Initiation

Coagulation is initiated by a membrane-bound lipoprotein called tissue factor. This is usually expressed on subendothelial TF-bearing cells, such as stromal fibroblasts; in the absence of vessel injury it is separated from the vessel lumen. After injury, a complex is formed between TF and factor VIIa (TF/VIIa), which activates factors IX and X. Factor Xa, in association with cofactor Va, forms “prothrombinase” complexes on the surface of the TF-bearing cell, which activates a small amount of thrombin (factor IIa)¹² (Fig. 18-1). The purpose of this small generation of thrombin and Xa is to activate platelets and factors V and VIII.

FIGURE 18-1 Initiation. The TF/VIIa complex on the tissue factor–bearing cell activates factors IX and X. The factor Xa/Va complex, known as the “prothrombinase” complex, forms small amounts of thrombin (factor IIa).

(Reproduced from Hoffman M. Remodeling the blood coagulation cascade. J Thromb Thrombolysis 2003;16:17-20.)

Tissue factor pathway inhibitor (TFPI) and antithrombin III (ATIII) provide a localizing function on factor Xa by inhibiting any factor Xa that becomes dissociated from the TF-bearing cell. Factor IXa is not localized to the cell in the same way.

Low levels of IX and X activation occur in the absence of tissue injury without causing clot formation. The process only leads to amplification when damage to the vasculature allows intravascular platelets and a complex formed by factor VIII and von Willebrand factor (VIII/vWF) to adhere to the extravascular TF-bearing cells.¹¹

Amplification

Small quantities of thrombin are generated on the TF-bearing cells. This sets up the subsequent propagation phase, during which thrombin is generated in large quantities (Fig. 18-2). This thrombin has many major functions:

FIGURE 18-2 Amplification. Small amounts of thrombin (factor IIa) set the stage for large-scale generation of thrombin in the propagation phase. Small amounts of thrombin activate platelets, as well as other important coagulation enzymes and cofactors.

(Reproduced from Hoffman M. Remodeling the blood coagulation cascade. J Thromb Thrombolysis 2003;16:17-20.)

Propagation

Propagation occurs on the surface of activated platelets that are recruited to the site in large numbers. Activated factor IX (from both initiation phase and provided by factor XI on the platelet) binds to VIIIa. The resultant IXa/VIIIa complex activates factor X on the platelet surface. This factor Xa associates with factor Va and forms the prothrombinase complex. The prothrombinase complex causes a “burst” of thrombin generation to cause clotting via fibrinogen (Fig. 18-3).¹¹

FIGURE 18-3 Propagation. Factor Xa is formed locally by the VIIIa/IXa complex on the surface of the activated platelet. The resulting Xa/Va prothrombinase complex causes a burst of thrombin (factor IIa) generation.

(Reproduced from Hoffman M. Remodeling the blood coagulation cascade. J Thromb Thrombolysis 2003;16:17-20.)

This sequence explains why children with hemophilia bleed despite having a normal TF/VIIa complex; the factor Xa during the initiation phase is broken down by ATIII and TFPI if it dissociates from the TF-bearing cell. It is therefore unable to activate the “burst” of thrombin generation that normally occurs in this propagation phase. The factor Xa needs to be generated on the platelet itself via the IXa/VIIIa complex.¹¹

Of note, an alternative pathway is initiated by contact factors (XII, XI, prekallikrein, and high-molecular-weight kininogen [HMWK]). It is of no physiologic importance in terms of coagulation activation; however, it provides important acceleration loops through feedback activation of factors VIII, IX, and XI¹⁴ and is important in fibrinolytic and inflammatory pathways.¹⁵

Clot Inhibition

The clot is confined to the site of injury by direct and indirect thrombin inhibitory systems. The direct system comprises ATIII, α₂-macroglobulin, and heparin cofactor II (HCII). ATIII and HCII activities are accelerated in the presence of heparin.

Several indirect systems inhibit thrombin, such as the protein C–protein S–thrombomodulin (TM) and the TFPI systems. Thrombin binds to TM on the surface of intact endothelial cells and can no longer cleave fibrinogen to form fibrin. The TM/thrombin complex is neither able to activate platelets nor activate factors V and VIII. Instead, this complex activates protein C, which binds to the cofactor protein S and inactivates factors Va (on the surface of endothelial cells and platelets) and VIIIa.^11,14 TFPI bound to endothelial surfaces can form complexes with factor Xa that inhibit factor VIIa. Thrombin generation is therefore inhibited.¹⁴

Fibrinolysis

Fibrinolysis (the breakdown of fibrin into soluble degradation products) is mediated by the proteolytic enzyme, plasmin. Plasmin is formed from an inactive zymogen, plasminogen, which is produced in the liver. This process is controlled by activators and inhibitors. The principal plasminogen activators are tissue plasminogen activator (tPA) and urokinase (uPA). Although overlap exists between the functions of these activators, they have distinct physiologic roles. uPA is primarily involved in a wide variety of extracellular processes, including remodeling of tissues. tPA is the major intravascular activator of fibrinolysis and is discussed in more detail later.¹⁶ The main inhibitory proteins are plasminogen activator inhibitor-1 (PAI-1), antiplasmins (α₂-PI and α₂-macroglobulin), and thrombin activated fibrinolysis inhibitor (TAFI)¹⁷ (Fig. 18-4).

FIGURE 18-4 The main fibrinolytic pathway, leading to the breakdown of fibrin into fibrin degradation products (FDP). It is initiated by release of tissue plasminogen activator (tPA) from endothelial cells in response to contact activation (shaded area). Plasminogen needs to be bound to fibrin in order to allow conversion to plasmin. sc-tPA and tc-tPA refer to single- and (more active) two-chain tPA, respectively. Endogenous fibrinolysis inhibitors are shown in blue boxes: TAFI is thrombin activated fibrinogen inhibitor, PAI-1 is plasminogen activator inhibitor, and α₂-PI is α₂ plasmin inhibitor. Interactions with the coagulation system are shown by red arrows.

As well as cleaving fibrin, plasmin metabolizes a number of other proteins, including the platelet receptor for fibrinogen (glycoprotein IIb/IIIa) and fibrinogen.¹⁸ In addition, plasmin accelerates its own production by metabolizing the conversion of single chain plasminogen activators to more active two-chain versions. The action of plasmin on fibrin produces a series of degradation products some of which convey anticoagulant properties; this effect is achieved by preventing polymerization of fibrinogen and by inhibition of platelet function.

tPA is released by vascular endothelium of small blood vessels. Release is increased in the presence of stimuli such as trauma, endotoxins, ischemia, or normal exercise. This effect is mediated via contact activation (through the kallikrein system) and also by a series of other substances, including thrombin. Once released, tPA is rapidly metabolized by the liver with a half-life of approximately 5 minutes.¹⁷ Fibrin binds both plasminogen and tPA and greatly accelerates the conversion of plasminogen to plasmin (facilitating its own degradation, but also localizing the process to areas of clot). An alternative mechanism for plasminogen activation by tPA exists through binding to receptors expressed by certain cells (endothelium, white cells and some tumor cells); the importance of this in health or disease is unclear.

Excessive fibrinolysis can directly result from excess production of fibrin, as in disseminated intravascular coagulation. This is termed secondary hyperfibrinolysis and in this context the fibrinolysis is considered beneficial because it prevents widespread vascular occlusion. Therapy is directed at replacement of consumed clotting factors, inhibition of excessive coagulation, and treatment of the underlying cause. Primary hyperfibrinolysis can occur during cardiac bypass, massive blood loss, trauma, and liver transplantation.¹⁹ During the anhepatic stage of liver transplant surgery, there is hyperfibrinolysis because of failure to metabolize tPA. On reperfusion of the liver, a further surge of tPA occurs that can take several hours to return to normal. In coagulopathic patients, reduced thrombin formation may lead to reduced production of TAFI (important in inhibition of membrane bound plasmin), whereas conversion of single- to two-strand tPA by plasmin may further sustain the process. Individual susceptibility is likely to be important and may have a genetic component.^20,21

Two groups of drugs are used clinically to inhibit fibrinolysis:

Synthetic lysine analogues are fairly specific inhibitors of plasminogen activation, working by competitively binding to lysine-binding sites on the plasminogen molecule (Fig. 18-5). This blocks the binding of plasminogen to fibrin; a required step for the conversion of plasminogen to plasmin by plasminogen activators.²² At larger doses, they may have additional effects through direct inhibition of plasmin; this includes inhibition of the plasmin mediated effects on platelets.

FIGURE 18-5 Diagrammatic representation of the mode of action of the synthetic lysine analogues tranexamic acid and ε-aminocaproic acid.

(Reproduced from Mahdy AM, Webster NR. Perioperative systemic haemostatic agents. Br J Anaesth 2004;93:842-58.)

Aprotinin is a less specific inhibitor of proteolytic enzymes; it has actions on the kallikrein–kinin (contact) system, as well as enzymes involved in coagulation and fibrinolysis. In addition, aprotinin may be associated with greater preservation of platelet function, as well as an antiinflammatory effect. This wider spectrum of effects from aprotinin may have additional benefits over the antifibrinolytic lysine analogues. Some additional effects may not necessarily be of benefit; for example, the contact system may have protective effects against ischemic reperfusion injury that are inhibited by aprotinin.¹⁵

Genetics of Bleeding

The hemophilias are a group of genetic diseases that cause excessive bleeding, often in response to minor trauma. von Willebrand disease and hemophilia A are the most common variants (see also Chapter 9), associated with low levels of von Willebrand factor and factor VIII respectively. A wide range of single gene defects resulting in deficiencies of single clotting proteins or regulatory proteins have now been described. A further group of single gene disorders may result in thrombophilic disorders, associated with abnormalities of inhibitory proteins.

Unexplained variation has been observed in bleeding between apparently similar patients in the absence of specific factor deficiency. The causes of this variation are likely to be multifold according to nuances of surgical technique and subtle difference in disease process and therapy. It might appear counterintuitive that genetic factors have a significant role in acquired bleeding resulting from surgery; however, recently there has been increased interest in the interplay of genetic and environmental factors in the progress of acquired diseases. The genetics of most clotting proteins has been described, and common variations within populations have been revealed for some. Of these, a common polymorphism of PAI-1 has been well described. PAI-1 is an important endogenous inhibitor of fibrinolysis, and deficiency is associated with increased bleeding.³³ The G5/G5 polymorphism is common (about 20% in European populations) and is associated with lower levels of PAI-1. It has been linked (though not consistently) to bleeding after heart surgery and to increased benefit from use of antifibrinolytics.^20,21 Should these findings be substantiated, then PAI-1 may still be an exceptional example. In general, the influence and consequences of genetic determinants of bleeding are likely to be more subtle. In infancy an additional factor in the mix may be the relationship between developmental and genetic factors. Many clotting proteins are present in infants as isoforms distinct from those in adults. This implies a different gene expression in the young. It is possible that understanding genetically determined variations in bleeding will increase our understanding of bleeding and, perhaps, allow us to guide therapy for the individual patient. The practical applications of such techniques, however, remain speculative.

Routine Anticoagulation

Failure of Hemostasis Associated with Cardiac Surgery

Complex abnormalities occur in the coagulation system in cardiac surgery, owing to the profound surgical insult, hypothermia, acid-base disturbance, blood transfusion, anticoagulants, CPB (contact activation, platelet dysfunction, and hemodilution), and in some cases, deep hypothermic cardiorespiratory arrest. In addition children and infants with congenital heart disease may have preexisting coagulation defects or be taking medications that affect coagulation before surgery (Table 18-1).

TABLE 18-1 Causes of Excessive Bleeding after Pediatric Cardiac Surgery

Antiplatelet drugs such as aspirin or clopidogrel may exacerbate bleeding. The clinician must balance the risk of perioperative bleeding against the risk of drug discontinuation when deciding if and when to withhold the drug before surgery. In most cases aspirin can be safely discontinued 5 days before surgery. Prostaglandin E₁ can inhibit platelet aggregation, acting in synergy with endothelial cell–derived factors (nitric oxide and prostacyclin) at clinically relevant concentrations,⁶⁰ although this effect was too subtle to detect in vitro using the TEG.³² It is usually not possible to stop the prostaglandin E₁ infusions, because neonates are dependent on its use to maintain ductal patency. For elective surgery in children on oral anticoagulant therapy, it is usually possible to transfer to heparin before surgery. An INR of less than 1.5 is usually considered acceptable for surgery. If urgent correction of anticoagulants is required, prothrombin complex concentrates, together with vitamin K, are more effective than FFP.⁶¹ FFP should only be used if prothrombin complex concentrates are not available.

In the child with cyanotic congenital heart disease, hemostasis is further impaired because of polycythemia, low platelet count and altered function, reduced factors V, VII, and VIII, and increased fibrinolysis.^32,62 The degree of derangement is related to the degree of cyanosis.⁶³ Preexisting coagulopathy may also occur in children with severe underlying illness or poor nutritional status.

Despite improvements in design and materials of the CPB circuit, abnormal activation of the coagulation and fibrinolytic systems persist. The normal balance of coagulation and fibrinolysis is particularly delicate in children, and more susceptible to exogenous perturbations.⁶⁴ Shortly after CPB is established in children, all hemostatic protein concentrations are decreased (to a variable degree) as a result of dilution.⁶⁵ Because of the relative volume of the circuit in relation to the size of the child, it is not surprising that the magnitude of the effect of hemodilution is greater than in adults.⁴⁶

Reductions in coagulation factor concentrations after cardiac surgery contribute to bleeding in adult cardiac patients,⁶⁶ and a reduction in platelet count is well described on initiation of bypass. In adults, the platelet count is reduced on CPB by approximately 50%.⁶⁷ This is also attributed primarily to dilution, although other factors include organ sequestration, mechanical disruption, and adhesion to the circuit.⁶⁸ A relatively greater reduction will occur in smaller children, although newer combined filter and oxygenators should considerably reduce the circuit volume and, hence, the degree of dilution. Significant platelet dysfunction also occurs.^46,69 Low platelet counts have been strongly associated with increased bleeding.^10,31,70

Further activation of platelets and denaturing of plasma proteins will occur at interfaces between blood and air, or on contact with extravascular tissue. CPB pump suction and spilling of blood into the pericardium will further contribute to coagulopathy. Greater coagulopathy is also seen during prolonged bypass and when deep hypothermic circulatory arrest is employed.⁷¹

The major mechanism for activation of the coagulation cascade during CPB is thought to be the “extrinsic” TF pathway, which is activated as a result of surgical trauma and inflammation.^42,72,73 Inflammatory mediators such as tumor necrosis factor (TNF) and interleukin-1 induce expression of TF on endothelial cells and monocytes. The “intrinsic” coagulation system is also activated when factor XII is adsorbed onto the surface of the CPB circuit, causing activation of complement, neutrophils, and the fibrinolytic system via kallikrein.⁴² Although the intrinsic system has little role in initiating coagulation, activations of kinins will lead to increased fibrinolysis and inflammation.¹⁵

The blood loss in the first 24 hours after cardiac surgery can vary between 15 and 110 mL/kg, although the risk of excessive blood loss is greatest among those weighing less than 8 kg, younger than 1 year of age, and children undergoing complex surgery.¹⁰ The common preoperative and intraoperative risk factors for excessive bleeding are summarized in Table 18-2.

TABLE 18-2 Common Predictive Factors for Bleeding

CPB, Cardiopulmonary bypass.

Williams GD, Bratton SL, Ramamoorthy C. Factors associated with blood loss and blood product transfusions: a multivariate analysis in children after open-heart surgery. Anesth Analg 1999;89:57-64.

Blood loss and transfusion requirements in pediatric cardiac surgery vary inversely with age; neonates bleed more and receive more products per kilogram than any other age group.⁷⁴ There are some correlations between coagulation tests before and during CPB and postoperative blood component transfusion requirements, although the sensitivity and specificity of the tests are not sufficient to justify routine coagulation tests while on CPB. Of all the tests, a platelet count of less than 108,000/mm³ while on CPB yielded the greatest sensitivity (83%) and specificity (58%) for predicting excessive blood loss.⁷⁰

Blood Products

Frequently it is necessary to administer blood products on a largely empirical basis; laboratory tests describe only parts of the coagulation process³² and are practically too slow to direct the anesthesiologist in real time as to the requirement for blood components. The TEG is a dynamic whole-blood test that is frequently used as a near-patient test to assess clot elasticity properties and, hence, more precisely delineate the bleeding and homeostasis profile. The role of TEG during pediatric heart surgery has been recently reviewed.⁷⁶ The use of treatment algorithms based on TEG can limit transfusion of blood products in adults^77–81 and children.⁸² In children, it may not be appropriate to use protocols originally designed for adults. An alternative approach has recently been proposed.⁷⁶

Platelets

Platelet dysfunction and thrombocytopenia are common after cardiac bypass in infants. Hence, in the presence of bleeding, transfusion of platelets is logical. Coagulation variables that relate to platelet number and function (such as TEG maximum amplitude) are corrected by infusion of platelets, and clinical experience indicates that platelet transfusion reduces bleeding. This has been confirmed in a small study of children after cardiac surgery.³¹ A platelet count of less than 108,000/mm³ has been identified as a predictor of bleeding, whereas clot strength measured by TEG reduces steeply at platelet counts below 120,000/mm³. The current recommendation of targeting a platelet count of approximately 100,000/mm³ appears to be logical. It would also appear reasonable to use platelets for the initial treatment of presumed coagulopathic bleeding (in the absence of specific clotting tests).

Fresh Frozen Plasma

The case for use of fresh frozen plasma (FFP), either as empirical treatment of bleeding, or guided by coagulation tests, is considerably weaker than for platelets. The use of FFP in cardiac surgery is based on the observation that the concentration of clotting factors are often low in bleeding patients, especially in the period following bypass. The PT (with its derived measure, INR) is the most common test used to detect the presence and gauge the severity of clotting factor–deficient coagulopathy. Unfortunately, observation studies have shown that the PT correlates poorly with clinical bleeding, and that transfusion of plasma often achieves no measurable change in the INR, nor is of any known clinical benefit (particularly if the INR was only marginally raised to less than 1.7).⁸³ Despite this, FFP is frequently transfused in the absence of either bleeding or significantly raised PT.⁸⁴ Administration to patients with higher PT values will be more effective in correcting clotting tests, however large volumes may be required,⁸³ even in the absence of ongoing loss of clotting factors. The situation is further complicated, in that coagulation tests that are initially normal will deteriorate during severe bleeding.

A systemic review of the use of FFP to treat or prevent bleeding resulting from acquired coagulopathy failed to demonstrate benefit, although the trials examined were small, used a wide range of outcomes, and were conducted in heterogeneous populations.⁸⁵ In a small observational study of children undergoing heart surgery, a number of patients had coagulopathic bleeding after transfusion of platelets; if these patients were then given FFP, the bleeding increased, whereas if cryoprecipitate was given bleeding decreased.³¹

It is possible that the lack of efficacy may be related to the dose of FFP used. The dose of FFP most frequently recommended (10 to 15 mL/kg)^84,86 may be inadequate to restore factor concentration.⁸³ Of note, dosage of plasma is inexact. Hence, interinstitution practices differ considerably in defining both a threshold INR for prophylactic plasma administration and the actual dosage of plasma prescribed.⁸⁷

Modeling of FFP administration in major trauma suggests that larger doses (30 to 40 mL/kg) may be required to increase or maintain factor concentrations⁸⁸; however, administration of such a large volume is often undesirable in the absence of severe ongoing bleeding. Addition of FFP to the bypass circuit or during modified ultrafiltration may allow administration of much larger doses; benefit appears to be confined to younger children.^89,90 It is always important to document the specific reasons for FFP administration.

Cryoprecipitate

Not all coagulation factors are of equal importance during bleeding. Fibrinogen is present in much greater concentrations than other clotting factors, and while other factors are mainly involved in initiating or amplifying thrombin formation, fibrinogen is a substrate for the production of fibrin. Deficiencies of fibrinogen are reflected in reduced strength of clot and are associated with increased bleeding.⁹¹ Cryoprecipitate is the most concentrated fibrinogen replacement widely available, with a fibrinogen concentration 4 to 8 times that of FFP; however, a unit of cryoprecipitate contains less fibrinogen than a unit of FFP, and multiple cryoprecipitate units will be required in larger patients. One cryoprecipitate unit for each 10 kg should raise fibrinogen concentration by 0.5 to 1.0 g/L. As well as fibrinogen, cryoprecipitate is a source of von Willebrand factor and factors VIII and XIII, although the clinical significance of this to bleeding is unclear. Cryoprecipitate has been effective in treating bleeding resistant to platelet concentrates alone during pediatric heart surgery.³¹

Fractionated Human Blood Products

Products such as FFP or cryoprecipitate can be considered as only crudely purified blood products. It is possible to produce more refined products containing high concentrations of only a single clotting protein or relatively standardized concentrates of selected proteins. Examples of these are prothrombin complex concentrates and fibrinogen concentrate (e.g., Riastap, CSL Behring, King of Prussia, Pa., USA; Haemocomplettan P, CSL Behring, Marburg, Germany). These products are human blood products produced from pooled plasma; they are presented as a powder requiring reconstitution for use. They do not require freezing or cross matching, which simplifies their supply, storage, and administration. Risk of viral transmission should be low⁹² because of sourcing of plasma from low-risk populations, pooling of plasma from many individual donors (reducing viral load resulting from a single infected donor), and pasteurization. Although the effect of pasteurization on prion infection is less certain, the risk of transmission would be expected to be similarly low. Recombinant factor concentrates (as opposed to factor concentrates of human origin) are increasingly available and avoid cross infection. One such agent, recombinant activated factor VII (rVIIa), is discussed later in this chapter (see also Chapter 10). Although the agents discussed in this section are used in a similar way to traditional factor supplementation to reestablish near physiologic concentrations, rVIIa is used very differently to produce factor levels greatly in excess of “normal.”

Prothrombin complex concentrates (PCCs) are mixtures of factors II (prothrombin), VII, IX, and X, plus the coagulation inhibitors Protein S and C. They are indicated when urgent reversal of oral anticoagulation (with warfarin or related drugs) is required. Dose is specific for each make of PCC, and is dependent on the child’s INR and the target for correction (generally an INR of 1.5). Older reports of lack of efficacy or risk of thrombosis probably relate to older brands of PCCs with low levels of factor VII or of inhibitors.⁹³ PCCs are more efficacious than FFP for warfarin reversal.⁶¹ They have also been used in other causes of acquired coagulopathy and bleeding, including after adult heart surgery.^94,95 Potentially these products are superior to FFP in such situations; however, routine use for treatment of acquired bleeding in children (other than reversal of warfarin) cannot be recommended until further evaluation is conducted. Dosing in this situation is also uncertain.

The importance of fibrinogen depletion in coagulopathy and bleeding has been increasingly appreciated.^96,97 It has also been recognized that traditional cut-off values for supplementation of fibrinogen during bleeding (1.0 g/L) are likely to be too low and that further advantage can be gained from higher targets; possibly as high as 2.0 g/L. This may be difficult to achieve with traditional blood products. FFP contains only physiologic concentrations of fibrinogen and only modest increases may be achieved, while cryoprecipitate will require administration of product from several donors to achieve significant increases in fibrinogen concentration (especially in larger patients). Human fibrinogen concentrate has been available in continental Europe for some time and has recently been licensed within the United Kingdom and the United States for treatment of congenital fibrinogen deficiency. Unlike cryoprecipitate, they contain only fibrinogen and a direct comparison in other forms of coagulopathy has not been made. Recent reports have demonstrated advantages to its use in adult cardiac surgery and other causes of bleeding.^98,99 Reports of its use in children with acquired coagulopathy are limited. Combining use of fibrinogen concentrates with rapid measurement of fibrinogen concentration (using modified TEG) is an attractive approach.¹⁰⁰

The most obvious use for fibrinogen concentrates is as an alternative to cryoprecipitate. Their apparent safety and ease of administration may, however, make other applications possible. A small randomized study comparing prophylactic administration of fibrinogen concentrate to placebo in adult cardiac patients demonstrated a reduction in bleeding.⁹⁹ Alternatively, it has been proposed that targeting larger fibrinogen concentrations with these agents may reduce the requirement for platelet transfusion. Animal work has demonstrated the effectiveness of this approach in models of bleeding and thrombocytopenia.¹⁰¹ The product of fibrinogen concentration and platelet count may be a better predictor of bleeding in children than either measure alone.⁹¹ If the concentration of fibrinogen is high, the amplitude of TEG recordings can be normal, even in the absence of platelets. In vivo, it would be expected that there are limits to this phenomenon; at some point, platelets are required for formation of clot. Such an approach requires proper clinical evaluation before being more widely adopted. Recombinant fibrinogen in fibrin sealants is currently undergoing phase II trials, although widespread availability of systemic recombinant fibrinogen is still some years away.¹⁰² With proper assessment, recombinant fibrinogen has the potential to greatly alter the management of severe bleeding and coagulopathy.

Deficiency of factor XIII during and after cardiopulmonary bypass has been described, although the importance of this finding is uncertain.^103–105 Supplementation with human or recombinant factor XIII has been demonstrated to reduce bleeding after adult heart surgery, although possibly only in the presence of factor deficiency.^104,106,107 The action of factor XIII is to catalyze the formation of cross bridges between fibrin molecules. It also has a wider function in promoting fibroblast proliferation, and it has been used in the treatment of chylothorax after pediatric heart surgery (with possibly only minor benefit).¹⁰⁸ Recombinant factor XIII is currently being considered for approval by the FDA, and wider availability may lead to greater use of this agent. Evidence to support widespread use is currently not available.

Pharmaceutical Agents to Reduce Bleeding

Blood products are, in most cases, effective at treating coagulopathic bleeding, but may be associated with deleterious effects on the patient. Drug use is aimed at reducing bleeding and blood product requirements. These drugs are principally the two classes of antifibrinolytics:

More recently rVIIa has been employed as a rescue therapy for severe bleeding unresponsive to traditional therapy using blood products. Desmopressin has also been used in the past and is discussed later.

The Effectiveness of Synthetic Antifibrinolytics during Pediatric Heart Surgery

A meta-analysis comparing TXA and placebo in children undergoing heart surgery² concluded that antifibrinolytics reduced bleeding and that TXA was as effective as aprotinin. Figure 18-6 shows a summary of the studies included in this meta-analysis^109–116; the figure has been updated to include a more recent trial.¹¹⁵ There were 1061 children entered into nine studies, and results demonstrated a mean reduction in bleeding of approximately 14 mL/kg (95% confidence interval [CI] of 12.7 to 15.6 mL). This is a clinically useful reduction in blood loss. The meta-analysis also demonstrated a small reduction in transfusion of red blood cells with use of TXA.

FIGURE 18-6 A forest plot of trials of synthetic antifibrinolytic drugs (ε-aminocaproic acid [EACA] and tranexamic acid [TXA]) to reduce bleeding during pediatric heart surgery. Number in parentheses after each study refers to reference from main text. *Study is from The All Indian Institute of Medical Science.

(Forest plot generated using Review Manager (RevMan) Version 5.1. Copenhagen: The Nordic Cochrane Centre; The Cochrane Collaboration, 2011.)

Data concerning more clinically important outcomes (e.g., reexploration or “life-threatening” hemorrhage) are lacking. Analyses have a high degree of statistical heterogeneity (I² = 78%), indicating differences in the populations, interventions, or study methodology.^115–117 This makes pooling of results less valid and makes it impossible to exclude factors that could introduce bias to conclusions. The patients and protocols are clinically variable with differing drugs, dosing regimens and patient populations. Only 163 of the patients did not have cyanotic heart disease; data on 6-hour blood loss was available for 144 of these children and did not demonstrate a significant reduction in bleeding (mean reduction 1. 8 mL/kg, 95% CI of −0.8 to 4.5 mL/kg).^115–117 The majority of studies come from a single institution (The All Indian Institute of Medical Science),^111–114 and these patients represented a specific population undergoing relatively late correction of cyanotic disease. Polycythemia, and presumably severe cyanosis, was common, with a mean hemoglobin concentration of 19 g/dL.

Another feature of the studies^115–117 is that only one study recruited infants less than 2 months of age.¹¹⁶ The mean age of patients was 40 months. This may be important, because the fibrinolytic system in neonates differs from adults; plasminogen levels are about 50% of adult values, plasminogen activation occurs more slowly, and levels of inhibitor (PAI-1) are normal.²⁵ This may make neonates relatively resistant to fibrinolysis. In turn, this may mean that antifibrinolytic medication is less effective, although in vitro experiments using neonatal plasma do not demonstrate a difference in the ability of synthetic antifibrinolytics to inhibit fibrinolysis.^118,119

In conclusion, current pediatric data support the use of synthetic antifibrinolytics in cyanotic children older than 2 months of age, but data supporting more widespread use are unclear. While there is substantial evidence to support the use of these drugs in adults,¹²⁰ extrapolation of conclusions to children is uncertain and has implications with regard to drug development.¹²¹ If it can be demonstrated that there are similarities for both the disease process (e.g., hyperfibrinolysis as a contributor to hemorrhage) and drug effect at different age groups (e.g., hyperfibrinolysis reduction), then adult data for efficacy could more reliably be used in children. Hyperfibrinolysis is most likely part of an inflammatory response to cardiopulmonary bypass and would be expected to occur in children (other than neonates) in a similar way to adults. It would also be expected that antifibrinolytic drugs function in a similar way in children and adults. It seems reasonable to extrapolate adult data of efficacy to children. A more difficult question is whether this reduction in bleeding is worthwhile when balanced against adverse effects (discussed later). Small children have more bleeding (relative to body weight) compared to adults. This does not in itself imply that an intervention to reduce bleeding will have greater effect in children. It has been suggested that factors other than fibrinolysis make a greater contribution to bleeding in children.⁹¹

Synthetic Antifibrinolytic Medications

There is considerable uncertainty about the dose of these drugs. A recent survey of pediatric cardiac anesthetists in the UK demonstrated a 50-fold variation in TXA dose used during surgery (from a single dose of 5 mg/kg to a cumulative dose of 250 mg/kg).¹²² In the studies summarized in Figure 18-6 there was also wide variation.

Rational dosing of a drug is determined by its PK and how the resulting plasma (or tissue) concentration translates to a clinical effect (PD). TXA has a volume of distribution (Vd) of 9 to 12 L in adults. The drug is water soluble and poorly protein bound. Most of the drug is excreted unchanged by the kidneys, so the dose should be reduced in renal failure. The terminal elimination half-life is about 2 hours,¹²³ although CPB will affect this. In adults a loading dose of 12.5 mg/kg followed by 6.5 mg/kg/hour with 1 mg/kg added to the pump prime maintains the concentration of TXA greater than 52.5 µg/mL.¹²⁴ However, neonates and infants will be expected to have a relatively larger Vd and a greater change in Vd with initiation of CPB. In addition, clearance will be reduced in neonates because of renal immaturity. The effects of the surgical insult, CPB, hypothermia, ultrafiltration, and fluid resuscitation on the PK and PD of TXA remain to be established.

The target concentration of TXA is not clear. A plasma concentration of greater than 100 µg/mL will produce complete inhibition of fibrinolysis in vitro, however a concentration of 30 µg/mL accounts for most of the pharmacologic effect of the drug.¹²⁵ It is difficult to translate these in vitro effects into clinical dosing recommendations. Dosing regimens that involved multiple doses were more effective than a larger single bolus at the start of cardiac surgery, in terms of sternal closure time, blood loss, and blood product requirement, despite similar total doses.¹¹³ The administration of a loading dose (50 mg/kg repeated on bypass) in addition to an infusion of 15 mg/kg/hour achieved mean tranexamic concentrations above 100 µg/mL, although concentrations may be below this concentration after bypass.¹²⁶ This dosing regimen reduced bleeding compared with placebo, and greater tranexamic concentrations were associated with less bleeding (although this does not demonstrate causality). Dosing regimens using a smaller total dose (30 to 50 mg/kg in divided doses) have also been effective. A detailed PK study of TXA in children undergoing heart surgery is currently under way and may help to improve decision making about dosing.

The use of EACA during surgery to correct a congenital heart defect was first described in 1969.¹²⁷ The dose of EACA is based on a small PK study in children undergoing heart surgery.¹²⁸ Most of the drug is excreted unchanged by the kidneys, while about 35% of EACA undergoes hepatic metabolism. The terminal elimination half-life is 1 to 2 hours.¹²⁹ Children have a greater weight-adjusted clearance and Vd compared with adults before, during, and after CPB. Consequently, larger doses are used in children: 75 mg/kg over 10 minutes, 75 mg/kg in the CPB circuit, and 75 mg/kg/hour as an infusion. This regimen maintains concentrations of EACA above the presumed “therapeutic” concentration of 260 µg/mL.¹²⁸ While the target concentration is not clear, this triple-dosing regimen appears to be clinically effective in children. One study (using three doses of 100 mg/kg EACA before, during, and after CPB) demonstrated a reduction in postoperative blood loss, reduced blood and platelet transfusion requirements, and reduced reexploration rate compared with control subjects.¹¹⁴ The PK have not been studied in infants and neonates.

Serine Protease Inhibitor: Aprotinin

The use of aprotinin has become a source of considerable controversy since publication of data that appeared to demonstrate increased morbidity and mortality with its use in adults undergoing heart surgery.^3,130 This is discussed further, later, but has lead to a reduced availability of the drug in most countries. If aprotinin remains available, the decision to continue to use it must rest on a balance of the benefits (principally reduction in bleeding) and uncertain risks. More precisely, one must consider how this risk–benefit balance compares with that of the synthetic antifibrinolytics, which have more certain risk profiles.

Which Antifibrinolytic Drug for Pediatric Cardiac Surgery?

A meta-analysis of trials in adult patients demonstrated no difference between EACA and TXA. Aprotinin was more effective than either TXA or EACA, and although that effect appeared marginal, only aprotinin was demonstrated to reduce the risk of reoperation.¹²⁰ Direct comparison between TXA and EACA has demonstrated no difference in children.^112,152 Low-dose aprotinin (10,000 KIU/kg followed by 10,000 KIU/kg/hour) has been compared with EACA (100 mg/kg on induction, 100 mg/kg in the pump prime and 100 mg/kg on weaning CPB).¹¹⁰ There was a reduction in postoperative blood loss and in transfusion requirements in both groups compared with control but no significant difference between the two drugs. These data are difficult to interpret because substantially larger doses of aprotinin are frequently used in current practice. More recently, TXA (100 mg/kg before, during, and after CPB) was compared with aprotinin (30,000 KIU after induction, 30,000 KIU in the pump, and 30,000 KIU after weaning off bypass). There was a reduction in the time to sternal closure, transfusion requirements, and blood loss compared with control, but no significant difference between the two drugs. When the two drugs were combined at the same doses, they showed no additional benefit.¹⁰⁹ A recent retrospective study confirmed these findings.¹⁵³

There remains uncertainty about the comparative effectiveness of lysine analogues and aprotinin; data suggest that there are no differences or only marginal differences. The choice of antifibrinolytic will, therefore, depend on the adverse effect profiles and cost. The incidence of important adverse effects is also uncertain. Both mechanistically and from the available data it would appear that the recent “safety” concerns about aprotinin in adults may not be applicable to children. A more certain risk of aprotinin is anaphylaxis with repeated exposure. This is a concern because many pediatric cardiac patients (and a large proportion of those at a increased risk of severe bleeding) will undergo repeat surgery. The greater cost of aprotinin will also be difficult to justify in the absence of clear advantage over lysine analogues. In the absence of further studies, it is the authors’ conclusion that if an antifibrinolytic is indicated, lysine analogues are preferred over aprotinin.

Use of Antifibrinolytic Drugs in Noncardiac Surgery

A recent trend has been towards increased use of antifibrinolytic drugs (especially TXA) during noncardiac surgery; principally major orthopedic surgery (e.g., spinal instrumentation, see also Chapter 30) and craniofacial surgery (see also Chapter 33). Aprotinin is less commonly used because of reduced availability and the safety concerns discussed previously.

There are 27 trials that have examined the use of TXA during orthopedic surgery in adults¹²⁰; they report a 50% reduction in the relative risk of requiring a blood transfusion. There was a reduction in the incidence of total blood loss of over 400 mL in 20 of these trials. Almost all these studies were conducted in adults undergoing major joint replacement surgery. In one study of adults (n = 147) undergoing posterior spinal fusion, TXA reduced total blood loss by 25%; however, the reduction in blood products transfused was not statistically significant.¹⁵⁴ The popularity of antifibrinolytic use during scoliosis surgery in children and adolescents is reflected in two surveys, which show use in 70% to 80% of hospitals.^155,156 It is unclear whether this use is mostly confined to higher-risk cases; however, in the authors’ institution TXA is currently used during all scoliosis surgery as part of a program aimed at reducing the use of blood products. Two meta-analyses have examined this question (looking at 6 and 7 studies, respectively); no individual study had more than 45 patients.^2,157 These meta-analyses conclude that antifibrinolytic drugs reduce bleeding and volume of red cells transfused. To date there has been no single prospective trial that compares different antifibrinolytics. A retrospective comparison of TXA and EACA appeared to show superiority of TXA; however, doses used were not comparable.¹⁵⁸ The two meta-analyses failed to establish any difference between the drugs.^2,157 It is also not possible to determine whether reported benefits hold for all children undergoing scoliosis repair. Bleeding is greater in children with secondary scoliosis (including patients with Duchenne muscular dystrophy),¹⁵⁹ those with hemostatic disorders, and in those undergoing more extensive surgery; these might, therefore, be expected to see a greater advantage in terms of bleeding reduction. At least one study examined children with idiopathic scoliosis and demonstrated a reduction in blood loss with use of aprotinin (see also Chapter 30).¹⁶⁰ As with cardiac studies there is a considerable range in the dose of drug used; in a survey of practice in the UK, TXA dose ranged from three- to fivefold, whereas in published trials, the dose varied 10-fold.¹⁵⁵

Antifibrinolytics are also used widely in craniosynostosis surgery. In a survey of North American practice, they were used in 20% of hospitals for strip craniotomy, rising to 30% for more complex repairs.¹⁶¹ Four small trials (total 141 children) have addressed the use of these agents for this indication; three using TXA^162–164 and one using aprotinin.¹⁶⁵ All of these trials have demonstrated reductions in blood transfused and in bleeding. A recent publication demonstrated a greater than 50% reduction in the volume of blood transfused and in the proportion of children transfused (70% vs. 37%).¹⁶² Although all these studies were relatively small, it would appear likely that antifibrinolytics produce a useful reduction in blood loss for craniosynostosis surgery.

Antifibrinolytic Medication and Trauma

There is no evidence in children concerning the use of antifibrinolytic medications to treat massive hemorrhage following trauma. The CRASH 2 study (a blinded randomized control trial of over 20,000 patients) demonstrated a reduction in mortality with TXA in adult trauma patients with, or at risk of, significant hemorrhage; mortality was reduced from 16% to 14.5%.¹⁶⁶ Mortality due to bleeding was reduced from 5.7% to 4.9%. No difference was seen in rates of vascular occlusive events. Despite the size and complexity of the study, it appears to have been properly randomized and blinded. It is difficult to see an explanation for the reduction in mortality, other than a reduction in bleeding. It is therefore likely that TXA can be efficacious in the treatment of established bleeding (at least in the context of trauma). Ideally, further studies would be required to establish the effectiveness of TXA in the treatment of established bleeding of other causes in children; however, it is unlikely a controlled trial of pediatric trauma would be large enough to be adequately powered. In the absence of such studies, and in the absence of definite evidence of toxicity, it would be reasonable to consider TXA as part of the treatment of established severe bleeding resulting from trauma or surgery. Some caution is required, and given the uncertainty concerning dosing, it would be premature to recommend the widespread use of the drug in young children. When used, it should be reserved for severe bleeding; for example, bleeding requiring treatment with non–red cell blood products. An important caveat to this is that administration greater than 3 hours after injury appeared to increase mortality.¹⁶⁷ The significance of this is unclear; however, such late administration should be avoided.

Authors’ Recommendations for Use of Antifibrinolytics

There are 2260 children and more than 25,000 adults who have been recruited into trials that investigated antifibrinolytics during surgery. As a result, we can conclude that these drugs reduce bleeding during a variety of surgeries. However, the dose remains uncertain, important adverse effects unclear, the choice of drug open to question, and their effectiveness in different patient subgroups and clinical situations debatable. Further studies should be designed very carefully to address these questions. Establishing the safety of interventions, in particular with respect to uncommon but devastating complications, is a particular challenge. In the absence of this data it is not possible to make solid recommendations. One of the author’s current practice is as follows:

Desmopressin

Desmopressin acetate (1-desamino-8-d-arginine; DDAVP) is a synthetic analogue of vasopressin. The production of desmopressin involves alteration in the chemical structure of naturally occurring vasopressin. In the process, the antidiuretic effect is enhanced and the vasopressor effect is virtually eliminated. Desmopressin is more resistant to enzymatic cleavage, and hence the duration of action is prolonged to 6 to 24 hours.¹²³ Desmopressin potently causes endothelial release of factor complexes VIII/protein C and VIII/vWf. It has been used in mild hemophilia, von Willebrand disease, coagulopathy of uremia, liver failure, and in adults undergoing cardiac and spinal fusion surgery (see also Chapter 10).

During the early phase of coagulation, platelets bind via the glycoprotein receptor Ib to vWf on the damaged endothelium. IV desmopressin releases vWf, thus enhancing the binding of platelets to damaged endothelium. The maximum effect of desmopressin is observed at a dose of 0.3 µg/kg. It must be given after cessation of the extracorporeal circuit to prevent unwanted platelet activation.⁴⁸

Several meta-analyses of its use in adults have been published, the most recent showing that desmopressin was associated with a small reduction in blood loss, but no benefit in terms of repeat sternotomy, proportion of patients requiring transfusion, or mortality.¹³¹ It was also associated with a 2.4-fold increase in risk of myocardial infarction.

In children, desmopressin at 0.3 µg/kg failed to demonstrate any benefit for both non–high-risk¹⁶⁹ and high-risk cardiac surgery.¹⁷⁰ Younger children are not as capable of releasing vWf from endothelial storage sites as are older children, and the maximal release of vWf caused by the operative stimulus cannot be enhanced by desmopressin.¹⁷¹ Potential adverse sequelae from desmopressin include fluid retention, hyponatremia, tachyphylaxis, tachycardia, and mild hypotension.¹⁷² Desmopressin is not currently recommended in pediatric surgery,⁶⁴ although one could postulate that it could be reserved for those with ongoing bleeding with evidence of platelet function abnormalities, such as prolonged bleeding time or decreased maximum amplitude values on a TEG.¹⁷³

Factor VIIa

Recombinant activated factor VII (rVIIa) is known to be safe and effective for treatment and prevention of hemorrhage in children with hemophilia who have circulating inhibitors to replacement factors (see also Chapters 9 and 10). It is also used in patients with Glanzmann thrombasthenia refractory to platelet transfusion, and in patients with Factor VII deficiency. Regulatory approval has been granted in Europe and the United States for these conditions.

Off-label use of this drug has grown steadily over the last few years and now accounts for 97% of in-hospital use.¹²⁵ This has encompassed a number of indications involving treatment or prevention of acquired bleeding or coagulopathy^174–186 of varying severity. The most common off-label indications are treatment of bleeding related to cardiovascular surgery, trauma, and intracranial bleeds. Off-label use in children is also widespread. One retrospective study recorded 3655 administrations (in 39 U.S. pediatric hospitals over a 7-year period); 46% were in children admitted to a surgical specialty or to a pediatric intensive care unit, over 20% to cardiac surgery or cardiology. Administration was most common in children under 1 year of age.

For rVIIa to exert a beneficial action, it must generate a burst of thrombin at the sight of injury. Two mechanisms are likely to work in synergy to produce this. First, the TF pathway described previously is stimulated to augment generation of factor Xa. Second, high concentrations of rVIIa will bind directly to the surface of activated platelets, again activating factor Xa, leading to thrombin production (Figs. 18-7 and 18-8).¹⁸⁷

FIGURE 18-7 Schematic representation of the cell-based coagulation model and the proposed mechanism of how recombinant activated factor VII (rVIIa) can potentially improve coagulation in hemophiliacs. At supraphysiologic concentrations, rVIIa can bind to the phospholipid membranes of activated platelets, where it activates factor X independent of the tissue factor (TF) pathway, causing a large rise in thrombin at the platelet surface. It can therefore compensate for a lack of factor VIII or IX, a possible explanation for its effectiveness in platelet function disorders

(From Welsby IJ, Monroe DM, Lawson JH, Hoffmann M. Recombinant activated factor VIIa and the anaesthetist. Anaesthesia 2005;60:1203-12.)

FIGURE 18-8 Theoretical mechanisms by which recombinant activated factor VII could increase thrombin generation and fibrin deposition that may lead to improved hemostasis.

(From Welsby IJ, Monroe DM, Lawson JH, Hoffmann M. Recombinant activated factor VIIa and the anaesthetist. Anaesthesia 2005;60:1203-12.)

Usually such processes will be confined to the site of injury; however, in markedly proinflammatory states, expression of tissue factor on activated monocytes and platelets increases the risk of thrombosis distant from the injury. For this reason rVIIa is contraindicated in disseminated intravascular coagulation and should be used with caution in children on extracorporeal life support.¹⁸⁸ The importance of this to more routine bypass is unknown. The therapeutic effects of factor rVIIa begin at doses up to 10 times greater than physiologic concentrations of the endogenous factor. It is therefore not simply a “replacement” therapy of a deficient factor.¹⁴⁷

The action of rVIIa is to produce a surge of thrombin; however, because clots comprise fibrin, platelets, and red cells, it is unlikely that rVIIa will be effective in the absence of these substrates. Measures should be taken to ensure adequate fibrinogen concentration and platelet numbers before giving rVIIa. In addition, factors such as acidosis and hypothermia will reduce the efficiency of rVIIa (although mild to moderate hypothermia has only minor effects in vitro).¹⁴⁸ Given the expense of rVIIa, in addition to doubts over safety, it would appear to be wise to employ more established measures before its use, along with (as dictated by local agreements) consultation with a hematologist.

Evidence that rVIIa will reduce bleeding comes mainly from case reports. Numerous anecdotes appear to demonstrate dramatic reductions in bleeding in apparently catastrophic situations. Demonstrations of even occasional success, in the face of an otherwise hopeless situation, might itself be taken as a good reason to use the drug. Events are rarely, however, that clear cut and rVIIa is often used on patients at much lower risk of bleeding to death. Reported cases of off-label use in anesthesiology and surgical practice range from truly prophylactic use (in high-risk populations before evidence of severe bleeding), to use to control prolonged or severe (but not immediately life-threatening) bleeding, through to compassionate use in immediately life-threatening bleeding after exhaustion of all conventional treatments. The balance of risk and benefit will vary widely within these differing clinical scenarios.

Most controlled trials of rVIIa have focused on prophylaxis use or use in less severe bleeding. There are 26 randomized controlled trials that have examined off-label use of rVIIa, and 5 have focused on surgical use.^{150,189–192} A meta-analysis of these trials demonstrated no reduction in mortality and only relatively modest reductions in bleeding and transfusion.¹⁹³ A study¹⁹⁰ that examined patients with moderately severe bleeding (more than 200 mL/hr or greater than 2 mL/kg for 2 consecutive hours) following heart surgery in adults reported that although 2619 patients had consented to the trial, only 179 subsequently had sufficient bleeding to be randomized. The results demonstrated a reduction in reexploration, bleeding, and in transfusion requirements. The objective of the trial was to examine the safety of using rVIIa. Unfortunately, the trial was terminated early; although a trend to increased adverse effects was seen, the result was inconclusive. The authors of this study cautioned against wider use before further trials. A single randomized trial (n = 76) that examined prophylactic rVIIa administration (40 µg/kg) in infants undergoing heart surgery,¹⁹¹ demonstrated neither efficacy nor toxicity in this group. The study was small but larger than several of the adult studies that demonstrated an effect. Whether this demonstrates a genuine difference between adults and infants or is related to the relatively low dose rVIIa used is unclear. The use of rVIIa in children has been recently reviewed.¹⁸⁸

Concerns about adverse effects of this drug have grown. Two studies have attempted to systematically review the data on toxicity.^194,195 The rate of thrombosis was greater (10.2% vs. 8.7%) in those treated with rVIIa, but this increase was largely confined to the elderly. The second study examined data from observational trials with comparator groups; mortality was not affected by treatment with rVIIa, although the risk of thrombosis was increased in some groups (including adult cardiac patients). Both these reviews combined data from studies on very different patient populations, and some caution is required in their interpretation. The risk of thrombosis in children is uncertain and data inconclusive.¹⁸⁸ Neonates treated with either FFP or rVIIa for similar indications had a similar incidence of thrombosis (7%).¹⁹⁶ Thrombotic complications occurred in 10.8% of children who received off-label rVIIa; the overall mortality in those receiving rVIIa was 34%.¹⁹⁷ It is not possible to know how this was affected by use of rVIIa. It is clear that both thrombotic complications and mortality are common in patients with bleeding serious enough to warrant use of this drug. Use of rVIIa will very likely increase the risk of thrombosis, although the impact of this (and the benefit of the drug) will vary in different clinical situations.

The dose regimen for off-label use of rVIIa is yet to be firmly established. One of the difficulties is that there is no satisfactory laboratory test to monitor its effectiveness,^198,199 and factor VII activity does not always predict efficacy.²⁰⁰ Although it is known that the PT, aPTT, and TEG improve after rVIIa is given in liver surgery,²⁰¹ they cannot reliably be used to determine a dosing regimen. If the main effect of rVIIa is at the site of injury, clinical observation still remains the best indication of effect.¹⁹⁸ The dose used often closely relates to the dose indicated in hemophilia patients (90 µg/kg), although in practice doses are often rounded to the nearest vial.²⁰¹ Although Warren et al. recommended a lower dose of 40 to 60 µg/kg, a dose of 40 µg/kg has been shown to have no effect.¹⁹¹ Larger doses of 60 to 90 µg/kg may prove more effective.

The justification to use larger doses is supported by differences in PK. In older children with hemophilia, the terminal elimination half-life of rVIIa is less than in adults (1.3 vs. 2.7 hours), and the clearance is greater (67 vs. 37 mL/kg/hour).^202,203 When prophylactic rVIIa at 100 µg/kg was given to 10 preterm infants between 23 and 28 weeks gestational age every 4 hours for 72 hours, two umbilical artery catheter thromboses were reported and no embolic events or increase in intraventricular hemorrhage rate compared with those not given rVIIa.²⁰⁴ There were no other treatment-related side effects, despite this large repeated dose, which suggests that a large study in neonates is warranted.

Since the previous edition of this book, there have been a considerable number of publications related to the use of rVIIa. Unfortunately, information, especially that concerning specific indications, remains sparse. It would appear that rVIIa can be effective for the treatment of severe bleeding refractory to other treatments, but it is certainly not universally effective. The risk of thromboembolic complications is substantial in those recovering from severe bleeding, and rVIIa is likely to increase this risk. There seems to be no reason to think that these considerations do not apply to children, although the balance of risk and benefit may be very different in children. It is the authors’ current view that rVIIa should only be used for treating life-threatening bleeding that has proved refractory to other treatments. Outside of properly conducted trials, it should not be used for prevention of bleeding, treatment of less severe bleeding, or as an alternative to blood products. Its use should not delay surgical reexploration if indicated. Every attempt should be made to ensure an adequate platelet count, adequate fibrinogen concentration, correction of acidosis, and near normothermia before use.