Strategies for Blood Product Management, Reducing Transfusions, and Massive Blood Transfusion

Red Blood Cell–Containing Components

Blood components containing RBCs are indicated for the treatment of symptomatic deficits of oxygen-carrying capacity. PRBCs are the most widely available RBC-containing blood component, although in settings where the collection facilities do not have the capability to make components, whole blood may be the only component available. Donor whole blood is collected in a preservative-anticoagulant solution that contains citrate, phosphate, dextrose (glucose), and adenine (CPDA) or just citrate, phosphate, and dextrose (CPD). In the latter case, the platelet-rich plasma is removed after centrifugation of the whole blood unit, and a solution containing adenine, dextrose, and occasionally mannitol is added to the PRBCs. The additive-solution systems permit storage for 42 days (compared with 35 days for CPDA) and better preservation of 2,3-diphosphoglycerate (DPG) levels. The characteristics of the CPDA and additive-solution PRBCs and of whole blood at the time of outdate are shown in Table 12.4 ; the hematocrit is reduced in the additive-solution PRBCs and the total volume is increased, but the red cell mass remains the same.

RBCs carry glycoconjugate antigens of the ABH histo-blood group system on the cell surface that are determined by three common alleles at the ABO locus on chromosome 9. During the first year of life, infants begin to elaborate alloantibodies to whichever A or B antigens they lack. These isoagglutinins are invariably present after a few months and constitute a formidable immunologic obstacle to transfusion or transplantation across this ABO barrier. The RBCs for transfusion must be compatible with the ABO isoagglutinins of the intended transfusion recipient. Similarly, components with a large volume of plasma (e.g., whole blood, FFP, apheresis platelets) must be compatible with the A or B surface antigens expressed on the recipient's RBCs. PRBCs must be ABO compatible with the recipient, whereas whole blood must be ABO identical because of the larger volume of donor plasma and, hence, AB isoagglutinins . Table 12.5 summarizes the permissible combinations.

Only RBCs express the Rh(D) antigen. Rh(D)-positive patients may receive Rh(D)-positive or Rh(D)-negative RBCs. Rh(D)-negative patients are routinely given Rh(D)-negative RBCs for any elective transfusions, but in the setting of massive transfusion it may be necessary to switch to Rh(D)-positive RBCs to preserve the supply of Rh(D)-negative RBCs. The blood bank usually determines when to make this substitution based on the inventory and does so more quickly for a patient who is a male or a postmenopausal female ( Table 12.6 ). The objective is to avoid exposing a female with childbearing potential to Rh(D)-positive RBCs and possibly triggering the production of the anti-D alloantibody, which is responsible for the most severe forms of hemolytic disease of the newborn. Table 12.7 shows the common initial volume of PRBCs needed to increase the hemoglobin level by 2 to 3 g/dL.

The changes that occur to RBCs during storage under conventional blood bank conditions have been well described. These observations have generated physiologically plausible hypotheses about how such changes may impair the function of the banked RBCs in vivo. The reduced hemoglobin level of 2,3-DPG and its corresponding decrease in the P ₅₀ value may reduce the ability of stored RBCs to relinquish bound O ₂ compared with 2,3-DPG–replete RBCs. The depletion of nitric oxide (NO) may reduce the vasodilatory properties of the RBCs, hence impairing their ability to maintain the patency of the small vessels in the microcirculation and blood flow to the tissues. Numerous changes in the composition and behavior of the RBC plasma membrane, including the loss and oxidation of membrane lipids and proteins and the rearrangement of some membrane constituents, correlate with changes in the shape and elasticity of the RBC membrane. The loss of elasticity in particular can impede the rapid movement of the RBCs through the microcirculation.

These hypotheses and some supportive data from animal models have led to a number of observational clinical studies (mostly in trauma, critical care, colorectal surgery, and cardiac surgery) of outcomes after using stored PRBCs, but the results have been inconclusive. One prospective, observational study from 30 North American Centers in 296 children younger than 18 years of age who received blood stored 14 days or longer reported increased multiple organ dysfunction (OR 1.87) and increased pediatric intensive care unit (PICU) stay (~3.7 days) but no difference in mortality. A small study of pediatric cardiac surgical patients found that children who received blood older than 3 days required additional RBC and FFP transfusions but this study was underpowered. About half of such observational studies found a statistical association between an unfavorable clinical outcome measure and the transfusion of RBCs that had been stored for a greater time. However, no such association was reported in the other half of the studies, including two that were extensions of previous studies with positive findings. A small number of randomized, controlled trials (RCTs) addressed this issue without statistically significant differences in outcomes between patients receiving RBCs stored for different amounts of time, although two of them were underpowered. Knowing that RBCs change during storage raises the question of whether these changes affect children in a clinically meaningful way, a question that remains unanswered by the observational studies.

In the past few years, several RCTs have addressed these issues in four different patient populations, two of which were in pediatric cohorts. A multicenter RCT conducted in Canada randomized low–birth-weight neonates in ICUs to receive PRBCs stored 8 days or less or the standard of care, which was to provide aliquots from one unit of PRBCs to each infant until transfusion was no longer required or the donor unit was depleted. There were no differences in the incidence of infections, bronchopulmonary dysplasia, necrotizing enterocolitis, death, or the composite between the two groups. Children in Uganda between the ages of 6 months and 6 years who presented with severe anemia (hemoglobin <5 g/dL) and lactic acidosis (lactate >5 mmol/L) were randomly assigned to receive PRBCs stored 10 days or less or 35 days longer. There were no differences in lactate clearance, left ventricular strain, as assessed by β-naturietic peptide or, in a subset of patients, correction of cerebral tissue oxygen saturation as measured by near-infrared spectroscopy. Two studies in adults compared clinical outcomes after transfusion with PRBCs stored for different periods of time. Patients 12 years of age or older undergoing complex cardiac surgery and very likely to require PRBC transfusion were randomly assigned to receive PRBCs stored 10 days or longer or 28 days or less. No differences were observed in the change in the Multiple Organ Dysfunction Score (MODS) or mortality at 7 or 28 days, in length of ICU or hospital length of stay, or in the frequency of serious adverse events. Adult patients in ICUs in Canada were randomly assigned to receive PRBCs stored 8 days or less versus the standard of care, which was to issue the oldest units first. No differences were found for 30- or 90-day mortality, ICU or hospital length of stay, changes in MODS, or several other clinical endpoints. There is no apparent clinical benefit derived by transfusing units of PRBCs, which have been stored for a period of time that is substantially shorter than that of the PRBCs routinely supplied by our current inventory practices for these vulnerable populations.

Platelets

Platelets may be obtained from a unit of whole blood or collected by apheresis. Platelets from whole blood are separated by centrifugation and suspended in 40 to 60 mL of plasma at a concentration that is two to four times greater than in the circulation. Each unit contains a minimum of 5.5 × 10 ¹⁰ platelets and is stored at 20°C to 24°C with gentle continuous agitation for a maximum of 5 days. One unit of whole blood−derived platelets can be expected to increase the platelet count in an 18-kg child by 15,000/mm ³ and in a 70-kg adult by 5000 to 10,000/mm ³ . A unit of platelets obtained by apheresis contains at least 3 × 10 ¹¹ platelets in 200 to 400 mL of plasma, or the equivalent of approximately 6 units of whole blood−derived platelets. A common dose for children is 0.1 to 0.3 unit/kg of body weight, or 10 to 15 mL/kg (see Table 12.7 ); this dose usually produces an increment of 30,000 to 90,000/mm ³ . However, in several prophylactic platelet dose trials for medical causes of bleeding, doses equivalent to the standard dose of 1 pheresis unit (or 6 units of whole blood−derived platelet concentrates) per meter squared, half of this dose and double this dose were compared in 1272 adult and children who received at least 1 platelet transfusion. Blood losses were determined with the World Health Organization (WHO) bleeding scale: grade 0 = no bleeding, grade 1 = petechiae, grade 2 = mild blood loss, grade 3 = gross blood loss, and grade 4 = debilitating blood loss. No differences were observed in bleeding outcomes (WHO grade ≥2) among the three doses. The subset of 200 children were found to have a greater risk for bleeding than adults for unknown reasons, but bleeding in this age group also did not differ by platelet dose received and was seemingly unrelated to platelet counts. The recommended platelet dose may be reduced in the near future, but this change awaits further discussion by the blood transfusion and hemostasis community. This trial assessed the use of platelets to prevent bleeding events and did not address patients undergoing surgical procedures with ongoing bleeding.

In the setting of dilutional thrombocytopenia with ongoing blood loss or a consumptive coagulopathy (e.g., disseminated intravascular coagulation), larger doses (≥0.3 unit/kg) may be required to boost the platelet count above 50,000/mm ³ . Because platelets are suspended in plasma that contains the anti-A and anti-B isoagglutinins, they should be ABO compatible with the recipient's RBCs. Some blood donors have high-titer isoagglutinins that can produce hemolysis in transfusion recipients if a large enough volume of plasma is given. The transfusion of plasma-incompatible, whole blood–derived platelets to adult recipients does not produce clinically significant hemolysis because the volume of plasma given is so small relative to the plasma volume of an adult. However, apheresis platelets (and whole blood–derived platelets for small children) should be ABO compatible with the recipient's RBCs. Matching for Rh(D) antigen is not necessary for apheresis platelets because platelets do not express Rh antigens and they contain virtually no RBCs. However, whole blood–derived platelets may contain enough RBCs to provoke Rh alloimmunization, so platelets from Rh(D)-negative donors are given preferentially to Rh(D)-negative recipients with childbearing potential. If a premenopausal female receives whole blood–derived platelets from an Rh(D)-positive donor, Rh immune globulin (Rhogam) can be administered within 72 hours to prevent alloimmunization. Platelets should never be withheld in an emergency situation because of Rh(D) incompatibility.

Platelets are essential to hemostasis associated with the vascular injury of surgery and are necessary for the control of surgical bleeding. Platelets are also required for the maintenance of an intact endothelial barrier to spontaneous blood loss. The number of platelets required to provide adequate hemostasis in the surgical setting is much greater than the number needed to provide prophylaxis against spontaneous hemorrhage. A platelet count of 40,000 to 50,000 /mm ³ is considered adequate to prevent spontaneous bleeding or bleeding from minor invasive procedures (e.g., lumbar puncture, line placement) in an otherwise stable child. If overt signs of bleeding are present or a more significant hemostatic challenge in the form of a surgical procedure is imminent, sustaining a level of 30,000 to 50,000/mm ³ for several days may be required. A target level of 50,000/mm ³ is appropriate in the setting of massive transfusion. Platelets may also be required for children with adequate counts but in whom platelet function is impaired in some forms of congenital heart disease and following cardiopulmonary bypass. Many medications (e.g., aspirin; other nonsteroidal antiinflammatory agents, including ibuprofen and naproxen; dipyridamole; platelet P2Y12 receptor blockers such as clopidogrel or prasugrel; or glycoprotein IIa/IIIb receptor inhibitors such as abciximab, eptifibatide, or tirofiban; serotonin uptake antagonists such as Zoloft) and some medical conditions (e.g., renal failure with blood urea nitrogen levels above 60 mg/dL) cause abnormal platelet function, which may interfere with surgical hemostasis, in which case it may be necessary to maintain the platelet count at a somewhat greater concentration, at least until the effect of the medication dissipates or the child's platelets have largely been replaced by banked platelets. In a few settings, such as intracranial, ophthalmic, and otologic surgery, even greater concentrations (100,000/mm ³ ) may be sought, although the appropriate threshold in these settings has not been examined in detail.

There is no clear-cut threshold value below which the platelet count predicts clinical bleeding in the perioperative period. Each child must be individually assessed by constantly observing the surgical field for evidence of abnormal bleeding. Unfortunately, we lack a well-validated bedside tool to assess platelet function. The utility of the thromboelastogram and other devices to measure platelet function under controlled flow conditions, such as microfluidic flow devices and the platelet function analyzer (PFA-100), are currently under investigation but are not at this time able to predict risk for hemorrhage. The standard technique for diagnosis and evaluation of thrombocytopathies remains Born-O'Brien platelet aggregometry, but it is not useful in the intraoperative or intensive care setting. Most commonly, dilutional thrombocytopenia rather than a newly acquired platelet function defect is the cause in the operative setting and in massive transfusion.

A child occasionally presents for surgery with a previously characterized platelet dysfunction that may be associated with bleeding. If the child has a normal platelet count, it is reasonable practice to ensure that the blood bank has an adequate platelet supply available for the operating room and to withhold transfusion of platelets until pathologic bleeding occurs.

Several additional points should be considered :

• 1.

  Not all hospitals have platelets in the inventory. Unless the need is anticipated before surgery, platelets may not be available when they are required. Therefore, a preemptive request for platelets may have to be arranged.

• 2.

  For children who are thrombocytopenic before surgery, platelets should be infused just before the surgical procedure to ensure the greatest concentrations during the time of peak demand. The start of the procedure should not be delayed to obtain the results of a posttransfusion platelet count.

• 3.

  Platelets should be filtered only by large-pore filters (≥150 µm) or leukocyte-reduction filters (if indicated). Micropore filters may adsorb large numbers of platelets, thereby diminishing the effectiveness of the platelet transfusion.

• 4.

  Platelets are suspended in plasma, which may help to replenish coagulation factors other than factors V and VIII, which are labile, and factor VII, which has an especially brief half-life.

• 5.

  Platelets should not be refrigerated or placed in a cooler with ice before administration, because cold-exposed platelets are rapidly cleared from the circulation.

Special Processing of Cellular Blood Components

Leukocytes collected with whole blood donations partition into both the platelet and the PRBC components; few intact leukocytes are present in FFP. Passenger leukocytes are responsible for most febrile, nonhemolytic transfusion reactions, human leukocyte antigen (HLA) alloimmunization, and transmission of cytomegalovirus (CMV). To prevent the complications from these leukocytes, blood components should be passed through leukocyte-reduction filters that effectively remove leukocytes (by a 2–3 log reduction) immediately after collection (prestorage leukoreduction) or at the bedside (pretransfusion). This technique is superior to washing or freezing-deglycerolizing, which was used in the past. Table 12.8 lists those patients who may benefit from leukocyte-reduced cellular components (i.e., PRBCs or platelets). It should be noted that anaphylaxis may very rarely occur when using white-cell reduction filters.

CMV transmission can also be reduced by screening donors for CMV exposure (testing for antibody to CMV), although leukocyte reduction is the more widely used approach. Even though primary CMV infection is benign in children with intact immune systems, some pediatric populations are at risk for developing systemic disease and should be protected from blood-borne CMV transmission. Only patients who have not previously been infected with CMV (i.e., CMV-seronegative) are at risk. Children who are particularly vulnerable to systemic CMV infections are listed in E-Table 12.1 .

Transfused lymphocytes may mediate a graft-versus-host process in some recipients with impaired cellular immunity. Because this process involves the bone marrow as well as the usual targets (i.e., skin and gastrointestinal tract), the fatality rate is substantial. Transfusion-associated graft-versus-host disease (TA-GVHD) can be prevented by exposing cellular blood components to gamma irradiation that disables the donor lymphocytes. Children who are considered to be at risk for TA-GVHD and who should receive irradiated cellular components are listed in E-Table 12.2 . This complication can also occur in children with intact immune systems in the unusual circumstance when the transfusion donor is homozygous for an HLA haplotype that is shared with the recipient. In this case, the recipient's immune system, although fully functional, cannot recognize the donor lymphocytes as foreign. The donor lymphocytes mount a GVHD attack on the recipient's tissues, recognizing the mismatched haplotype. This situation is more likely to occur when the donor is a blood relative of the recipient. It is for this reason that blood and HLA-matched platelets donated by family members are routinely irradiated.

Because irradiation damages the cell membrane, irradiated RBCs lose K ⁺ at a greater rate than usual. As a result, the shelf life of irradiated RBC is only 28 days (from a maximum permitted of 42 days). The problem of increased amounts of circulating K ⁺ can be obviated by irradiating units close to the time of issue or washing the unit if it was irradiated early in the storage period.

Fresh Frozen Plasma

FFP represents the fluid portion of whole blood that is separated and frozen within 8 hours of collection. After thawing at 37°C, which usually requires 30 minutes, it may be administered within 24 hours if stored at 1°C to 6°C. The volume of 1 unit varies from 180 to 300 mL and represents 7% to 10% of the coagulation factor activity in a 70-kg patient. It contains all of the clotting factors and regulatory proteins at approximately the native concentration, but after 6 hours at 1°C to 6°C, the concentrations of the labile factors V and VIII begin to diminish, as does that of the short-lived factor VII. FFP does not provide functional platelets, nor does it contain leukocytes or RBCs. Thawed FFP may be used for transfusion up to 7 days after thawing; however, it must be labeled as thawed plasma to indicate that it has reduced levels of factors V, VII, VIII, and protein S.

FFP should be ABO compatible with recipient red cells because it contains the anti-A and anti-B isoagglutinins appropriate to the donor's ABO blood group. If the recipient's blood type is not known, plasma from a donor with blood type AB, which contains neither anti-A nor anti-B, may be administered. Because the citrate anticoagulant is present in the plasma, rapid administration of FFP is more likely to be associated with citrate toxicity than the transfusion of components with smaller volumes of plasma (e.g., PRBCs).

FFP is all too often administered without justification by evidence-based medicine. One major surgical indication for FFP is to correct the coagulopathy associated with massive blood transfusion (see Table 12.7 ). Other indications include correction of a prolongation in the prothrombin time (PT) before surgery or in the setting of bleeding, the emergency reversal of warfarin, or the presence of a specific congenital or acquired coagulation protein deficiency for which a factor concentrate or a recombinant factor is not available (e.g., for factor X deficiency). Administration of vitamin K should not be overlooked in children with hepatic insufficiency, who have been exclusively breastfed, who have been treated with warfarin, broad-spectrum antibiotics (which often eliminate normal vitamin K–producing gastrointestinal flora), or total parenteral nutrition for inadequate oral caloric intake, or who have had a prolonged hospitalization. Correction of a mild increase in the PT (e.g., international normalized ratio [INR] <1.5) is rarely necessary. Fig. 12.1 shows the relationship between the level of coagulation factors and the in vitro clotting time—in this case, the PT. Relatively modest levels of coagulation factors can support normal hemostasis, even though the PT is prolonged. When the PT is markedly increased (see Fig. 12.1 , point A), the transfusion of 1 unit of FFP, which increases the coagulation factor levels by 7% to 10% in an adult, dramatically decreases the PT. When the PT is only mildly prolonged, as at point C, where factor levels are already adequate for hemostasis, infusion of 1 unit of FFP (in an adult) decreases the PT to a much smaller extent. This small decrease in the PT does not clinically improve hemostasis.

Cryoprecipitate

Cryoprecipitate is prepared by thawing FFP at 4°C to 10°C and removing most of the plasma, leaving behind precipitated protein that is then resuspended in a small volume of residual plasma (15–25 mL) and refrozen. This component contains 20% to 50% of the factor VIII from the original unit of plasma. It also contains von Willebrand factor (vWF), fibrinogen (approximately 250 mg), and factor XIII. Cryoprecipitate is indicated for the treatment of factor XIII deficiency, dysfibrinogenemia, and hypofibrinogenemia (see Table 12.7 ). It has not been used for the treatment of von Willebrand disease or hemophilia A since the advent of clotting factor concentrates and recombinant factor VIII. Plasma concentrates of factor XIII and fibrinogen have been licensed in the United States for use in patients with these congenital deficiencies.

Pathogen Inactivation/Reduction

Techniques to inactivate or reduce the level of infectious agents have been in place for plasma derivatives such as intravenous (IV) immunoglobulin and plasma-derived clotting factors for over 20 years. More recently, pathogen inactivation technologies have been applied to RBCs, FFP, and platelets with the goal of eliminating a wide range of infectious organisms. There are several potential advantages of these technologies over screening by donor history and testing for specific organisms, particularly for those pathogens that frequently cause asymptomatic infection, have a long serologic “window” period before screening tests become positive, are newly emerging, or are completely new and unrecognized.

There are two general approaches to pathogen inactivation: methods that disrupt lipid membranes (solvent detergent treatment and methylene blue plus visible light exposure ) and methods that target RNA and DNA (or nucleic acid techniques) [amotosalen or riboflavin plus ultraviolet (UV) light exposure, or UV light exposure alone]. The nucleic acid–targeted techniques also inactivate leukocytes and eliminate the risk of transfusion-transmitted graft versus host disease, which may make gamma irradiation of cellular components for susceptible patients unnecessary.

However, these techniques do have limitations. Even though inactivation techniques can achieve a 5 or 6 log reduction in infectious particles, they may not be completely effective in components with high pathogen loads. In addition, the lipid-targeted techniques do not inactivate nonlipid enveloped viruses such as hepatitis A virus, hepatitis E virus, and parvovirus B19, and none of the techniques inactivates prions. These procedures also damage or deplete plasma proteins and platelets, which may reduce their effectiveness. Although there is no evidence to date of major adverse effects, there is the potential for toxicity such as the formation of plasma or membrane protein neoantigens, or from long-term effects of exposure to amotosalen. Finally, these treatments are relatively complex and expensive.

In the United States, several pathogen-inactivated component systems have been licensed by the Food and Drug Administration (FDA): plasma and platelets treated with amatosalen and UV-A light; plasma treated with riboflavin and UV light ; and solvent/detergent-treated plasma. There are currently no licensed systems for pathogen inactivation of RBCs, although several are under investigation.

Plasma-Derived Factor Concentrates and Recombinant Factors

The most commonly administered factor concentrate is factor VIII, used in the treatment of hemophilia A. Children with hemophilia can have many problems related to their disease, including splenomegaly, abnormal liver function, and joint disease related to hemarthrosis. In the past, the use of pooled plasma products was associated with very high rates of transmission of viral hepatitis (especially HCV) and HIV. The use of more rigorous viral removal and inactivation processes and the introduction of recombinant factor VIII and IX products have greatly reduced these problems. Initial concerns that there may be an increased incidence of inhibitors in children who receive recombinant therapy compared with plasma-derived factor therapy have been confirmed by the SIPPET trial (Survey of Inhibitors in Plasma-Product Exposed Toddlers), which showed that toddlers treated with plasma-derived factor VIII containing VWF had a nearly twofold lower incidence of inhibitors than those treated with recombinant factor VIII. However, the study failed to show a significant difference with regard to high-titer inhibitors and did not include any of the newer concentrates that have appeared since the trial started in 2010. Therefore, final conclusions regarding the use of recombinant factors remain unclear. Mild hemophilia A usually responds well to desmopressin (1-deamino-8- d -arginine vasopressin [DDAVP]) therapy.

New extended half-life factor concentrates are now available for both hemophilia A and B. For factor VIII and factor IX concentrates, these include fusion to either albumin or the monomeric Fc fragment of immunoglobin G1 (IgG1). Conjugation with polyethylene glycol (glycosylation) is another modification technique. The optimal use of these concentrates is under current investigation.

Von Willebrand disease is routinely treated with DDAVP or plasma-derived factor VIII concentrates that are also rich in vWF, such as Humate-P, Alphanate, Koate DVI, and Wilate, with dosing in ristocetin cofactor units per kilogram, not factor VIII units. A newly available alternative is a recombinant VWF, although this product is so far only approved for those 18 years or older. In children who have von Willebrand disease and are resistant to DDAVP or for whom DDAVP is contraindicated (e.g., central nervous system [CNS] bleeding, allergic reaction, brain tumor, recent CNS surgery), it is reasonable to withhold treatment with blood-derived products until surgery has begun unless surgery is performed in an area where even minor bleeding can produce serious complications. These children often do not demonstrate pathologic bleeding. Adjunctive therapies that can further limit hemorrhage include use of the antifibrinolytic agent ε-aminocaproic acid (Amicar) and tranexamic acid (Cyklokapron), a competitive inhibitor of plasminogen, both being administered orally or intravenously, and topical hemostatic agents, including topical collagen and fibrin glues.

Children with hemophilia B (i.e., Christmas disease or factor IX deficiency) are managed with recombinant human factor IX and highly purified factor IX (preparations with various amounts of factors VII, X, and prothrombin) that are treated to inactivate or remove viruses. Careful planning of any surgical procedure for these children includes close communication with the child's hematologist to ensure optimal therapy while reducing unnecessary transfusions (see Chapter 10 ).

Prothrombin Complex Concentrates

Prothrombin complex concentrates (PCC) have been used to rapidly reverse vitamin K–based anticoagulants in the setting of significant hemorrhage, especially in the CNS. They consist of either three-factor (II, low VII , IX, and X, proteins C and S) or four-factor (II, high VII , IX, and X, proteins C and S) human plasma-derived products. These have been primarily used in adults; pediatric experience is limited. The advantages appear to be a more rapid reversal than the administration of FFP or vitamin K (without, however, any difference in clinical outcomes) and for some patients, reduced volume of administration. One systematic review concluded that four-factor PCC more reliably corrected the INR than three-factor PCC, whereas another suggested that protocols based on body weight offer an advantage over individual physician decisions. However, in the setting of intracranial bleeding in patients taking warfarin, four-factor PCC has not proven superior to FFP. The efficacy of PCCs in the operating room setting to treat perioperative coagulopathy is unclear.

Desmopressin

DDAVP, a synthetic analog of vasopressin, can increase the levels of factor VIII:C (i.e., coagulant activity) and factor VIII:vWF in children with mild hemophilia A or von Willebrand disease. An IV dose of 0.3 µg/kg (maximum 20 µg; a subcutaneous preparation is available in Europe) increases the levels of both factors twofold to threefold within 30 to 60 minutes, with a half-life of 3 to 6 hours. Intranasal DDAVP is also effective, but onset is less rapid and, in younger children for whom a sustained inhalation may be more difficult, part of the dose may find its way into the gastrointestinal tract, bypassing nasal blood vessels. Between 80% and 90% of children with von Willebrand disease are responders, and affected children should be tested for their responsiveness to IV DDAVP. This treatment is best suited to treat bleeding from surgical procedures, which ceases within 2 to 3 days. When bleeding continues beyond this period, as is the case with some orthopedic procedures, daily IV Humate P (or Alphanate or Koate DVI) can obviate possible tachyphylaxis with DDAVP. Products rich in the vWF allow better control over peak concentrations of factor VIII. When in excess of 200%, factor VIII predisposes to postoperative deep venous thrombosis and pulmonary embolism.

DDAVP has been used to treat the coagulopathy associated with platelet dysfunction, uremia, and cirrhosis. It may reduce elective surgical bleeding when the potential for blood loss is substantial, such as in cardiac surgery and spinal fusion. Although initial reports apparently demonstrated a benefit in patients who did not have a preexisting coagulopathy, other controlled studies failed to show an effect despite increases in factor VIII:C and vWF, and its use for these indications has largely been abandoned. Because of the potential for hyponatremia from water retention, use of DDAVP is avoided in children younger than 2 years, in children with CNS lesions, including a brain tumor, history of CNS irradiation, or recent neurosurgery or CNS trauma.

Albumin, Dextrans, Starches, and Gelatins

Solutions of several high–molecular-weight molecules (i.e., colloids) have been used for volume replacement, although systematic reviews have determined they offer no advantage over crystalloid solutions. These colloids include albumin, dextrans, starches, and gelatins.

Albumin has the longest track record and the fewest adverse effects. In the past, dextrans (i.e., high– and low–molecular-weight glucose polymers) were administered for volume expansion and hemodilution in children, but currently their primary use is for antithrombosis, although their value for even this indication is questionable.

Starches are branched polysaccharide polymers available in high–, medium–, and low–molecular-weight ranges (480,000-70,000 Da). Although they expand blood volume, they also alter hemostasis by diluting clotting factors and impairing platelet function and the coagulation cascade. In addition, starches accumulate in the reticuloendothelial system and carry the potential for unknown long-term adverse effects. Minor coagulation changes have been reported when the dose exceeded 20 mL/kg. A 6% hydroxyethyl starch (HES 130/0.4) yielded clinical and physiologic profiles similar to those for 5% albumin in volumes up to 16 mL/kg in noncardiac surgery and in volumes up to 50 mL/kg in cardiac surgery, although at smaller cost. A meta-analysis of randomized controlled trials of hydroxyethyl starch concluded that mortality, creatinine, and blood loss did not increase, although platelet counts and the duration of ICU stay significantly decreased. The authors recommend against their use in pediatric patients. Several major reviews regarding the use of starches and gels in adult patients who are critically ill have raised substantive concerns regarding adverse effects on coagulation and renal function, and they found inadequate overall safety data, even for the third-generation products. If these concerns have been raised in adult populations, we should hold even greater concern about their use in children.

Gelatins are polypeptides derived from bovine collagen that seem to have a minimal effect on coagulation and provide reasonable plasma volume expansion. However, life-threatening anaphylactic or anaphylactoid reactions have been reported, and their use in children remains somewhat limited. A systematic review identified a lack of safety and efficacy data in neonates and children.

Red Blood Cell Substitutes

Blood substitutes offer the promise of agents with universal compatibility, minimal infectious risks, and prolonged shelf life (years rather than days) to carry oxygen to vital organs. Early efforts to develop these products involved human, bovine, and genetically engineered hemoglobin polymer solutions, perfluorocarbons, and lipid-encapsulated hemoglobin. Most failed in clinical trials because of severe complications such as renal failure, stroke, and vasoconstriction. The last of the hemoglobin-based oxygen carriers under investigation, Hemospan (Sangart, Inc., San Diego, CA), failed in clinical trials and is no longer being investigated. Liposome-encapsulated hemoglobin is currently under investigation as a means to carry oxygen to compromised tissues such as cerebral infarction. At this juncture, none of these attempts at investigation of blood substitutes have reached the pediatric population.

Coagulopathy

The coagulation system involves platelets, coagulation proteins, and localized tissue factor, which initiate all steps of hemostasis. Fig. 12.2 shows that an initial step is platelet adhesion to a wound or site of vessel wall injury, with adhesion being mediated by the vWF through its receptor on the platelet, the glycoprotein Ib–glycoprotein IX complex (GPIb-IX), and fibrinogen through the fibrinogen receptor on the platelet, the glycoprotein IIb–glycoprotein IIIa complex (GPIIb-IIIA). In flowing blood, initial platelet attachment is facilitated by vWF, whereas platelet spreading and more secure (shear stress-resistant) platelet-platelet aggregation is driven by fibrinogen and by the GPIIa-GPIIIa complex. However, there is also evidence that platelets attach even to intact endothelium, which has an activated phenotype, as after inflammatory cytokine exposure or sepsis. Platelets attach to the endothelium through high-molecular-weight von Willebrand multimers. Fibrinogen is attached to endothelium through upregulated integrins and selectins.

Initial platelet hemostasis (i.e., platelet plug formation) is accompanied by the local generation of fibrin, which is the end product of at least three surface-active enzyme complexes. Clotting is initiated by the tissue factor/factor VIIa surface-active enzyme complex and amplified by the factor VIIIa/IXa/X and factor II/Va/Xa complexes. A mural platelet thrombus, which includes platelets and fibrin, then forms a scaffold on which healing of the vessel wall can take place. The fibrin component of a platelet thrombus forms beneath, not above aggregating platelets, as was previously thought. The scaffold is removed when it is no longer needed by thrombolysis and the effects of macrophages.

The surface-active enzyme complexes ( Fig. 12.3 ) are active on the phospholipid surfaces provided by platelets, leukocytes, and endothelial cells but not in the bulk of the blood. Initially it was believed that formation of the tissue factor/factor VIIa complex, as shown, was critical for the action of recombinant FVIIa (rFVIIa) and for activation of factors X and IX. However, the high doses of rFVIIa necessary to achieve clinical hemostasis are far in excess of those required to saturate available tissue factor, suggesting that rFVIIa must also operate in large part independent of tissue factor, especially in view of the observation that rFVIIa can bind to platelets directly via platelet anionic phospholipid or via platelet GPIb. In this regard, the conventional coagulation cascade shown in Fig. 12.4 is oversimplified, although it is a convenient approach to understanding the PT and partial thromboplastin time (PTT). All of the steps must be considered in the milieu of flowing blood, such that the high-velocity gradients of arterioles (i.e., mucous membranes of the uterus, gastrointestinal tract, upper respiratory tract, oral cavity, and gums) and arteries favor thrombi with a greater proportion of platelets (i.e., white thrombi), and low-velocity gradient states, such as those found in stasis or blood accumulation within a body cavity, favor a greater proportion of red cells (i.e., red thrombi). This explains why a patient with von Willebrand disease, characterized by a defect in the protein that allows blood platelets to adhere to a wound, tends to bleed from mucous membranes, sites of high-velocity (arteriolar) gradients, whereas hemophiliacs tend to bleed into joint spaces and muscle planes, sites of low- or near-zero–velocity gradients.

The coagulopathy associated with massive blood transfusions is usually attributable to the dilution of clotting factors or platelets, or both. The point at which the deficiency in clotting factors is sufficient to produce a coagulopathy depends on the volume of blood lost and the type of blood component transfused (i.e., PRBCs or whole blood). Dilutional thrombocytopenia sufficient to cause clinical bleeding depends on the starting platelet count and the volume of blood replaced ( Fig. 12.5 ). In some cases, the cause of bleeding is a consumptive coagulopathy such as fibrinolysis or disseminated intravascular coagulation (DIC). In other scenarios, bleeding is caused by hypothermia, severe metabolic acidosis, poor tissue perfusion, and the release of tissue factors. Body temperature should be maintained by using efficient blood-warming devices, acidosis should be treated, and normovolemia and cardiac output should be restored to prevent a coagulopathy from developing.

Dilutional Thrombocytopenia

In an effort to formulate a plan to manage children who require massive blood transfusions, we must rely on our clinical experience and data extrapolated from adults and limited pediatric studies. A study of adult trauma patients during the Vietnam War reported that the onset of clinical bleeding occurred after about 15 units of whole blood or 1.5 blood volumes had been transfused. The incidence of coagulopathy was unrelated to an abnormal PT or PTT but correlated closely with a platelet count of less than 65,000/mm ³ . Studies of massive blood loss with whole blood replacement also support the conclusion that the coagulopathy at these levels of blood loss results from thrombocytopenia rather than a clotting factor deficiency. Fig. 12.6 compares the calculated reduction in platelet count with the observed decline in platelet count in adults and children; when normalized for blood volumes shed, the observed changes were nearly identical. The observed and calculated decrements differ because platelets are mobilized from the bone marrow, spleen, lungs, and lymphatic tissues. The platelet count usually does not decrease to concentrations that may cause bleeding in children until 2.0 to 2.5 blood volumes have been shed or until 20 to 25 units of whole blood are transfused in adults. Clinical bleeding does not usually occur in children whose platelet counts exceed 50,000/mm ³ , despite blood losses as great as 5.0 blood volumes ( Fig. 12.7A ). Consequently, children should be monitored for thrombocytopenia and possible transfusion of platelets or clotting factor deficiency (see further) after the loss of the first 1.0 to 1.5 blood volumes. After the platelet count has decreased to 50,000/mm ³ , it is likely that approximately one platelet dose (i.e., 6 units for an adult or 10-15 mL/kg for a child) will be required for each blood volume replaced. If a coagulopathy develops earlier than expected (i.e., before a 1.0-blood volume loss), a search should be initiated for other causes of bleeding, such as increased arterial or venous pressure in the surgical field or DIC.

The starting platelet count is important for estimating how much blood loss can be tolerated before critical thrombocytopenia occurs. For example, with a starting platelet count of 600,000/mm ³ , dilutional thrombocytopenia is unlikely to occur until 4.0 or more blood volumes have been shed, whereas with a starting count of 100,000/mm ³ , dilutional thrombocytopenia should be anticipated after 1.0 blood volume has been lost (see Fig. 12.7B ). Prophylactic transfusion of platelets typically is not indicated without documented evidence of dilutional thrombocytopenia, visible microvascular bleeding, and ongoing blood loss, although it should be anticipated based on the starting platelet count and the volume of blood lost.

Although the primary platelet defect in massive transfusion is thrombocytopenia, some data suggest that platelets may not function normally (i.e., thrombocytopathy) after massive trauma or in the presence of hypothermia. This has not been our experience in the children we studied whose temperature remained within the normal range. The only simple test to assess platelet function is the bleeding time. However, this test is also sensitive to thrombocytopenia and its predictive value is of equivocal utility. The PFA-100 test shows less potential as a rapid screening tool than it once did, because the device uses citrated blood warmed to 37°C and is relatively insensitive to milder defects in platelet-vessel wall interaction, such as that in mild von Willebrand disease. There remains a critical need for a point-of-care device or simple test to assess platelet function, as noted previously. Currently, the platelet count is our best indication for the need for platelet transfusions in situations involving rapid blood loss. Other devices such as thromboelastography to measure whole blood clotting have been used to guide transfusion therapy, but their efficacy in improving outcomes has not been established.

In several in vitro and animal model systems, recombinant factor VII (rFVIIa) activates factors IX and X on the surface of activated platelets, probably through the binding of rFVIIa to the platelet membrane (from which rFVIIa can also be taken up into storage sites within the platelet) and subsequent recruitment of circulating tissue factor. Although this has significantly improved hemostasis in hemophiliacs with inhibitors to factor VIII, there is limited evidence that rFVIIa reduces mortality for off-label use, as in cardiovascular surgery, trauma, and intracerebral hemorrhage. Dosing in children appears to be greater than in adults, but this has not been systematically examined and is anecdotal. This factor increases the risk of thromboembolism, hence its use should also consider the possibility of this sequela.

Basic clotting studies (e.g., PT, PTT, fibrinogen, platelet count) should be performed before elective surgery when major blood loss can be anticipated to determine the cause of underlying coagulopathies and provide adequate quantities of blood components.

Factor Deficiency

Laboratory results for developing deficiencies in clotting factors are integral when managing component therapy in massive transfusions. The PT (for the extrinsic system) measures the adequacy of factors VII, X, and V; prothrombin; and fibrinogen, whereas the PTT (for the intrinsic system) measures the adequacy of factors XII, XI, IX, VIII, X, and V; prothrombin; and fibrinogen (see Fig. 12.4 ). Banked whole blood contains normal plasma concentrations of all of the clotting factors and regulatory proteins, with the exception of factors V and VIII (20%–50% of normal at the time of outdate), as well as factor VII. For a coagulopathy to develop because of a clotting factor deficiency, factor VIII must be less than 30% of the normal concentration and factor V less than 20% of normal. For these to occur, at least 3.0 blood volumes must be exchanged with whole blood. In this scenario, the first coagulation test that is abnormal is the PTT because factor VIII is diluted to less than 30%.

If blood loss is replaced with PRBCs, as is the current practice with modern blood banking techniques, the amount of plasma that is transfused is minimal because ~70% was sequestered in the FFP fraction when it was separated. Massive replacement of blood loss with PRBCs and no other blood products quickly dilutes all of the clotting factors, including fibrinogen (see Fig. 12.4 ). Data have confirmed that the PT and PTT are prolonged in children with multiple clotting factor deficiencies (e.g., during massive transfusion) at concentrations of clotting factors that are greater than in children with single clotting factor deficiencies (e.g., congenital coagulopathies). This was also documented in adult patients who were transfused exclusively with PRBCs and crystalloid; the dilution of multiple clotting factors correlated with the volume of blood and crystalloid transfused. Replacing 1.0 to 1.5 blood volumes with PRBCs and crystalloid exclusively dilutes clotting factors to approximately 30% of normal. Because moderately prolonged PT and PTT values exist without overt signs of clinical bleeding, administration of FFP should be initiated with the onset of a clinical coagulopathy. However, the anesthesiologist should anticipate that PRBCs and crystalloid solutions or albumin will dilute the concentration of clotting factors so that FFP is begun after ~1.0 blood volume of blood loss has been replaced to avoid falling behind in the clotting indexes. Documented deficiency of fibrinogen (<80 mg/dL) may also be corrected by transfusing FFP, but marked deficiency, particularly in the presence of a consumptive coagulopathy (e.g., DIC, fibrinolysis), may require the addition of cryoprecipitate (0.2 to 0.4 unit/kg).

Our experience with 26 children (12 ± 4 years old, weight of 41.9 ± 15.8 kg) who underwent 22 Harrington rod procedures, three tumor excisions, and one Whipple procedure received no FFP or whole blood despite losing between 0.5 and 1.0 blood volume. They exhibited no clinical signs of coagulopathy. Slight prolongations of the PT or PTT occurred when the blood loss was equal to 1.0 blood volume or less ( Table 12.10 ). Two children who lost 1.5 to 2.0 blood volumes exhibited prolonged PT and PTT values. The only child who developed signs of a clinical coagulopathy lost 2.0 blood volumes.

The magnitude of the increase in PT or PTT that is predictive of a clinical coagulopathy is not well defined. However, the consensus panel of the National Institutes of Health and others suggest that when either clotting index exceeds 1.5 times normal (or INR >2.0), it should be considered pathologic. Our studies suggest that the PT and PTT are prolonged to more than 1.5 times normal when the blood loss is 1.5 blood volumes or more and the blood loss has been replaced with only PRBCs and crystalloid or 5% albumin. Our clinical practice is to initiate FFP after 1.0 blood volume blood loss ( Table 12.11 ). At that point, FFP is administered in a ratio of 1 unit for every 2 units of PRBCs transfused. The indications for and timing of FFP depend on which blood product has been transfused, the volume of that transfusion as it relates to the child's blood volume, and whether the blood loss will continue perioperatively. The PT, PTT, fibrinogen concentration, and platelet count should be measured after each blood volume has been replaced and used to guide the need for additional FFP and platelets.

Recombinant factor VIIa (rFVIIa; NOVOSeven) is approved for use in the United States for hemophiliac patients with high-titer inhibitors and congenital factor VII deficiency. Anecdotal reports describe its effectiveness in controlling hemorrhage in a variety of other settings including congenital heart disease. However, in several randomized clinical trials during partial hepatectomy, liver transplantation, prostate surgery, pelvic (orthopedic) surgery, trauma, and upper gastrointestinal tract bleeding rFVIIa failed to confer any benefit. In a large randomized clinical trial of rFVIIa in children with intracranial hemorrhage, the largest-dose group showed only a small improvement in hematoma expansion, and 10% also experienced thromboembolic complications. In adults who received rFVIIa, the frequency of major thromboembolic complications was 1.4% to 10%, including acute myocardial infarction and stroke. Until controlled trials demonstrate a clear benefit for its use, rFVIIa should be used with great caution for off-label indications, and even then only for life-threatening bleeding.

Several studies have compared transfusion strategies in adults for treating trauma incurred during combat in the Middle East. A consensus conference and other reviews on massive transfusion concluded that the evidence did not support the up-front use of a 1 : 1 : 1 ratio of units of PRBC, FFP, and platelets, but it did support the early use of an antifibrinolytic medication (i.e., tranexamic acid). The consensus conference also recommended an integrated approach to managing massive transfusion that included rapid provision of PRBCs, the use of antifibrinolytics, and a foundation ratio of blood components directed by the results of standard coagulation testing (e.g., PT, PTT, platelet count, fibrinogen) or clot viscoelasticity, or both.

The dilutional coagulopathy associated with massive blood transfusion is reasonably predictable. When using whole blood, dilutional thrombocytopenia usually develops first and may occur as early as after the first blood volume has been replaced (if the initial platelet count is <50,000/mm ³ ). In most cases, clotting factors (particularly factors V and VIII) are not diluted until the blood loss exceeds three blood volumes. On the other hand, when PRBCs are used to replace blood loss, all the clotting factors and platelets may be diluted after as little as 1.0 blood volume is lost. However, the predictable coagulopathy of dilution is only an approximate guide. The PT, PTT, fibrinogen, and platelet count should be assessed during massive transfusions to guide replacement therapy.

Disseminated Intravascular Coagulation and Fibrinolysis

DIC and fibrinolysis are frequently associated with shock, trauma, and other forms of tissue damage, with release of procoagulants (e.g., tissue factor) and fibrinolytics (e.g., tissue plasminogen activator). In the presence of massive blood loss, these processes must be differentiated from dilutional coagulopathy. Differentiation may be difficult, because both are associated with pathologic oozing of blood in the surgical field and each may result in prolongation of the PT and PTT, as well as thrombocytopenia. With massive replacement using whole blood or PRBCs and adequate FFP, the fibrinogen concentration should remain normal; with uncompensated (acute) DIC, it may be decreased. However, replacing the blood loss with PRBCs, albumin, and crystalloid also leads to a reduction in fibrinogen.

The most helpful test for DIC and fibrinolysis is documentation of a significant increase in the level of d -dimer, a small peptide fragment generated during the digestion of fibrin by ongoing thrombolysis (i.e., through plasmin), along with evidence on the peripheral blood smear of schistocytes and helmet cells (i.e., microangiopathic hemolytic anemia). Abnormal RBCs and RBC fragments are believed to arise from the slicing action of immobilized fibrin strands in the microcirculation, although the precise mechanism remains unknown. A scoring system to screen for potential DIC has been developed but not evaluated in the operating room setting. If pathologic oozing in the surgical field is observed and 1.0 blood volume or less has been lost in a child who had a normal platelet count and PT and PTT values preoperatively, the child may have developed a consumptive coagulopathy.

The most effective treatment for DIC is to eliminate the cause, such as correcting shock, acidosis, or sepsis. Heparin therapy remains controversial even in children with thrombotic manifestations of DIC. It is not advisable in children with active bleeding, especially in the operative setting.