation where platelets

become dysfunctional due to the trauma of the

bypass machine. A prothrombin time (PT) will

provide global information regarding the so called

extrinsic clotting system (factors VII, X, V, prothrom-

bin and fibrinogen); an acceptable value would be an

INR of <1.5 (less than 1.5 times normal). An activated

partial thromboplastin time (PTT) provides global

information regarding the so-called intrinsic clotting

system (factors XII, XI, IX, VIII, X, V, prothrombin and

fibrinogen); a value <50 s (less than 1.5 times normal)

does not generally require treatment. The presence of

a significant increase in fibrin split products com-

bined with a low fibrinogen value generally indicates

disseminated intravascular coagulopathy. In this

situation, cryoprecipitate or fresh frozen plasma

may provide the needed increase in factors. More

sophisticated tests such as the platelet function

analyzer (PFA) will provide a measure of platelet

function that is more reliable than bleeding times in

children. Thromboelastography (TEG) will provide

information regarding the rapidity of clot formation

and clot lysis as well as a rapid means for determin-

ing the effectiveness of each intervention. Neither

PFA nor TEG have been found to have a positive

predictive value for perioperative bleeding in most

settings. Management of specific coagulation abnor-

malities such as hemophilia and von Willebrand

disease are beyond the scope of this review but new

approaches to coagulation are presented below.

Table 3 presents our indications for monitoring

based upon the anticipated blood loss. Recent advan-

ces in transfusion therapy such as the development of

recombinant clotting factor replacement therapy may

dramatically alter our current approach to patients

undergoing massive blood transfusions resulting

in either dilutional thrombocytopenia or multiple

clotting factor deficiencies. Such therapy will allow

specific clotting factor replacement and reduce expo-

sure to homologous blood products and the potential

for disease transmission or transfusion reactions.

B. Recombinant factor VIIa

A novel but at present enormously expensive

approach to treat bleeding is the use of recombinant

factor VIIa (rFVIIa) (NovoSeven�, Novo Nordisk

Pharmaceutical Inc., Princeton, NJ, USA). This prod-

uct was developed for the management of bleeding

complications in hemophilia patients who also have

alloantibodies against factors VIII and IX (5–8).

However, it is likely that further research will find

many additional indications for patients with coag-

ulopathy from a variety of causes (Figure 2) (9). The

current model of the coagulation cascade is based

upon the observation that formation of a tissue

factor-factor VIIa complex (TF-FVIIa) is the major

initiating event of in vivo hemostasis (Figure 3). This

paradigm differs from the traditional coagulation

cascade involving separate intrinsic and extrinsic

pathways and helps to explain the importance of the

presence of adequate amounts of critical factors (i.e.

VII, X, V, VIII and XI) while deficiencies of others

(i.e. XII) do not result in clinical bleeding. Tissue

factor is normally located deep within vessel walls

and is not usually exposed to the general circulation

until an injury occurs. Injury to tissue results in

release of TF from the endothelium. Tissue factor

then forms a complex with activated factor VII

(VII fi VIIa). Factor VIIa in turn activates factor X

(X fi Xa) and prothrombin, which promotes the

production of thrombin and the formation of fibrin

clots at the site of injury (10). It appears that rFVIIa

Table 2

Changes in prothrombin (PT) and partial thromboplastin (PTT)times during massive blood transfusions in children receivingeither packed RBCs and albumen or packed RBCs and crystalloid(66)

Blood volumes lost Baseline 0.5 0.75 1.0

Prothrombin timen 26 16 12 10Mean ± SD 10.9 ± 0.96 12.5 ± 0.77 13.2 ± 0.76 13.6 ± 0.98Range 9.3 – 12.0 11.4 – 14.0 11.4 – 14.2 11.9 – 15.8

Partial thromboplastin timeMean ± SD 31.8 ± 4.4 38.0 ± 4.9 40 ± 5.4 45.1 ± 13.1Range 25 – 45.9 28.1 – 59.6 33 – 51.5 25.6 – 60

816 S.L. BARCELONA ET AL.


also activates factors IX and X on the surface of

activated platelets (IX fi IXa & X fi Xa), leading

to the generation of additional amounts of thrombin

(11–17). Activated platelets serve as a surface site for

activation of factors VIII (VIII fi VIIIa), and XIa, as

well as further interaction with activated factors X

and V (V fi Va) to create thrombin. Effective

hemostasis requires full thrombin generation in the

presence of activated platelets. Thus rVIIa adminis-

tration promotes hemostasis at several locations

within the coagulation cascade and may be useful

in normal patients as well as patients with thromb-

ocytopenia (18–20).

The pharmacokinetics of rFVIIa have been most

extensively studied in patients with hemophilia. It

is important to note that the half-life of rFVIIa is

shorter (1.3 vs. 2.7 h) and the clearance faster (67

vs. 37 mlÆkg)1Æh)1) in pediatric patients compared

with adults (21). There may also be differences in

patients who have ongoing blood loss compared

with hemophilia patients (21). Studies of small

numbers of cirrhotic patients and those undergoing

liver transplant have revealed altered pharmacoki-

netics as well (21). Initial dosing regimens for

hemophiliacs undergoing surgery include an intra-

venous dose of 90–120 lgÆkg)1 of rFVIIa (22).

Studies completed in trauma and surgical patients

have used somewhat higher dosing (up to

200 lgÆkg)1) (23). Follow-up dosing intervals of 2–

4 h because of the relatively short circulating half

Table 3

Estimated predicted blood lossand recommended monitoringand equipment

Predicted blood loss Recommended monitors/equipment

<0.5 blood volumes Routine monitoring0.5–1.0 blood volumes Routing monitoring + urine catheter1.0 blood volumes or greater Routine monitoring + urine catheter

+ CVP + arterial line1.0 blood volumes or greaterwith potential for rapid blood loss

Routine monitoring + urine catheter+ CVP + arterial line + large bore IV+ rapid infusion device

Severe head injury Routine monitoring + urine catheter+ CVP + arterial line + large bore IV

Major trauma with unknown severity Routine monitoring + urine catheter+ CVP + arterial line + large bore IV+ rapid infusion device

Modified from Cote CJ, Dsida RM. Strategies for blood product management and transfusionreduction, In: A Practice of Anesthesia for Infants and Children, 3rd edn. Philadelphia, PA: WBSaunders Co., 2001.

Factordeficiency

Generalhemostasis

Plateletdisorder

Lowplateletcount

Defectiveplateletfunction

SurgicalbleedingAcute

bleeding

Multipleclottingfactors

Singlefactor

rVIIa

GI

CNS Post-bypass

Massivetransfusion

Figure 2

Proposed likely indications forthe use of recombinant factor VIIa(NovoSeven�). Note that thisagent is not yet approved for usein children except for the treat-ment of hemophilia von Wille-brand disease with high titerinhibitors or for congenital factorVII deficiency. However manyclinical trials in adults and anec-dotal reports in children suggestpotential use for a wide variety ofconditions resulting in patho-logical bleeding. See text fordetails. Modified from (9).



life of NovoSeven, adds considerably to the ex-

pense of using this agent.

Satisfactory hemostasis with rFVIIa administra-

tion has been achieved in critically ill massively

transfused trauma victims, patients undergoing

cardiopulmonary bypass, those sustaining liver

injury or undergoing liver transplantation, as well

as uncontrolled gastrointestinal or intraabdominal

hemorrhage (16, 24–32). Neurosurgical patients with

coagulopathy have also received benefit from rFVIIa

administration in order to correct coagulopathy

prior to surgical intervention to evacuate intracranial

bleeding (33).

The use of rFVIIa has been shown in adults to

decrease prothrombin time and increase hemostasis

at the site of injury (34–36). The sensitivity of plasma

assays such as the prothrombin time (PT) to even the

smallest quantities of activated factor VII do not

correlate with clinical outcomes and thus makes

these tests unsuitable for monitoring responses to

rFVIIa therapy (37). A more promising strategy has

been the use of TEG. These whole blood assays

measure several aspects of effective hemostasis,

including the rate of thrombin generation, as well

as fibrin clot strength, elasticity and degradation in

the presence of activated platelets (38–40). This

approach has been far more reliable in measuring

responses to rFVIIa in hemophilia patients with

high-titer factor VIII inhibitors, in normal subjects,

and has also been used for monitoring anticoagula-

tion therapy.

The administration of rFVIIa provides the advant-

age of hemostasis at the site of injury, without

systemic activation of the coagulation cascade (11–

16, 41). Although rFVIIa has proven useful in

multiple situations, at the time of this writing, it is

not United States Food and Drug Administration

(FDA) approved for use in trauma or surgical

procedures. The initial intention and only FDA

approved use is for patients with hemophilia or

von Willebrand disease who have developed inhib-

itor antibodies or patients with congenital factor VII

deficiency. There are limited pediatric data at this

point but anecdotal cases suggest some promise for

treating a variety of conditions with pathological

bleeding (27, 33, 42–51). Because rFVIIa uses recom-

binant technology, no material of human origin or

blood product is used in its production (17). It,

VASCULAR INJURY

TF exposureVII

VIIa + TF

IXX Xa + VaVWF-FVIII

Fibrinogen Fibrin Clotformation

Prothrombin

Platelet activationVIIIa + IXa

THROMBIN

Figure 3

Schematic diagram of current model of the coagulation cascade. Hemostasis is initiated following tissue factor (TF) release and activation offactor VII (FVIIa). FVIIa mediates coagulation through the generation of thrombin, activation of factor X (FXa), and activation of factor IX(FIXa). Activation of factors V, VIII, and IX occurs on the surface of activated platelets in response to thrombin generation.



therefore, may aid in the control of surgical bleeding

in Jehovah’s Witness patients and help to avoid

transfusion. It also is less immunogenic and elimin-

ates the risk of viral or bacterial contamination that

can occur with blood product transfusion. It would

appear that this new weapon for treating coagulop-

athy will be available to anesthesiologists in the near

future. It is also likely that the cost of such treatment

will be quite enormous thus mandating very clear

indications. Potential concerns regarding the recom-

binant products are allergic reactions, inhibitor

formation, and thromboses (6).

C. Platelet infusions

In general, most pediatric patients will have platelet

counts within the normal range at the time surgery is

needed, however, those with chronic inflammatory

diseases and patients with functional or surgical

asplenia may have high platelet counts. Patients

with infections, on myelosuppressive chemotherapy,

or those with ongoing coagulopathies will be at

risk for thrombocytopenia. For patients who are

otherwise normal, a platelet count greater than

50 000 mm)3 is required in order to maintain hemo-

stasis (1). Patients who are chronically thrombo-

cytopenic will often not manifest bleeding

tendencies even when the platelet count is in the

10 000–20 000 mm)3 range. Immune or nonimmu-

nological-based peripheral platelet destruction may

shorten the survival of transfused platelets. None-

theless, perioperative platelet transfusions may be

appropriate if significant bleeding is anticipated. In

these patients a preoperative transfusion plan

should be developed in consultation with the child’s

oncologist/hematologist.

IV: Specific circumstances

A. Trauma

Trauma related injuries are the leading cause of

death in pediatric patients. Exsanguination, either at

the site of the accident, or in the hospital, accounts

for 39% of all traumatic deaths (52). The major

source of morbidity and mortality in pediatric

trauma victims is oxygen deprivation to vital organs.

Because both anemia and hypovolemia decrease

oxygen delivery, appropriate transfusion and

volume resuscitation are critical. Although rapid

infusion of fluids to a trauma patient who is

hemorrhagic and hypotensive may seem obvious,

the end points for fluid administration to hemody-

namically stable victims of blunt trauma and severe

head injury are vague and not well established. The

overtransfusion of such patients may result in

cerebral edema, pulmonary edema, or coagulopathy.

The evaluation and treatment of the pediatric

trauma patient differs from that of the adult primar-

ily because of differences in physiology. The evalu-

ation of the volume status of the pediatric trauma

patient may be difficult since children are able to

maintain a normal blood pressure in a supine

condition with a loss of up to 20% of their blood

volume. In pediatric trauma victims, postural hypo-

tension and a narrow pulse pressure may be more

sensitive signs of hypovolemia than tachycardia or

systolic hypotension. Invasive monitors can be

extremely helpful in this population, and should

never be omitted secondary to patient size. The

insertion of an arterial line greatly facilitates beat to

beat blood pressure monitoring while a central

venous line provides valuable information regarding

cardiac filling pressures particularly when the blood

loss is such that it cannot be quantitated, e.g. internal

bleeding, neurosurgical trauma. Our experience has

been that a 2–3 mmHg reduction in the central

venous pressure may reflect the loss of as much as

20% of the circulating blood volume. Likewise a

diminished arterial pulse contour (loss or migration

of the dicrotic notch) may identify hypovolemia that

would not be recognized by vital signs alone (53).

Metabolic acidosis secondary to hypoperfusion and

low or concentrated urine output are additional

indicators of hypovolemia. Proper assessment of

volume status aids interpretation of obtained hemo-

globin levels. A volume depleted child may have a

normal hemoglobin value. Once volume resuscita-

ted, however, a more realistic lower hemoglobin

level may reveal significant anemia.

Like adults, large amounts of blood may be lost

internally secondary to a long bone fracture, retro-

peritoneal bleeding or blunt abdominal trauma.

Unique to the child is the potential for substantial

bleeding from closed head trauma. Infants and

toddlers have open cranial sutures which allows

for collection of a substantial amount of intracranial

blood. Young pediatric patients may suffer from



hemorrhagic shock secondary to closed head injury,

while older children and adults are unlikely to

encounter this result.

Volume resuscitation should be initiated with

large bore intravenous access. It may be difficult to

access peripheral veins in a hypovolemic child.

Short length, thin walled catheters have been

successfully placed in the femoral or antecubital

veins of dehydrated pediatric patients (54). The

intraosseous route of fluid delivery may be life

saving for delivery of fluids, medications, and

dilute blood products when intravenous access

cannot be obtained (55–57).

The efficacy of colloid administration instead of

crystalloid has been debated for many years. Hyper-

tonic saline and dextran solutions have been demon-

strated to increase survival in hypotensive patients

with severe head injury as well as patients with

penetrating trauma when compared with standard

resuscitation with normal saline (58, 59). However,

no randomized controlled studies of pediatric

patients have been published. In addition hypertonic

saline/dextran solutions as well as hydroxyethyl

starch are potential causes of coagulopathy in

trauma patients (60). These solutions appear to

compromise clotting by interfering with factor VIII

as well as von Willebrand factor, resulting in a von

Willebrand syndrome (61, 62).

The severely injured child may require massive

blood transfusion. The metabolic consequences of

massive transfusion including hypothermia, hyper-

kalemia, hypocalcemia, and hypomagnesemia are

discussed in part I. Coagulopathy is also more likely

to develop in a massively transfused child who has

suffered from hypoperfusion or hypothermia (63).

Although dilutional thrombocytopenia has been

labeled as the primary reason for coagulopathy in

the massively transfused patient, most studies dem-

onstrating this point were carried out using whole

blood (4, 64). Since most transfusion therapy, espe-

cially in the trauma patient, is now administered as

packed RBCs, dilution of clotting factors may be an

equally important cause of coagulopathy. In fact, the

dilution of clotting factors has been demonstrated to

become clinically important (with PT and PTT

approaching 1.5–2 times normal levels) when 1.5

blood volumes have been transfused (Table 2) (65–

67). On the other hand, platelet counts which begin

as normal (>150 000 mm)3) are unlikely to be the

cause of clinically significant coagulopathy until

more than two blood volumes have been transfused

(Figure 1) (1).

B. Neonates

Neonates pose some unique challenges with respect

to anesthesia and blood products. One potential

complication of RBC transfusion that occurs less

frequently in neonates compared with older children

and adults is a major hemolytic reaction. For the first

3–4 months of life, infants are unable to form allo-

antibodies to RBC antigens (68–70). After initial

screening for ABO/Rh determination, repeated

screening for antibodies is generally not necessary.

With the exception of exposure to the Rh-D antigen,

hemolytic reactions due to ABO incompatibility

rarely occur in young infants because of their imma-

ture immune systems. Therefore, administration of

RBCs to infants less than 4 months of age may be

either type specific or O negative and do not require

further cross matching. Beyond the first 4 months of

life, hemolytic reactions remain an important cause

of transfusion related morbidity and mortality.

Premature infants may be more susceptible to

hypocalcemia hypothermia with large volume blood

transfusions. As noted previously, transfusion-rela-

ted graft versus host disease occurs more often in sick

neonates and may require special consideration for

the use of directed donor components from close

relatives (pretransfusion irradiation) (69, 71, 72).

Filtration to leukocyte-deplete blood products prior

to transfusions should be routine in neonates and is

likely to reduce the incidence of febrile nonhemo-

lytic transfusion reactions, transmission of CMV

infection, as well as transfusion-mediated immuno-

modulation.

C. Cardiac surgery

Children are prone to more hemostatic disturbances

after cardiopulmonary bypass (CPB) than adults.

Children with cyanotic congenital heart lesions are

at high risk for hypoxemia and resulting polycythe-

mia. Hypoxemia prolongs bleeding times by negat-

ively affecting platelet function. Patients with

congenital heart disease have also demonstrated a

high incidence (19%) of abnormal preoperative PT

and PTT values, as well as abnormalities in von



Willebrand factor (73, 74). Decreased clotting factor

production may be caused by hepatic congestion

secondary to heart failure. Disseminated intravascu-

lar coagulopathy may consume clotting proteins as

well. Multiple factors, which may be identified

preoperatively, contribute to these patients’

increased risk for perioperative bleeding.

Neonates are at highest risk for increased bleeding

after cardiac surgery in part because of immaturity

of the coagulation system (75, 76). In addition, many

newborns undergo complex repairs that require

longer cardiopulmonary bypass (CPB) times, deep

hypothermia, or circulatory arrest, all of which have

been demonstrated to result in increased perioper-

ative bleeding (76–79); the risk of perioperative

bleeding increases with decreasing patient age.

The relatively large volume of CPB pump prime

significantly dilutes the small child’s clotting factor

and platelet levels. Modern CPB pumps use much

smaller volumes and are less likely to contribute to

coagulopathy due to hemodilution (80). Ultrafiltra-

tion of the blood remaining in the CPB circuit helps

to increase the patient’s hemoglobin as well as

maintain high levels of clotting factors and platelet

numbers (81, 82). Alternatively, intraoperative

hemodilution by removing whole blood from the

patient to be given after CPB as well as bloodless

priming of the CPB pump have successfully limited

patients’ exposure to homologous blood compo-

nents after cardiac surgery (83–88). Prebypass

removal of blood for later transfusion should pre-

serve platelet function by eliminating the exposure

of the platelets to the damaging effects of the bypass

machine while also providing replacement of clot-

ting factors. This technique is limited to larger

pediatric patients. Intraoperative blood salvage (i.e.

cell saver) may also be utilized in these patients,

however, ultrafiltration yields higher levels of func-

tional platelets and clotting factors (85–88).

Aprotinin, a serine protease inhibitor, prevents

fibrinolysis and platelet activation normally seen

after CPB (89–91). When given by high dose regi-

men, it decreases blood loss in pediatric cardiac

patients undergoing complex surgery or resternoto-

my (92). In addition to its hemostatic benefits,

aprotinin has also been linked to improved postop-

erative oxygenation resulting in less postoperative

ventilation time (89). This may be attributed to its

systemic anti-inflammatory effects. Some studies

demonstrated no benefit of aprotinin in children

(93), and raised concerns regarding allergic potential

and renal impairment (94–96). However, other stud-

ies have found a low risk to benefit ratio and

reduced cost with the use of high dose aprotinin in

this pediatric surgical population (89, 92, 94, 97–99).

The risk of allergic reaction upon secondary expo-

sure to aprotinin in the pediatric population is less

than half that of adults’ (1.2% vs. 2.7%) (100). To

further lessen this incidence, the time between

aprotinin exposures is recommended to be greater

than 6 months (101). In high risk patients, a test dose

of 10 000 KIU should be given when surgeons are

prepared to initiate bypass quickly, and preemptive

histamine blockade should be considered (101).

D. Tonsillectomy and adenoidectomy

Tonsillectomy is one of the most common pediatric

surgical procedures. Although perioperative trans-

fusion is rare, the procedure carries potential for

life threatening hemorrhage. The majority of early

bleeding episodes occur within the first 4–6 h after

the procedure (102–105). Late bleeding typically

occurs 5–10 days postoperatively (102–105). Late

bleeding is commonly associated with an episode

of early bleeding (104). In either situation, a

significant amount of blood may be swallowed by

the child, which makes assessment of its severity

difficult. The child who is hypovolemic from

decreased oral intake secondary to pain may have

an artificially elevated hematocrit. Careful assess-

ment of the patient’s volume status and laboratory

values are imperative to a proper evaluation and

preparation for reoperation. Blood pressure and

heart rate determination in both the supine

and sitting position may shed light upon the child’s

intravascular volume status.

No universal practice of preoperative screening

for increased risk of bleeding has been adopted by

all otolaryngologists. Multiple studies have demon-

strated that neither preoperative bleeding history

nor routine laboratory screening with PT, PTT, and

bleeding time are reliable predictors of postoperative

bleeding (106–109). However, laboratory screening

may be useful in patients who have a history

suggestive of a major bleeding disorder. Even in

these individuals, screening tests such as the PT and

PTT may fail to identify some patients with bleeding



tendencies. The platelet function analyzer (PFA) is a

relatively new tool which simulates primary hemos-

tasis and is particularly well suited for evaluating

suspected cases of von Willebrand’s disease or

intrinsic platelet defects (110, 111). More formal

consultation with a hematologist is strongly recom-

mended to determine the most appropriate course of

preparation for surgery. For example, patients with

the most common type of von Willebrand disease

can undergo successful, nonhemorrhagic adenoton-

sillar surgery if treated appropriately with DDAVP

[desmopressin (1-deamino-8-DD-arginine vasopres-

sin)]. Other subtypes, however require the use of

cryoprecipitate or plasma-derived factor VIII con-

centrate containing von Willebrand factor (112).

Bleeding in patients with intrinsic platelet defects

can be prevented or controlled with platelet trans-

fusions. Without proper family and patient history

and subsequent testing, these patients may not be

identified preoperatively and may suffer from

marked, uncontrollable bleeding. It appears that

meticulous surgical technique is the most important

factor assuring that postoperative hemostasis is

maintained regardless of the patient’s history.

The use of ketorolac for postoperative pain control

in tonsillectomy patients has been debated in the

literature. Although an effective analgesic, concerns

regarding its antiplatelet activity have limited its

use. Several studies have demonstrated increased

perioperative bleeding associated with ketorolac use

for analgesia following tonsillectomy, however,

none identify the need for further intervention in

these patients (113–117). Ketorolac has been demon-

strated to lower pain scores when compared with

placebo (118). It is unclear whether or not ketorolac

use changes clinical outcome in the pediatric adeno-

tonsillar surgery population. However, because of

its theoretical risk of inhibiting platelet function, it is

prudent to reserve its use until after the surgery has

been completed and hemostasis obtained. After

initial hemostasis has been obtained, the influence

of platelet function upon subsequent hemorrhage

should be less (119). The next generation of cox-2

inhibitors will not have this antiplatelet effect and

should provide analgesia without respiratory

depression or the potential for bleeding. However,

other methods of pain control have been shown to be

equally efficacious without the potential risk for

increased bleeding (114, 115, 117).

E. Orthopedics

The vast majority of pediatric orthopedic procedures

are accomplished without the need for blood trans-

fusions. However, the inability to use a surgical

tourniquet (such as femoral osteotomy or spinal

fusion cases) can result in clinically important blood

loss. The incidence and severity of the blood loss is

often dependent upon the skills of the surgeon. A

number of techniques such as hemodilution and

controlled hypotension have been advocated to

reduce the exposure to exogenous blood products,

however concerns for impaired spinal cord blood

flow during spinal instrumentation have altered our

practice. We no longer advocate combined con-

trolled hypotension and extreme hemodilution. If

we reduce blood pressure it is only reduced to a

modest range (mean arterial pressure 65–75 mmHg)

and we generally do not allow the hemoglobin to fall

below 8 gÆdl)1. If the patient develops abnormal

bleeding in an early part of the procedure it is vital

to check the patient’s position to assure that the

support bolsters have not changed position resulting

in increased intraabdominal pressure thereby

increasing venous pressure by mechanical obstruc-

tion.

One further concern is that after major spine

fusion procedures there may be the potential for

continued bleeding from the raw surfaces of bone;

therefore, it is our practice to transfuse these patients

so that their hemoglobin at the end of surgery is in

the 10–11 gÆdl)1 range. The use of aprotinin, epsilon

aminocaproic acid, tranexamic acid, or DDAVP has

failed to consistently demonstrate decreased blood

loss (120).

F. Neurosurgery

One study has demonstrated that pediatric patients

undergoing neurosurgical procedures may have a

hypercoagulable state postoperatively as demonstra-

ted by TEG (121). However, the primary cause of

problems related to hemostasis in neurosurgical

procedures are attributable to massive blood loss

and subsequent dilutional coagulopathy or dissem-

inated intravascular coagulopathy (DIC). Neural

tissue injury due to trauma can also result in tissue

thromboplastin release and subsequent activation

of the coagulation cascade via FVIIa. Children



undergoing resection of vascular malformations and

vascular tumors (e.g. choroid plexus adenoma) as

well as trauma patients with epidural or subdural

hematomas are at great risk for massive and rapid

blood loss. Children with traumatic brain injury and

shaken baby syndrome are at particular risk for the

development of DIC. In each of these cases it is

prudent for the clinician to use invasive monitoring

(arterial and central venous monitoring) as well as

large bore intravenous catheters with appropriate

blood warming devices (see part I). It is rare for

children undergoing elective resection of brain

tumor to develop a coagulopathy unless it is due

to dilution of clotting factors.

G. Burns

Children with burn injuries can develop a variety of

clotting factor abnormalities. These abnormalities

depend in part upon the extent of the burn injury,

the presence or absence of sepsis, and the volume of

blood shed during reconstructive surgery. Early

after massive burn injury there is both a consumpt-

ive coagulopathy and a microangiopathic hemolytic

process (122, 123). Thus anemia, thrombocytopenia,

and evidence of coagulopathy are common. After

the initial 3–5 days following burn injury the typical

anti-inflammatory response occurs and patients

develop marked increases in fibrinogen, platelets,

as well as a variety of clotting factors (122–124).

Platelet counts over 1000 000 mm)3 and fibrinogen

values over 2 gÆdl)1 may be observed. Despite these

abnormalities it is rare for pediatric burn victims to

suffer thrombotic events. Conversely with the onset

of sepsis there may be a sudden fall in the platelet

count. Excision of burn wounds can also involve

rapid and massive blood loss; some of this blood loss

may be reduced by using saline with dilute concen-

trations of epinephrine injected under both the

donor and recipient sites (clysis) (125). In addition,

because the skin has been damaged it no longer

provides the normal insulation to the body such that

these patients are particularly prone to hypothermia.

Thus, although uncomfortable, it is wise to use a

very warm operating room (35�C). The administra-

tion of blood products should be guided by serial

platelet counts and evaluation of the PT and PTT

(66). In general, abnormal bleeding does not occur if

the platelet count is maintained above 50 000 mm)3

(1). It should also be kept in mind that central

venous lines, especially those with multiple ports,

will be inadequate to rapidly administer RBCs since

there is so much resistance (126). Short-term femoral

venous catheterization for the procedure with a

large bore short catheter provides a better means for

rapid RBC transfusion.

H. Kidney transplantation

Patients with renal failure have anemia for multiple

reasons. As kidney failure progresses, erythropoietin

production decreases, red blood cell life span shor-

tens secondary to hemolysis from circulating toxins,

and iron and folic acid deficiencies ensue (127). In

addition, hemodialysis increases the severity of

anemia secondary to chronic blood loss within the

dialysis circuit and progressive reduction in ery-

thropoietin levels compared with peritoneal dialysis

(128). Nevertheless, this anemia (typical hematocrit

25%) is chronic, and therefore, generally well toler-

ated if the patient is euvolemic. The threshold for

preoperative transfusion of these patients should be

lower than for the general population in whom

perioperative anemia is more likely to be acute and

less well tolerated. Historically, blood transfusion

was performed pretransplant for the purpose of

inducing immunosuppression. With the advent of

more controlled medical immunosuppressive ther-

apy (e.g. cyclosporin), the practice of automatic

pretransplant transfusion of these patients has

mostly fallen out of favor. In fact, there is some

evidence that blood transfusion to pediatric trans-

plant recipients may actually be contributory to the

increased incidence of rejection in this population

(129). Alternatively, recombinant erythropoietin [Ep-

ogen� (rhEPO), Amgen, Inc., Thousand Oaks, CA,

USA] has successfully increased hemoglobin levels

of pediatric renal failure patients over a period of a

few months (128). Erythropoietin therapy may aid in

increasing the patient’s oxygen carrying capacity if

begun enough in advance of a planned renal

transplant (e.g. a living donor).

Patients with renal failure are at increased risk for

platelet dysfunction secondary to uremia thereby

increasing bleeding tendencies. Although thrombo-

cytopenia is not uncommon, platelet dysfunction

appears to be more problematic. There are multiple

causes of platelet dysfunction in this population,



some of which can be at least partially treated with

dialysis (130, 131). Platelet transfusion as well as

DDAVP may be helpful in the patient with a

continued prolonged bleeding time despite dialysis

(132).

Pediatric renal transplants differ from those of

adults intraoperatively secondary to size discrepan-

cies between donor organ and recipient. It is com-

mon for small pediatric patients to receive a kidney

from an adult donor. In this situation, it is possible

for a large amount of blood (150–250 ml) to be

sequestered into the organ upon vascular unclamp-

ing. This may be a significant portion of the pediatric

recipient’s blood volume and may result in hypo-

volemia and hypotension upon reperfusion. Central

venous pressure monitoring is helpful in anticipa-

tion of this problem and also because urine output is

an unreliable indicator of volume status in these

patients. In some circumstances a unit of packed

RBCs is administered prior to reperfusion of the

donor kidney.

I. Liver transplantation

Pediatric liver transplant patients are often anemic.

Chronic disease, nutritional deficiencies, and occult

blood loss contribute to decreased hemoglobin

levels. These patients are typically coagulopathic

for a variety of reasons including but not limited to

hypersplenism and decreased clotting factor pro-

duction. Malabsorption of vitamin K leads to

decreased production of vitamin K-dependent fac-

tors II, VII, IX and X resulting in elevated PT and

PTT. Thrombocytopenia may be the result of splenic

sequestration secondary to portal hypertension, DIC,

or dilution from increased plasma volume. Although

preoperative coagulation tests have not been shown

to correlate with intraoperative blood loss during

pediatric liver transplantation, they are important

indicators of the severity of liver disease and help to

establish a baseline to which intraoperative values

can be compared (133, 134). Hemoglobin levels,

platelet numbers, PT and PTT values and whole

blood clotting assays such as the thromboelastogram

(TEG) are used to guide perioperative blood product

administration.

Pediatric liver transplant patients are at even

higher risk than adults for requiring massive trans-

fusion. Infants have demonstrated increased trans-

fusion needs during liver transplant than older

children and adults (135). Previous abdominal sur-

gery has been linked to increased blood use during

liver transplantation (134). The most common etiol-

ogy of pediatric liver failure is biliary atresia for

which previous abdominal surgery (i.e. Kasai por-

toenterostomy), is often performed prior to trans-

plant. The use of reduced-sized donor livers is

common in children because they often receive a

lobe or segment of a living related donor organ, or a

split cadaver organ which can serve two patients.

The use of reduced sized grafts has also been

associated with increased intraoperative bleeding

(136).

Adequate venous access is critical for these

patients. Large bore intravenous lines should be

placed in the upper extremities because the inferior

vena cava may be cross-clamped intraoperatively. If

adequate access in the upper extremities is not

possible, large bore access may be placed via the

subclavian or jugular veins. Great care must be taken

when inserting large central venous cannulas in the

neck region if coagulopathy exists as hematoma

formation may be difficult to control. Ultrasound

guided catheter insertion may be very advantageous

in this population (137, 138). The hospital blood

bank must have well-organized and well-equipped

resources to handle the potential massive transfu-

sion needs of these patients (multiple blood vol-

umes). Our blood bank defines massive blood

transfusion according to Table 4. For this popula-

tion, our practice is to have one blood volume of

cross-matched packed RBCs, five to 15 units of fresh

frozen plasma and two to four plasma pheresis

packs of platelets. After one blood volume of blood

Table 4

Units of crossed matched blood transfused within 24 h afterwhich cross-match is no longer required and type specific redblood cells may be utilized

Patient weight (kg) Packed red blood cell units transfused per 24 h

0.5–7 27–15 3

16–20 421–25 526–30 631–35 736–40 841–50 9>50 10



transfusion a cross-match is not required and type

specific packed RBCs are used. The anesthesiologist

must be prepared for complications of massive

blood transfusion including hypothermia, hypocal-

cemia, hypomagnesemia, and hyperkalemia. Citrate

induced hypocalcemia is a particular problem in

these patients because of decreased ability to meta-

bolize citrate secondary to liver failure and the

inability to metabolize citrate during the anhepatic

stage of the transplantation. If rapid and massive

transfusion is expected, a calcium chloride infusion

may be prudent to avoid hemodynamic instability

caused by rapidly changing ionized calcium levels.

V: Reduction of allogeneic bloodadministration

There are multiple methods aimed at reducing

perioperative homologous blood administration

(139, 140). For planned surgery in which blood

transfusion is expected (e.g. scoliosis repair), the

possibilities of autologous blood donation, erythro-

poietin therapy, and cell salvage should be consid-

ered. Autologous blood donation involves the

removal of the patient’s blood a week or more prior

to surgery. This blood is then treated and processed

by a blood bank center and is released for that

specific patient’s use on the day of surgery. This

process theoretically limits the likelihood of trans-

fusion related infectious disease (except bacterial)

and decreases the incidence of incompatibility

issues. The unit must be cross-matched to a fresh

sample of the patient’s blood prior to release.

However, clerical error in processing, labeling,

testing, and final checking of the unit is still possible,

and thus administration of what is deemed an

autologous unit is not without risk. In addition, the

metabolic consequences of transfusion of autologous

blood are the same as with allogeneic transfusion.

Autologous blood is often stored as whole blood and

will thus contain all clotting factors in normal

concentrations except labile factors V and VIII.

Whole blood will also carry an increased risk for

hyperkalemia and/or hypocalcemia if older than

7 days and rapidly administered. The hematocrit

will be lower than packed RBCs. Autologous dona-

tion has successfully decreased allogeneic blood

transfusion for children 8–25 kg undergoing ortho-

pedic or abdominal procedures (141).

An alternative method of autologous blood use

without predonation through a blood bank is acute

normovolemic hemodilution (ANH) at the begin-

ning of surgery and prior to the onset of surgical

blood loss. With this method, blood is removed

from the patient (usually from an arterial line) and

collected into appropriate storage bags (placed on a

scale to quantitate the volume removed). Careful

replacement of the removed blood volume with

crystalloid in a 3 : 1 ratio is important in order to

maintain normovolemia. The amount of blood to be

removed should be precalculated and the collection

bag weighed in order to avoid withdrawal of too

much blood. This is important when caring for

small children because one must think in the terms

of milliliters of blood instead of units of blood. The

anesthesiologist must also be careful not to overfill

the storage bag. The storage bags contain a speci-

fied amount of citrate as an anticoagulant for

450 ml of whole blood. If the bag is overfilled,

there may be insufficient anticoagulant to prevent

thrombus formation. After removal of this blood,

the patient then sheds intraoperative blood with a

lower hematocrit (i.e. fewer RBCs per plasma

volume lost to surgery) and the autologous blood

can be administered after the majority of blood loss

has been completed. The use of ANH has been

shown to effectively decrease the need for alloge-

neic blood transfusion in adolescent spine surgery

(142). If the removed autologous unit remains in

the operating room (preferably in a cooler to reduce

RBC metabolism and the potential for bacterial

growth), the likelihood of clerical error and incom-

patibility is minimized, however the metabolic

consequences of transfusion still exist.

Preoperative administration of erythropoietin has

been used to increase hemoglobin levels and reduce

the need for intraoperative or postoperative trans-

fusion (143). Many studies demonstrating the effic-

acy of erythropoietin have involved adults

undergoing orthopedic procedures (144–146). Ery-

thropoietin may also be used in combination with

autologous blood donation to further reduce the

need for allogeneic blood exposure. One small study

involving children undergoing open heart proce-

dures was successful in eliminating allogeneic blood

exposure through a combination of autologous

predonation and erythropoietin therapy over a

period of 2 weeks (147).



Intraoperative cell salvage allows surgical blood

loss to be collected, washed, and hemoconcentrated

in preparation for autologous donation. The qual-

ities of cell salvaged blood include an average

hematocrit of 50%, normal erythrocyte survival,

and minimal functional platelets (148). Contra-

indications for the use of this technique include

surgical field contamination with bacteria, amnionic

fluid, fat, or other potential clotting agents such as

topical thrombin (149). Originally, cancer was a

contraindication to intraoperative blood salvage for

fear of bloodborne tumor spread however an

analysis of six studies in which cell salvage was

used for cancer patients revealed no metastatic

spread (150). Also, the use of a leucocyte filter

further reduces this risk (150). There are no firm

guidelines in the literature for minimal expected

blood loss in order for cell salvage to be considered

cost effective. In our institution, we consider cell

salvage for liver transplant or scoliosis correction

patients over 30 kg.

Conclusions

Surgeons are now much more meticulous in main-

taining hemostasis in part because their patients

demand this and anesthesiologists are much more

circumspect with regard to transfusions thereby

minimizing our patient’s exposure to blood prod-

ucts. These papers were intended to present a

clinically oriented overview of perioperative blood

transfusion practices. Our purpose was to present

the topic in an easily understood format with a

generous use of tables and figures. We hope that we

provided useful strategies for a logical approach to

transfusion therapy and the associated complica-

tions as well as insight into the most recent under-

standing of the coagulation cascade and innovations

in recombinant clotting factor replacement. From

our viewpoint the most important issue is that the

blood pool is becoming increasingly safe through

better screening. The future is even more exciting

with the development of artificial blood, additional

recombinant clotting factors, and new medications

that further improve coagulation thereby further

reducing exposure to allogenic blood products. Our

hope is that blood banks and health care systems

around the world will have the resources to take

advantage of these wonderful innovations.

References1 Cote CJ, Liu LM, Szyfelbein SK et al. Changes in serial platelet

counts following massive blood transfusion in pediatricpatients. Anesthesiology 1985; 62: 197–201.

2 Cote CJ. Blood, colloid, and crystalloid therapy. Anesth ClinNorth Am 1991; 9: 865–884.

3 Barcelona SL, Cote CJ. Pediatric resuscitation in the operatingroom. Anesthesiol Clin North Am 2001; 19: 339–365.

4 Miller RD. Complications of massive blood transfusions.Anesthesiology 1973; 39: 82–93.

5 Butenas S, Brummel KE, Bouchard BA et al. How factor VIIaworks in hemophilia. J Thromb Haemost 2003; 1: 1158–1160.

6 Schlesinger KW, Ragni MV. Safety of the new generationrecombinant factor concentrates. Expert Opin Drug Saf 2002; 1:213–223.

7 Shapiro AD. Recombinant factor VIIa in the treatment ofbleeding in hemophilic children with inhibitors. Semin ThrombHemost 2000; 26: 413–419.

8 Seremetis S. Dose optimization of recombinant factor VIIa inthe treatment of acute bleeding in haemophilia-associatedinhibitors. Blood Coagul Fibrinolysis 2003; 14 Suppl 1: S29–S30.

9 Dejgaard A. Update on Novo Nordisk’s clinical trial pro-gramme on NovoSeven. Blood Coagul Fibrinolysis 2003; 14

Suppl. 1: S39–S41.10 Girard TJ, Nicholson NS. The role of tissue factor/factor VIIa

in the pathophysiology of acute thrombotic formation. CurrOpin Pharmacol 2001; 1: 159–163.

11 Midathada MV, Mehta P, Waner M et al. Recombinant factorVIIa in the treatment of bleeding. Am J Clin Pathol 2004; 121:124–137.

12 Gerotziafas GT, Chakroun T, Depasse F et al. The role ofplatelets and recombinant factor VIIa on thrombin generation,platelet activation and clot formation. Thromb Haemost 2004;91: 977–985.

13 Goodnough LT, Shander A, Brecher ME. Transfusion medi-cine: looking to the future. Lancet 2003; 361: 161–169.

14 Drews RE. Critical issues in hematology: anemia, thrombo-cytopenia, coagulopathy, and blood product transfusions incritically ill patients. Clin Chest Med 2003; 24: 607–622.

15 Laurian Y. Treatment of bleeding in patients with plateletdisorders: is there a place for recombinant factor VIIa? Path-ophysiol Haemost Thromb 2002; 32 Suppl. 1: 37–40.

16 Lynn M, Jeroukhimov I, Klein Y et al. Updates in the man-agement of severe coagulopathy in trauma patients. IntensiveCare Med 2002; 28 Suppl 2: S241–S247.

17 Jurlander B, Thim L, Klausen NK et al. Recombinant activatedfactor VII (rFVIIa): characterization, manufacturing, andclinical development. Semin Thromb Hemost 2001; 27: 373–384.

18 Poon MC. Management of thrombocytopenic bleeding: isthere a role for recombinant coagulation factor VIIa? CurrHematol Rep 2003; 2: 139–147.

19 Kjalke M, Ezban M, Monroe DM et al. High-dose factorVIIa increases initial thrombin generation and mediatesfaster platelet activation in thrombocytopenia-like condi-tions in a cell-based model system. Br J Haematol 2001; 114:114–120.

20 Hedner U, Ingerslev J. Clinical use of recombinant FVIIa(rFVIIa). Transfus Sci 1998; 19: 163–176.

21 Erhardtsen E. Pharmacokinetics of recombinant activatedfactor VII (rFVIIa). Semin Thromb Hemost 2000; 26: 385–391.

22 Abshire TC. Dose optimization of recombinant factor VIIa forcontrol of mild to moderate bleeds in inhibitor patients:



Improved efficacy with higher dosing. Semin Hematol 2004; 41Suppl. 1: 3–7.

23 Martinowitz U, Kenet G, Lubetski A et al. Possible role ofrecombinant activated factor VII (rFVIIa) in the control ofhemorrhage associated with massive trauma. Can J Anaesth2002; 49: S15–S20.

24 Al Douri M, Shafi T, Al Khudairi D et al. Effect of theadministration of recombinant activated factor VII (rFVIIa;NovoSeven) in the management of severe uncontrolledbleeding in patients undergoing heart valve replacementsurgery. Blood Coagul Fibrinolysis 2000; 11 Suppl. 1: S121–S127.

25 White B, McHale J, Ravi N et al. Successful use of recombinantFVIIa (Novoseven) in the management of intractable post-surgical intra-abdominal haemorrhage. Br J Haematol 1999;107: 677–678.

26 O’Neill PA, Bluth M, Gloster ES et al. Successful use ofrecombinant activated factor VII for trauma-associated hem-orrhage in a patient without preexisting coagulopathy.J Trauma 2002; 52: 400–405.

27 Markiewicz M, Kalicinski P, Kaminski A et al. Acute coagul-opathy after reperfusion of the liver graft in children correc-tion with recombinant activated factor VII. Transplant Proc2003; 35: 2318–2319.

28 Naik VN, Mazer CD, Latter DA et al. Successful treatmentusing recombinant factor VIIa for severe bleeding post car-diopulmonary bypass. Can J Anaesth 2003; 50: 599–602.

29 Dutton RP, Hess JR, Scalea TM. Recombinant factor VIIa forcontrol of hemorrhage: early experience in critically ill traumapatients. J Clin Anesth 2003; 15: 184–188.

30 Eikelboom JW, Bird R, Blythe D et al. Recombinant activatedfactor VII for the treatment of life-threatening haemorrhage.Blood Coagul Fibrinolysis 2003; 14: 713–717.

31 Aldouri M. The use of recombinant factor VIIa in controllingsurgical bleeding in non-haemophiliac patients. PathophysiolHaemost Thromb 2002; 32 Suppl. 1: 41–46.

32 Meijer K, Hendriks HG, de Wolf JT et al. Recombinant factorVIIa in orthotopic liver transplantation: influence on para-meters of coagulation and fibrinolysis. Blood Coagul Fibrino-lysis 2003; 14: 169–174.

33 Morenski JD, Tobias JD, Jimenez DF. Recombinant activatedfactor VII for cerebral injury-induced coagulopathy in pedi-atric patients. Report of three cases and review of the litera-ture. J Neurosurg 2003; 98: 611–616.

34 Martinowitz U, Kenet G, Segal E et al. Recombinant activatedfactor VII for adjunctive hemorrhage control in trauma.J Trauma 2001; 51: 431–438.

35 Martinowitz U, Holcomb JB, Pusateri AE et al. IntravenousrFVIIa administered for hemorrhage control in hypothermiccoagulopathic swine with grade V liver injuries. J Trauma2001; 50: 721–729.

36 Surudo T, Wojcicki M, Milkiewicz P et al. Rapid correction ofprothrombin time after low-dose recombinant factor VIIA inpatients undergoing orthotopic liver transplantation. Trans-plant Proc 2003; 35: 2323–2325.

37 Gabriel DA, Carr M, Roberts HR. Monitoring coagulation andthe clinical effects of recombinant factor VIIa. Semin Hematol2004; 41 Suppl. 1: 20–24.

38 Rajwal S, Richards M, O’Meara M. The use of recalcifiedcitrated whole blood – a pragmatic approach for thrombo-elastography in children. Paediatr Anaesth 2004; 14: 656–660.

39 Srinivasa V, Gilbertson LI, Bhavani-Shankar K. Thrombo-elastography: where is it and where is it heading? Int Anes-thesiol Clin 2001; 39: 35–49.

40 Whitten CW, Greilich PE. Thromboelastography: past, pre-sent, and future. Anesthesiology 2000; 92: 1223–1225.

41 Dunkley SM, Mackie F. Recombinant factor VIIa used to controlmassive haemorrhage during renal transplantation surgery;vascular graft remained patent. Hematology 2003; 8: 263–264.

42 Tobias JD, Berkenbosch JW, Muruve NA et al. Correction of acoagulopathy using recombinant factor VII before removal ofan intra-aortic balloon pump. J Cardiothorac Vasc Anesth 2002;16: 612–614.

43 Tobias JD, Berkenbosch JW, Russo P. Recombinant factor VIIato treat bleeding after cardiac surgery in an infant. Pediatr CritCare Med 2003; 4: 49–51.

44 Pychynska-Pokorska M, Moll JJ, Krajewski W et al. The use ofrecombinant coagulation factor VIIa in uncontrolled postop-erative bleeding in children undergoing cardiac surgery withcardiopulmonary bypass. Pediatr Crit Care Med 2004; 5: 246–250.

45 Egan JR, Lammi A, Schell DN et al. Recombinant activatedfactor VII in paediatric cardiac surgery. Intensive Care Med2003; 30: 682–685.

46 Brown JB, Emerick KM, Brown DL et al. Recombinant factorVIIa improves coagulopathy caused by liver failure. J PediatrGastroenterol Nutr 2003; 37: 268–272.

47 Chuansumrit A, Visanuyothin N, Puapunwattana S et al.Outcome of intracranial hemorrhage in infants with congen-ital factor VII deficiency. J Med Assoc Thai 2002; 85 Suppl. 4:S1059–S1064.

48 Chuansumrit A, Chantarojanasiri T, Isarangkura P et al.Recombinant activated factor VII in children with acutebleeding resulting from liver failure and disseminated intra-vascular coagulation. Blood Coagul Fibrinolysis 2000; 11 Suppl.1: S101–S105.

49 Almeida AM, Khair K, Hann I et al. The use of recombinantfactor VIIa in children with inherited platelet function disor-ders. Br J Haematol 2003; 121: 477–481.

50 Veldman A, Fischer D, Voigt B et al. Life-threatening hem-orrhage in neonates: management with recombinant activatedfactor VII. Intensive Care Med 2002; 28: 1635–1637.

51 Greisen G, Andreasen RB. Recombinant factor VIIa in pretermneonates with prolonged prothrombin time. Blood Coagul Fi-brinolysis 2003; 14: 117–120.

52 Sauaia A, Moore FA, Moore EE et al. Epidemiology of traumadeaths: a reassessment. J Trauma 1995; 38: 185–193.

53 Pinsky MR. Probing the limits of arterial pulse contour ana-lysis to predict preload responsiveness. Anesth Analg 2003; 96:1245–1247.

54 Idris A, Melker RJ. High-flow sheaths for pediatric fluidresuscitation: a comparison of flow rates with standardpediatric catheters. Pediatr Emerg Care 1992; 8: 119–122.

55 Goldstein B, Doody D, Briggs S. Emergency intraosseousinfusion in severely burned children. Pediatr Emerg Care 1990;6: 195–197.

56 Tighe SQ, Rudland SV, Kemp PM et al. Paediatric resuscita-tion in adverse circumstances: a comparison of three routes ofsystemic access. J R Nav Med Serv 1993; 79: 75–79.

57 Velasco AL, Delgado-Paredes C, Templeton J et al. Intraos-seous infusion of fluids in the initial management of hyp-ovolemic shock in young subjects. J Pediatr Surg 1991; 26: 4–8.

58 Wade CE, Grady JJ, Kramer GC. Efficacy of hypertonic salinedextran fluid resuscitation for patients with hypotension frompenetrating trauma. J Trauma 2003; 54 Suppl. 5: S144–S148.

59 Wade CE, Grady JJ, Kramer GC et al. Individual patientcohort analysis of the efficacy of hypertonic saline/dextran in



patients with traumatic brain injury and hypotension.J Trauma 1997; 42 Suppl. 5: S61–S65.

60 Maningas PA, Mattox KL, Pepe PE et al. Hypertonic saline-dextran solutions for the prehospital management of trau-matic hypotension. Am J Surg 1989; 157: 528–533.

61 Stump DC, Strauss RG, Henriksen RA et al. Effects ofhydroxyethyl starch on blood coagulation, particularly factorVIII. Transfusion 1985; 25: 349–354.

62 Strauss RG, Stump DC, Henriksen RA et al. Effects ofhydroxyethyl starch on fibrinogen, fibrin clot formation, andfibrinolysis. Transfusion 1985; 25: 230–234.

63 Patt A, McCroskey BL, Moore EE. Hypothermia-inducedcoagulopathies in trauma. Surg Clin North Am 1988; 68: 775–785.

64 Counts RB, Haisch C, Simon TL et al. Hemostasis in massivelytransfused trauma patients. Ann Surg 1979; 190: 91–99.

65 Murray DJ, Olson J, Strauss R et al. Coagulation changesduring packed red cell replacement of major blood loss.Anesthesiology 1988; 69: 839–845.

66 Cote CJ. Changes in prothrombin and partial thromboplastintimes during massive blood loss in children undergoingHarrington rod instrumentation. Section on Anesthesiology.American Academy of Pediatrics, 1988.

67 Murray DJ, Pennell BJ, Weinstein SL et al. Packed red cells inacute blood loss: dilutional coagulopathy as a cause ofsurgical bleeding. Anesth Analg 1995; 80: 336–342.

68 Cohen A, Manno C. Transfusion practices in infants receivingassisted ventilation. Clin Perinatol 1998; 25: 97–111.

69 Warren LJ, Simmer K, Roxby D et al. DNA polymorphismanalysis in transfusion-associated graft-versus-host disease.J Paediatr Child Health 1999; 35: 98–101.

70 Floss AM, Strauss RG, Goeken N et al. Multiple transfusionsfail to provoke antibodies against blood cell antigens inhuman infants. Transfusion 1986; 26: 419–422.

71 Manno CS. What’s new in transfusion medicine? Pediatr ClinNorth Am 1996; 43: 793–808.

72 DePalma L, Yu M, McIntosh CL et al. Changes in lymphocytesubpopulations as a result of cardiopulmonary bypass. Theeffect of blood transfusion. J Thorac Cardiovasc Surg 1991; 101:240–244.

73 Colon-Otero G, Gilchrist GS, Holcomb GR et al. Preoperativeevaluation of hemostasis in patients with congenital heartdisease. Mayo Clin Proc 1987; 62: 379–385.

74 Gill JC, Wilson AD, Endres-Brooks J et al. Loss of the largestvon Willebrand factor multimers from the plasma of patientswith congenital cardiac defects. Blood 1986; 67: 758–761.

75 Chambers LA, Cohen DM, Davis JT. Transfusion patterns inpediatric open heart surgery. Transfusion 1996; 36: 150–154.

76 Petaja J, Lundstrom U, Leijala M et al. Bleeding and use ofblood products after heart operations in infants. J ThoracCardiovasc Surg 1995; 109: 524–529.

77 Kern FH, Morana NJ, Sears JJ et al. Coagulation defects inneonates during cardiopulmonary bypass. Ann Thorac Surg1992; 54: 541–546.

78 Miller BE, Mochizuki T, Levy JH et al. Predicting and treatingcoagulopathies after cardiopulmonary bypass in children.Anesth Analg 1997; 85: 1196–1202.

79 Williams GD, Bratton SL, Ramamoorthy C. Factors associatedwith blood loss and blood product transfusions: a multivari-ate analysis in children after open-heart surgery. Anesth Analg1999; 89: 57–64.

80 von Segesser LK, Tozzi P, Mallbiabrrena I et al. Miniaturiza-tion in cardiopulmonary bypass. Perfusion 2003; 18: 219–224.

81 Friesen RH, Campbell DN, Clarke DR et al. Modified ultra-filtration attenuates dilutional coagulopathy in pediatric openheart operations. Ann Thorac Surg 1997; 64: 1787–1789.

82 Friesen RH, Tornabene MA, Coleman SP. Blood conservationduring pediatric cardiac surgery: ultrafiltration of the extra-corporeal circuit volume after cardiopulmonary bypass.Anesth Analg 1993; 77: 702–707.

83 Stein JI, Gombotz H, Rigler B et al. Open heart surgery inchildren of Jehovah’s Witnesses: extreme hemodilution oncardiopulmonary bypass. Pediatr Cardiol 1991; 12: 170–174.

84 Mossad E, Estafanous F. Blood use in cardiac surgery and thelimitations of hemodilution. Curr Opin Cardiol 1995; 10: 584–590.

85 Fukumura F, Sese A, Ueno Y et al. Haemostatic profile ofsmall children during and following cardiopulmonarybypass. Jpn J Thorac Cardiovasc Surg 2003; 51: 577–581.

86 Nakamura Y, Masuda M, Toshima Y et al. Comparative studyof cell saver and ultrafiltration nontransfusion in cardiacsurgery. Ann Thorac Surg 1990; 49: 973–978.

87 Dittrich S, Aktuerk D, Seitz S et al. Effects of ultrafiltrationand peritoneal dialysis on proinflammatory cytokines duringcardiopulmonary bypass surgery in newborns and infants.Eur J Cardiothorac Surg 2004; 25: 935–940.

88 Ootaki Y, Yamaguchi M, Yoshimura N et al. Efficacy of acriterion-driven transfusion protocol in patients havingpediatric cardiac surgery. J Thorac Cardiovasc Surg 2004; 127:953–958.

89 Mossinger H, Dietrich W, Braun SL et al. High-dose aprotininreduces activation of hemostasis, allogeneic blood require-ment, and duration of postoperative ventilation in pediatriccardiac surgery. Ann Thorac Surg 2003; 75: 430–437.

90 Pouard P. Review of efficacy parameters. Ann Thorac Surg1998; 65 Suppl. 6: S40–S43.

91 Huang H, Ding W, Su Z et al. Mechanism of the preservingeffect of aprotinin on platelet function and its use in cardiacsurgery. J Thorac Cardiovasc Surg 1993; 106: 11–18.

92 Coniff RF. The Bayer 022 compassionate-use pediatric study.Ann Thorac Surg 1998; 65 Suppl. 6: S31–S33.

93 Boldt J, Knothe C, Zickmann B et al. Comparison of twoaprotinin dosage regimens in pediatric patients havingcardiac operations. Influence on platelet function and bloodloss. J Thorac Cardiovasc Surg 1993; 105: 705–711.

94 Ranucci M, Corno A, Pavesi M et al. Renal effects of low doseaprotinin in pediatric cardiac surgery. Minerva Anestesiol 1994;60: 361–366.

95 Dobkowski WB, Murkin JM. A risk-benefit assessment ofaprotinin in cardiac surgical procedures. Drug Saf 1998; 18:21–41.

96 Cohen DM, Norberto J, Cartabuke R et al. Severe anaphylacticreaction after primary exposure to aprotinin. Ann Thorac Surg1999; 67: 837–838.

97 Carrel TP, Schwanda M, Vogt PR et al. Aprotinin in pediatriccardiac operations: a benefit in complex malformations and withhigh-dose regimen only. Ann Thorac Surg 1998; 66: 153–158.

98 D’Errico CC. Pharmacoeconomics analysis in a pediatricpopulation. Ann Thorac Surg 1998; 65 Suppl. 6: S52–S54.

99 Mossinger H, Dietrich W. Activation of hemostasis duringcardiopulmonary bypass and pediatric aprotinin dosage. AnnThorac Surg 1998; 65 Suppl. 6: S45–S50.

100 Dietrich W. Incidence of hypersensitivity reactions. AnnThorac Surg 1998; 65 Suppl. 6: S60–S64.

101 Dietrich W, Spath P, Ebell A et al. Prevalence of anaphylacticreactions to aprotinin: analysis of two hundred forty-eight



reexposures to aprotinin in heart operations. J Thorac Cardio-vasc Surg 1997; 113: 194–201.

102 Handler SD, Miller L, Richmond KH et al. Post-tonsillectomyhemorrhage: incidence, prevention and management.Laryngoscope 1986; 96: 1243–1247.

103 Crysdale WS, Russel D. Complications of tonsillectomy andadenoidectomy in 9409 children observed overnight. CMAJ1986; 135: 1139–1142.

104 Wei JL, Beatty CW, Gustafson RO. Evaluation of posttonsil-lectomy hemorrhage and risk factors. Otolaryngol Head NeckSurg 2000; 123: 229–235.

105 Postma DS, Folsom F. The case for an outpatient ‘‘approach’’for all pediatric tonsillectomies and/or adenoidectomies: a4-year review of 1419 cases at a community hospital. Otolar-yngol Head Neck Surg 2002; 127: 101–108.

106 Asaf T, Reuveni H, Yermiahu T et al. The need for routinepre-operative coagulation screening tests (prothrombin timePT/partial thromboplastin time PTT) for healthy childrenundergoing elective tonsillectomy and/or adenoidectomy. IntJ Pediatr Otorhinolaryngol 2001; 61: 217–222.

107 Zwack GC, Derkay CS. The utility of preoperative hemostaticassessment in adenotonsillectomy. Int J Pediatr Otorhinolar-yngol 1997; 39: 67–76.

108 Krishna P, Lee D. Post-tonsillectomy bleeding: a meta-analy-sis. Laryngoscope 2001; 111: 1358–1361.

109 Close HL, Kryzer TC, Nowlin JH et al. Hemostatic assessmentof patients before tonsillectomy: a prospective study. Otolar-yngol Head Neck Surg 1994; 111: 733–738.

110 Lippi G, Manzato F, Franchini M et al. Establishment ofreference values for the PFA-100 platelet function analyzer inpediatrics. Clin Exp Med 2001; 1: 69–70.

111 Fressinaud E, Veyradier A, Truchaud F et al. Screening forvon Willebrand disease with a new analyzer using high shearstress: a study of 60 cases. Blood 1998; 91: 1325–1331.

112 Derkay CS, Werner E, Plotnick E. Management of childrenwith von Willebrand disease undergoing adenotonsillectomy.Am J Otolaryngol 1996; 17: 172–177.

113 Judkins JH, Dray TG, Hubbell RN. Intraoperative ketorolacand posttonsillectomy bleeding. Arch Otolaryngol Head NeckSurg 1996; 122: 937–940.

114 Splinter WM, Rhine EJ, Roberts DW et al. Preoperativeketorolac increases bleeding after tonsillectomy in children.Can J Anaesth 1996; 43: 560–563.

115 Gunter JB, Varughese AM, Harrington JF et al. Recovery andcomplications after tonsillectomy in children: a comparison ofketorolac and morphine. Anesth Analg 1995; 81: 1136–1141.

116 Gallagher JE, Blauth J, Fornadley JA. Perioperative ketorolactromethamine and postoperative hemorrhage in cases of ton-sillectomy and adenoidectomy. Laryngoscope 1995; 105: 606–609.

117 Rusy LM, Houck CS, Sullivan LJ et al. A double-blind eval-uation of ketorolac tromethamine versus acetaminophen inpediatric tonsillectomy: analgesia and bleeding. Anesth Analg1995; 80: 226–229.

118 Romsing J, Ostergaard D, Walther-Larsen S et al. Analgesicefficacy and safety of preoperative versus postoperative ket-orolac in paediatric tonsillectomy. Acta Anaesthesiol Scand1998; 42: 770–775.

119 Agrawal A, Gerson CR, Seligman I et al. Postoperative hem-orrhage after tonsillectomy: use of ketorolac tromethamine.Otolaryngol Head Neck Surg 1999; 120: 335–339.

120 Kovesi T, Royston D. Pharmacological approaches to redu-cing allogeneic blood exposure. Vox Sang 2003; 84: 2–10.

121 Goobie SM, Soriano SG, Zurakowski D et al. Hemostaticchanges in pediatric neurosurgical patients as evaluated bythrombelastograph. Anesth Analg 2001; 93: 887–892.

122 Curreri PW, Hicks JE, Aronoff RJ et al. Inhibition of activesodium transport in erythrocytes from burned patients. SurgGynecol Obstet 1974; 139: 538–540.

123 Simon TL, Curreri PW, Harker LA. Kinetic characterization ofhemostasis in thermal injury. J Lab Clin Med 1977; 89: 702–711.

124 Cullen JJ, Murray DJ, Kealey GP. Changes in coagulationfactors in patients with burns during acute blood loss. J BurnCare Rehabil 1989; 10: 517–522.

125 Sheridan RL, Szyfelbein SK. Trends in blood conservation inburn care. Burns 2001; 27: 272–276.

126 Ikeda S, Schweiss JF. Maximum infusion rates and CVPaccuracy during high-flow delivery through multilumencatheters. Crit Care Med 1985; 13: 586–588.

127 Nissenson AR, Strobos J. Iron deficiency in patients with renalfailure. Kidney Int Suppl 1999; 69: S18–S21.

128 Montini G, Zacchello G, Baraldi E et al. Benefits and risks ofanemia correction with recombinant human erythropoietin inchildren maintained by hemodialysis. J Pediatr 1990; 117: 556–560.

129 Chavers BM, Sullivan EK, Tejani A et al. Pre-transplant bloodtransfusion and renal allograft outcome: a report of the NorthAmerican Pediatric Renal Transplant Cooperative Study.Pediatr Transplant 1997; 1: 22–28.

130 Butt ML, Shafi T, Farooqi I et al. Effect of dialysis on bleedingtime in chronic renal failure. J Pak Med Assoc 1998; 48: 242–244.

131 Remuzzi G, Livio M, Marchiaro G et al. Bleeding in renalfailure: altered platelet function in chronic uraemia only par-tially corrected by haemodialysis. Nephron 1978; 22: 347–353.

132 Mannucci PM, Remuzzi G, Pusineri F et al. Deamino-8-D-arginine vasopressin shortens the bleeding time in uremia.N Engl J Med 1983; 308: 8–12.

133 Bontempo FA, Lewis JH, Van Thiel DH et al. The relation ofpreoperative coagulation findings to diagnosis, blood usage,and survival in adult liver transplantation. Transplantation1985; 39: 532–536.

134 Carlier M, Van Obbergh LJ, Veyckemans F et al. Hemostasisin children undergoing liver transplantation. Semin ThrombHemost 1993; 19: 218–222.

135 Eckhoff DE, D’Alessandro AM, Knechtle SJ et al. 100 con-secutive liver transplants in infants and children: an 8-yearexperience. J Pediatr Surg 1994; 29: 1135–1139.

136 Kalayoglu M, D’Alessandro AM, Sollinger HW et al. Experi-ence with reduced-size liver transplantation. Surg GynecolObstet 1990; 171: 139–147.

137 Basford TJ, Poenaru D, Silva M. Comparison of delayedcomplications of central venous catheters placed surgically orradiologically in pediatric oncology patients. J Pediatr Surg2003; 38: 788–792.

138 Randolph AG, Cook DJ, Gonzales CA et al. Ultrasoundguidance for placement of central venous catheters: a meta-analysis of the literature. Crit Care Med 1996; 24: 2053–2058.

139 Goodnough LT, Brecher ME, Kanter MH et al. Transfusionmedicine. First of two parts–blood transfusion. N Engl J Med1999; 340: 438–447.

140 Goodnough LT, Brecher ME, Kanter MH et al. Transfusionmedicine. Second of two parts – blood conservation. N EnglJ Med 1999; 340: 525–533.

141 Mayer MN, de Montalembert M, Audat F et al. Autologousblood donation for elective surgery in children weighing 8–25 kg. Vox Sang 1996; 70: 224–228.



142 Copley LA, Richards BS, Safavi FZ et al. Hemodilutionas a method to reduce transfusion requirements inadolescent spine fusion surgery. Spine 1999; 24: 219–222.

143 Goodnough LT. Erythropoietin therapy versus red celltransfusion. Curr Opin Hematol 2001; 8: 405–410.

144 Rosencher N, Ozier Y. [Peri-operative use of EPO]. TransfusClin Biol 2003; 10: 159–164.

145 Tamir L, Fradin Z, Fridlander M et al. Recombinant humanerythropoietin reduces allogeneic blood transfusion require-ments in patients undergoing major orthopedic surgery.Haematologia (Budap) 2000; 30: 193–201.

146 Feagan BG, Wong CJ, Kirkley A et al. Erythropoietin with ironsupplementation to prevent allogeneic blood transfusion intotal hip joint arthroplasty. A randomized, controlled trial.Ann Intern Med 2000; 133: 845–854.

147 Horiba K, Itou Y, Terada H et al. Experience of predepositautologous blood transfusion and medication of recombinanthuman erythropoietin in pediatric open heart surgery. KyobuGeka 1991; 44: 1146–1150.

148 Ley SJ. Intraoperative and postoperative blood salvage.AACN Clin Issues 1996; 7: 238–248.

149 Napier JA, Bruce M, Chapman J et al. Guidelines for autolo-gous transfusion. II. Perioperative haemodilution and cellsalvage. British Committee for Standards in HaematologyBlood Transfusion Task Force. Autologous TransfusionWorking Party. Br J Anaesth 1997; 78: 768–771.

150 Elias D, Lapierre V, Billard V. Perioperative autotransfusionwith salvage blood in cancer surgery. Ann Fr Anesth Reanim2000; 19: 739–744.

Accepted 13 September 2004



transfusion 2

Documents