Massive transfusion in children and neonates


  • Yaser A. Diab,

    1. Division of Hematology, Center for Cancer and Blood Disorders, Children's National Medical Center, Washington, DC, USA
    Search for more papers by this author
  • Edward C. C. Wong,

    1. Division of Laboratory Medicine, Center for Cancer and Blood Disorders, Children's National Medical Center, The George Washington University School of Medicine and Health Sciences, Washington, DC, USA
    Search for more papers by this author
  • Naomi L. C. Luban

    Corresponding author
    1. Division of Laboratory Medicine, Center for Cancer and Blood Disorders, Children's National Medical Center, The George Washington University School of Medicine and Health Sciences, Washington, DC, USA
    • Division of Hematology, Center for Cancer and Blood Disorders, Children's National Medical Center, Washington, DC, USA
    Search for more papers by this author

Correspondence: Dr Naomi L. C. Luban, Division of Hematology, Center for Cancer and Blood Disorders, Children's National Medical Center, 111 Michigan Avenue, NW, Washington, DC 20010, USA.



Resuscitation of children and neonates with severe or refractory bleeding due to surgery or trauma often requires massive transfusion (MT). Findings from recent studies have led to a better understanding of the complex pathophysiology in massive haemorrhage and the effects of MT on haemostasis. Current management of the massively bleeding adult patient has evolved over the past few decades, shifting to early transfusion of products in a balanced ratio as part of MT protocols (MTPs). Paediatric data on successful management of MT are limited and the optimal transfusion approach is currently unknown, leading to practice variability among institutions, depending on resource availability and patients' needs. Here, we review new important concepts in the biology of massive bleeding and MT, outline important management principles and current practices, and highlight available relevant adult and paediatric data.

Massive transfusion (MT) is a term that generally describes infusion of blood products in significant volumes over a short time period (Dehmer & Adamson, 2010). Several different definitions of MT for adult patients exist, based on absolute or relative volumes of transfused blood and blood products, and the time frame for these transfusions (Levy, 2006; Malone et al, 2006; Raymer et al, 2012). Three definitions of MT are commonly reported in adult literature (Raymer et al, 2012):

1-Transfusion of ≥10 units of red blood cells (RBCs) within 24 h,

2-Transfusion of >4 units of RBCs in 1 h with anticipation of continued need,

3-Replacement by transfusion of 50% total blood volume (TBV) in 3 h.

A definition of MT in children is not well established and most definitions of MT in adults are not applicable to paediatric patients. The TBV in neonates and children varies considerably according to age (Barcelona et al, 2005) (Fig 1). As the TBV, volume loss, and volume replacement are usually estimated in children in relation to weight, any definition of MT in paediatric patients must be relative to the TBV of each individual patient. Moreover, dynamic definitions of MT that use lower volumes of blood/blood products transfused over shorter time frames are probably more useful clinically to guide management of massively transfused patients (Raymer et al, 2012). Hence, we suggest defining paediatric MT as transfusion of >50% TBV in 3 h, transfusion >100% TBV in 24 h or transfusion support to replace ongoing blood loss of >10% TBV per min. The earliest published report of a paediatric MT involved a male neonate who survived despite inadvertently receiving a single MT (c. 400 ml blood over a short time period <24 h) during the course of treatment for Rh haemolytic disease of the newborn (HDN) (Bloxsom, 1946). Similar to adult patients, MTs are most frequently required for resuscitation of patients with severe haemorrhage from any cause. Common causes of massive haemorrhage in paediatric patients include accidental and non-accidental trauma, major surgeries, particularly cardiac and spinal surgeries, selected neurosurgical procedures and liver transplantation, as well as severe gastrointestinal bleeding. Unexpected serious bleeding may also be encountered in children with unsuspected or undiagnosed bleeding diathesis (examples include spontaneous intracranial haemorrhage in a male neonate with haemophilia and severe postoperative bleeding in a child with von Willebrand disease). Additionally, double-volume exchange transfusion (ET) for the treatment of severe neonatal hyperbilirubinaemia (most often in the setting of HDN) and circuit priming during initiation of extracorporeal membrane oxygenation (ECMO) and cardiopulmonary bypass (CPB) in neonates and small infants often necessitate large volume blood transfusions that could, in effect, be considered ‘massive’ (Luban, 1995). While neonates receiving small volume transfusions over time may, in essence, undergo a full exchange of their TBV over time, such transfusion would not be considered MT. However, feto-maternal haemorrhage, vascular malformations, twin-to-twin transfusion and severe neonatal bleeding both internally, as in sub-galeal bleeding secondary to Ventouse delivery, or externally, as in bleeding from dislodged arterial catheter, can result in profound neonatal anaemia requiring MT. MT can be associated with a number of complications (Table 1) some of which require particular attention in newborns and toddlers who are at particularly increased risk for metabolic complications (Fasano et al, 2012).

Table 1. Transfusion-associated complications in massive transfusion and its management approacha
  1. AHTR, Acute haemolytic transfusion reaction; DHTR, Delayed haemolytic transfusion reaction; FNHTR, Febrile non-haemolytic transfusion reaction; RBCs, red blood cells; TRALI, transfusion-related acute lung injury; TRIM, transfusion-related immunomodulation; TA-MC, transfusion-associated microchimerism; TA-GVHD transfusion-associated graft versus host disease; PTP, post-transfusion purpura; HDN, haemolytic disease of the newborn; ET, exchanged transfusion; TACO, transfusion-associated circulatory overload; VTE, Venous thromboembolism; TANEC, transfusion-associated necrotizing entercolitis; NEC, necrotizing entercolitis.

  2. a

    Data from (Sihler & Napolitano, 2010; Fasano et al, 2012; Meyer & Uhl, 2012).

Transfusion reactions
AHTR/DHTRFNHTR is usually a diagnosis of exclusion
FNHTRAllergic reactions range from mild urticarial reactions without changes in vital signs to anaphylactic reactions (due to anti-IgA antibodies)
Allergic reactionsTransfuse group O, RhD negative leucoreduced RBCs and AB plasma
Immunological complications
TRALIMinimize transfusions once haemorrhage is controlled. Consider using RBCs with a shorter storage time and plasma from men and/or nulliparous women
TRIM/TA-MCMay be responsible for increased risk of bacterial infection
TA-GVHDGamma irradiation of blood products for at risk patients (infants <6 months, suspected/confirmed immunosuppressed patients)
PTPDue to a recent prior transfusion. May be treated with IVIG infusion, steroids plasmaphaeresis
Metabolic complications
HypocalcaemiaDue to citrate overload. Neonates are especially vulnerable due to immature parathyroid hormone response to prolonged hypocalcaemia. Monitor ionized calcium, correct accordingly
HypomagnesaemiaDue to the infusion of large volumes of magnesium-poor fluids and citrate overload. Monitor magnesium, correct accordingly
HyperkalaemiaDue to leakage from prolonged storage or irradiation of RBCs, increased risk in infants and in those with renal dysfunction and cardiac disease Monitor potassium, correct accordingly. Transfusing fresh RBCs (<5–10 d old) irradiated <24 h before issue or washing may minimize risk
HypokalaemiaDue to re-entry into transfused RBCs, release of stress hormones, metabolic alkalosis, or infusion of potassium-poor solutions Monitor potassium, correct accordingly
Metabolic alkalosisDue to citrate overload (citrate is metabolized in liver to bicarbonate). Monitor acid-base balance
Impaired glucose homeostasisNeonates, especially premature infants are at risk for both hyperglycaemia due to increased glucose load and stress reaction, and hypoglycaemia due to rebound hyperinsulinaemia. Neonates with HDN requiring ET are at particularly increased risk for rebound hypoglycaemia. Monitor glucose, correct accordingly if clinically indicated
AcidosisDue to hypoperfusion (primarily), citrate overload and liver dysfunction. Monitor acid-base balance, correct with aggressive resuscitative measures and alkalinizing agents
HypothermiaDue to exposure, infusion of cold fluids/blood products, opening of body cavities, decreased heat production, and impaired thermoregulatory control. Increased risk in neonates and infants due to more permeable skin and poor thermal regulation Warm patient, use blood and fluid warmers with temperature controls
Haemostatic defectsDue to multiple mechanisms leading to a complex coagulopathy. May increase risk of VTE.
InfectionsMaintain high index of suspicion
TACOIncreased risk in infants and in cardiopulmonary disease. Treat with oxygen, diuretics
TANECNEC in premature infants that arises within 48 h of a blood transfusion. Mechanism and definite casualty have not been established. Holding feedings during transfusion may decrease risk
Air embolismRare potentially fatal complication. Manufacturer's instructions should be strictly followed when using rapid infusion devices. Management of suspected cases consists of rapid identification and elimination of source and aggressive resuscitation as needed
Figure 1.

Estimated total blood volume in neonates, children and adults. Data from (Barcelona et al, 2005).

Management of the massively transfused paediatric patient, therefore, requires careful and ongoing close monitoring that is best achieved using a standardized management approach (Table 1). Published data on paediatric MT are limited to case reports, case series, small retrospective studies and single centre experiences. Hence, the optimal strategy for the selection of volumes and types of blood components to be administered and for clinical and laboratory monitoring in paediatric patients requiring MT remains unknown.

Massive transfusion and haemostasis-pathophysiological considerations

Current understanding of the mechanisms underlying haemostatic defects associated with massive haemorrhage and haemodilution is largely derived from animal studies and studies of adult trauma patients (Brohi et al, 2007; Chesebro et al, 2009). Further insights on the effects of the haemodilution associated with MT on haemostasis have been gained from studies of paediatric cardiac surgery and extracorporeal life support (Manno et al, 1991; Kern et al, 1992). It is important to remember that the haemostatic system in neonates and infants younger than 12 months old is physiologically immature and lacking adequate reserve (Andrew et al, 1987; Kuhle et al, 2003). Hence, the haemostatic effects of MT are more profound and the risk of bleeding may be further increased in this patient population (Eaton & Iannoli, 2011).

Patients requiring MT commonly exhibit significant and complex perturbations in haemostasis that is multifactorial and related to the volume and age of transfused blood, pre-existing haemostatic abnormalities, concomitant pathological changes, and additional therapeutic interventions (Reiss, 2000). The underlying mechanisms for these haemostatic defects are not completely understood and have been historically attributed to haemodilution caused by crystalloid fluid resuscitation and packed red blood cell transfusion with insufficient plasma coagulation factor replacement (Hewson et al, 1985). However, subsequent studies in both adult and paediatric trauma victims have demonstrated that a significant proportion of patients have an acute coagulopathy at presentation which is evident prior to instituting fluid resuscitation and transfusion therapy and that this ‘early coagulopathy’ is associated with poor outcome, irrespective of injury severity (Brohi et al, 2003; MacLeod et al, 2003; Hendrickson et al, 2012a; Patregnani et al, 2012). Moreover, it was shown that systemic anticoagulation and hyperfibrinolysis, mediated by hypoperfusion-induced activation of the thrombomodulin-protein C pathway, represent the hallmark haemostatic defects of this early trauma-related coagulopathy in adult trauma patients. The early coagulopathy of trauma is further perpetuated by additional factors present in this context including MT and ultimately leads to a vicious bleeding-coagulopathy cycle (Brohi et al, 2007; Sihler & Napolitano, 2009). While the inciting triggers for these early haemostatic disturbances include systemic tissue hypoperfusion, which follows direct injury to major blood vessels and/or organs, as well as release of tissue factor from injured tissues, subsequent pathological events probably occur almost simultaneously and affect all elements of haemostasis (platelets, procoagulants, natural anticoagulants, profibrinolytics, and antifibrinolytics) (Bolliger et al, 2010a; Sorensen & Fries, 2012) (Fig 2). Although a similar pathophysiological sequence could be operating in paediatric trauma patients, it is important to keep in mind that the aetiologies of trauma in children differ from those observed in adults with more blunt than penetrating injuries. Furthermore, non-accidental injury and birth trauma in neonates represents important causes for blunt head trauma in paediatric patients. Coagulation abnormalities are frequently found in adult and paediatric patients with isolated severe traumatic brain injury, which are also linked to adverse outcomes. It is postulated that local and subsequent systemic release of tissue factor from the injured brain parenchyma resulting in unregulated activation of the extrinsic coagulation pathway and progressing into a consumptive coagulopathy plays a significant role in the pathogenesis of early coagulopathy in this patient population (Lustenberger et al, 2010; Talving et al, 2011). MT in itself can have several deleterious effects on the haemostatic system. These negative effects of MT on haemostasis are dependent on the volume, age (labile coagulation factors V and VIII deteriorate with storage), and rate of infused blood/blood products. Haemodilution is inevitable when giving specific blood component therapy, even with 1:1:1 ratio of RBC:plasma:platelets that have been advocated in adult protocols for MT. This is because the standard process of making these specific components results in a loss of platelets and dilution of all components with preservative. Hence, recombining the components does not result in a product equivalent to whole blood (Sihler & Napolitano, 2010). The critical level of different haemostatic elements occurs at different time points during haemodilution. Fibrinogen deficiency, for example, seems to develop earlier after MT than any other coagulation factor deficiency (Hiippala et al, 1995). Recently, acquired factor XIII deficiency owing to haemodilution was also found to occur early during paediatric major surgery and may play an important role in MT coagulopathy (Haas et al, 2012). Factor XIII activities <60% may increase risk of bleeding (Sorensen & Fries, 2012). Moreover, severe thrombocytopenia (platelet count <50 x 109/l) is observed in most adult patients receiving ≥20 RBC transfusions, while only 3 of 26 paediatric patients developed severe thrombocytopenia and clinical bleeding after MT (Cote et al, 1985; Leslie & Toy, 1991). Plasma levels of the antifibrinolytic proteins α2 antiplasmin and plasminogen activator inhibitor-1 (PAI-1) are also progressively lowered by haemodilution and platelet-derived PAI-1 is also depleted due to dilutional thrombocytopenia while tissue plasminogen activator levels are increased, probably as a result of stress-mediated release from endothelial stores (Emeis, 1992; Bolliger et al, 2010b). Hence, the fibrinolytic activity remains intact during haemodilution and in combination with decreased factor XIII levels, leads to increased susceptibility of fibrin clot to plasmin digestion (Bolliger et al, 2010a).

Figure 2.

Pathophysiology of haemostatic abnormalities in massive transfusion. The haemostatic defects encountered in patients requiring massive transfusion have complex multifactorial pathogenesis. Injury to blood vessels and or organs during surgery or due to trauma leads to haemorrhage and tissue injury. Systemic hypoperfusion due to massive bleeding is associated with thrombomodulin (TM) activation with increased enodothelial expression. Thrombomodulin complexes with thrombin (TH) and the thrombin-thrombomdulin complex then activates protein C. In addition, thrombin bound to TM is no longer available to cleave fibrinogen. Hence, thrombin becomes diverted from its physiological predominantly procoagulant role to a pathological anticoagulant role. Activated protein C inhibits coagulation factors V and VIII and in excess also depletes plasminogen activator inhibitor -1 (PAI-1), reducing tissue plasminogen activator inhibition and accelerating the formation of plasmin. In addition, diversion of thrombin to protein C activation may also reduce activation of thrombin-activatable fibrinolysis inhibitor (TAFI). The end-result is an early coagulopathy characterized by systemic anticoagulation and hyperfibrinolysis. Several additional factors contribute to the progression of this early coagulopathy. Systemic hypoperfusion can cause shock-induced impairment of hepatic function. Massive tissue injury leads to consumptive coagulopathy and, sometimes, disseminated intravascular coagulopathy (DIC)-like syndrome, which are mediated by release of tissue factor (TF). Patients with prolonged hypoxia, hypovolaemia or hypothermia, and those with massive head injury or extensive muscle damage are at increased risk for consumptive coagulopathy. Acute anaemia caused by bleeding can impair platelet adhesion and aggregation by decreasing the number of platelets in the peripheral plasma. Bleeding with loss of coagulation factors and platelets also causes ‘loss coaguloapathy’. Administration of massive transfusions leads to dilutional coagulopathy and dilutional thrombocytopenia and can cause or worsen some of the metabolic derangements (acidosis, hypocalcaemia, hypothermia) observed in this context. Hypothermia, acidosis and hypocalcaemia cause further haemostatic impairment. The development of progressive coagulopathy as a net result creates a pathological viscous circle that usually manifests as the ‘lethal triad’ of refractory coagulopathy, progressive hypothermia, and persistent metabolic acidosis. Data from (Brohi et al, 2007; Chesebro et al, 2009; Sihler & Napolitano, 2009; Lustenberger et al, 2010; Davenport et al, 2011; Sorensen & Fries, 2012).

Additionally, MT at rapid rates or in the presence of shocked liver can lead to significant citrate overload, which in turn can lead to iatrogenic hypocalcaemia and acidosis. Significant hypocalcaemia (ionized calcium <0·6 mmol/l) and acidosis (pH <7·3) are detrimental to normal haemostasis (Martini & Holcomb, 2007; Lier et al, 2008). Lastly, because most blood products are normally stored at 1–6°C, rapid transfusion of large quantities can lead to or exacerbate the hypothermia that is frequently observed in patients requiring MT. For each 1°C drop in temperature, coagulation factor activity decreases by 10% and below 34°C there will be clinically relevant impairment of coagulation with prolongation of clotting times (Watts et al, 1998; Martini, 2007). Furthermore, hypothermia is associated with increased platelet pooling in the spleen and with impairment of platelet adhesion and aggregation at temperatures below 34°C (Watts et al, 1998; Kermode et al, 1999). Conversely, platelet aggregation is initially increased in mild hypothermia (Straub et al, 2007). Finally, while the coagulopathy in massively transfused patients leads to defective haemostasis which, in advanced stages, manifests as brisk microvascular bleeding from the operative field and persistent oozing from puncture sites, it is important to bear in mind that the risk of venous thromboembolism (VTE) after the initial resuscitation and correction of coagulopathy is increased in both adult and paediatric trauma patients (Geerts et al, 1994; Hanson et al, 2010). The mechanism of VTE in these patients is multifactorial and blood product transfusions as well as other haemostatic intervention are potentially one contributing factor (Geerts et al, 1994).

Massive transfusion - current practices and protocols

Massive bleeding

Contemporary management of massive haemorrhage has undergone a paradigm shift with a focus on damage control strategies including abbreviated surgical procedure to control bleeding, permissive hypotension to avoid disruption of thrombus; aggressive control of acidosis, hypocalcaemia, and hypothermia; and limitation of crystalloid products with early administration of plasma (Spinella & Holcomb, 2009). It has also become evident that coordination of all the parts of management of the massively transfused patient is best accomplished by a well defined MT protocol (MTP) (DeLoughery, 2010). The MTP is a comprehensive institutional plan that facilitates communication, ensures frequent laboratory monitoring and reduces delay and errors in ordering and administering blood/blood products. Development and implementation of a MTP requires a multidisciplinary team approach (trauma service, transfusion medicine and laboratory support) taking into account specific patient care needs and available resources (Shaz et al, 2009). Although the majority of trauma centres in the United States have adopted MTPs for management of adult trauma patients, protocols vary from centre to centre (Schuster et al, 2010). Approaches to blood product administration in MTP can be either (i) laboratory test result-based, (ii) predetermined using preset ‘transfusion packages’, or (iii) directed by real-time transfusion service physician (Shaz et al, 2009). Currently, a predetermined ratio approach is favoured because it can mitigate early coagulopathy by administering coagulation factors early in the resuscitation, decrease the amount of crystalloid administered, and remove the delay in ordering, preparing, and subsequent administration of blood products (Holcomb & Hess, 2006; Shaz et al, 2009). While the optimal ratio for transfused products in MTPs has not been well established, most adult MTPs employ a ‘physiological’ 1:1:1 transfusion ratio (i.e.: equal units of RBCs, fresh frozen plasma (FFP), and platelets) for blood component therapy (Schuster et al, 2010) based on retrospective data suggesting improved survival and decreased early hemorrhagic death with a FFP/RBC ratio approaching 1:1 (Borgman et al, 2007; Zink et al, 2009). Age of transfused products may be an important consideration for massively transfused patients. Storage of RBCs is associated with decreased intracellular 2,3-diphosphoglycerate content, decreased deformability, depletion of adenosine triphosphate and decreased nitric oxide availability (d'Almeida et al, 2000; Winslow & Intaglietta, 2008). Hence, aged RBCs may not deliver oxygen to the tissues as well as fresh RBCs (Tinmouth et al, 2006; Winslow & Intaglietta, 2008; Shaz et al, 2009). Furthermore, blood storage is associated with increased concentrations of several cytokines (interleukin (IL) 8, IL1, tumour necrosis factor-α, soluble CD 40L) in vitro, which may be associated with increased risk for inflammatory and immunmodulatory adverse events in MT (Zallen et al, 2000; Cognasse et al, 2006). This might explain the increased morbidity and mortality associated with increased number and storage duration of transfused blood (Zallen et al, 1999; Weinberg et al, 2008). The choice of anticoagulant-preservative solution used for storage of RBCs is another important consideration that is particularly relevant for neonatal MT. Although the concentrations of the additives (i.e.: citrate, phosphate, dextrose, sodium chloride, mannitol and adenine) in the extended-storage solutions (i.e.: additive units) licensed for use in the United States (Adsol, Optisol, and Nutricell) and United Kingdom (SAGM) are safe for most children and neonates receiving small volume transfusions, the risk for metabolic abnormalities due to potentially toxic levels of these additives may be increased in extremely ill premature neonates requiring MT (Luban et al, 1991; Strauss et al, 2000). Because there are no clinical data on the risk of metabolic derangements with large-volume RBC transfusions in neonates, some experts recommend avoiding use of RBCs stored in these extended-storage media for neonatal MT until such data become available (Fasano & Luban, 2008). Alternatively, solutes could be reduced by hard packing RBCs to a haematocrit of 80%, inverted storage, or washing of the RBC product (Fasano & Luban, 2008).

Due to limited paediatric data, evidence-based guidelines for MT in neonates and children with massive bleeding are lacking. Therefore, paediatric practices vary significantly among institutions (Ringer et al, 1998; Laverdiere et al, 2002). Unfortunately, the experience with paediatric-specific MTPs is, in comparison, extremely limited (Barret et al, 1999; Paterson, 2009; Dressler et al, 2010; Pickett & Tripi, 2011; Chidester et al, 2012; Hendrickson et al, 2012b). Moreover, blood banking logistics make incorporating paediatric MTPs challenging for institutions that are not free-standing children's hospitals, meaning that a single ‘adult’ MTP will need to be used for both adult and paediatric patients. However, adult protocols may not be appropriate for children, for whom transfusion practices would need to be weight-based and take into account known developmental haemostatic aspects of paediatric patients (Dehmer & Adamson, 2010). On the other hand, certain blood products, such as pre-thawed plasma, and new haemostatic therapies that are utilized in adult trauma centres are not typically available in most free-standing paediatric trauma centres due to cost-effectiveness considerations.

To date, only two single-centre studies have reported on experience with MTP in paediatric patients (Chidester et al, 2012; Hendrickson et al, 2012b). The study reported by Hendrickson et al (2012b) was a retrospective study in which data on clinical status, resuscitation volumes, and hospital course from 53 children after implementing paediatric MTP that utilized 1:1 ratio of FFP:RBC and alternating weight-adjusted amounts of aphaeresis platelets or cryoprecipitate units were compared with data from historical controls prior to instituting MTP (Hendrickson et al, 2012b). Despite failing to show improvement in outcome with MTP, probably due to small sample size, there were two important findings revealed in this study. First, the majority of patients (80% of controls and 72% of MTP patients) had at least one coagulation value outside of the normal range at presentation. Second, implementation of paediatric MTP with increased FFP: RBC ratios and early plasma transfusion to coagulopathic children is feasible (Hendrickson et al, 2012b). More recently, Chidester et al (2012) conducted a prospective cohort study of 55 children, of whom 22 patients received transfusions according to a MTP; 33 patients received blood at physician discretion. Similar to the study reported by Hendrickson et al (2012b), mortality was not significantly different between the two groups. However, the MTP group received a greater overall amount of blood products and was more likely to be severely injured. Surprisingly, thromboembolic events were observed exclusively in the non-MTP group, which the authors attributed to under-transfusion in those patients (Chidester et al, 2012). Importantly, despite utilizing a MTP, neither study was able to reach the protocol's goal of 1:1 ratio for FFP/RBC transfusions which was the result of delay in transfusing plasma due to additional time that was required for thawing.

At our institution, we have recently incorporated a predetermined MTP (Table 2) that is primarily intended for use in patients with massive traumatic or surgical haemorrhage. Our experience to date is limited because the protocol had been utilized in a small number of patients. Upon activation of the MTP, uncrossmatched group O Rh(D)-negative RBCs are issued if the necessary blood bank data are not available to permit release of type-specific units, according to a standard emergency release procedure. In extreme situations, additional modifications of released units (for example irradiation in a patient who requires irradiated products) may be bypassed at the discretion of the treating attending physician in consultation with transfusion medicine. Blood type (ABO/Rh), cross-match and antibody screen are performed immediately upon receipt of the patient's sample and the patient is switched to type-specific blood products as soon as the patient's ABO/Rh blood group is identified. After the initial emergency release or transfusion of RBCs, packages are issued that contain the designated amount of blood products based on patient weight (Table 2). Packages continue to be prepared in advance, after each previous package is issued, until the MTP is terminated by the treating attending physician. While packages are generally issued from the blood bank in units and not in aliquots, the treating attending physician determines the appropriate volume to be transfused to the patient. For emergency release of uncrossmatched RBCs, we dispense additive units >10 d old and <25 d old, because the use of additive solution units by virtue of having less plasma could potentially prevent transfusion of excess isoagglutinins which can interfere with serological testing. For subsequent RBC transfusions, units not more than 14 d old that are preserved in citrate-phosphate-dextrose-adenine (CPDA)-1 (i.e.: non-additive RBC units), depending on availability, are preferred for neonates and young infants.

Table 2. Children's national paediatric massive transfusion protocol©
Packages to Transfuse (In order below)0–4 kg Neonate (85 ml/kg)a5–9 kg Infant (85 ml/kg)a10–24 kg Young Child (75 ml/kg)a25–49 kg Older Child (70 ml/kg)a≥50 kg Teen/Young Adult (65 ml/kg)a
  1. RBC, Red Blood Cells (in units); FFP, Fresh Frozen Plasma (in units); PLT, Platelet concentrate (in equivalent units); Cryo, Cryoprecipitate (in units).

  2. a

    Formula to calculate total blood volume; # Exceeds total blood volume, no crossmatch required.

Emergency Release (A)½ RBC1 RBC2 RBC3 RBC5 RBC
2 Cryo3 Cryo4 Cryo6 Cryo8 Cryo
2 Cryo3 Cryo4 Cryo6 Cryo8 Cryo
2 Cryo3 Cryo4 Cryo6 Cryo8 Cryo

ECMO and cardiac surgery in neonates and young infants

The best transfusion approach in ECMO and cardiac surgery in neonates and small infants remains unclear (Eaton & Iannoli, 2011; Yuan et al, 2012). Typically, 1–2 units of ABO and Rh group-specific and cross-match-compatible RBCs, and 1 unit of group-specific FFP are used to prime the circuits (Friedman & Montenegro, 2004). Platelet concentrates are also usually included in the CPB prime. For cardiac surgery, additional blood product transfusions are also typically given during and post-bypass to secure haemostasis. RBCs used in these settings should be negative for sickle haemoglobin, because transfused RBCs may be exposed to hypoxia and severely abnormal metabolic conditions (Friedman & Montenegro, 2004). Relatively fresh RBCs should be used for the reasons discussed earlier and to minimize the risk of hyperkalaemia. While the use of fresh RBC units in paediatric cardiac surgery has been associated with improved outcomes (Ranucci et al, 2009; Manlhiot et al, 2012), studies of fresh whole blood (FWB) <48 h old or fresh reconstituted whole blood (FRWB) have reported conflicting results (Mou et al, 2004; Gruenwald et al, 2008). Hence, there currently exists no consensus on the use of FWB or RFWB for cardiac surgery in infants and decisions are often made by individual institutions, based on inventory, the overall activity of cardiothoracic service, and the patient complexity. Although anecdotal evidence suggests that additive RBCs are safe in ECMO, non-additive RBC units are generally recommended (Transfusion Task Force, 2007; Gibson et al, 2004; Luban et al, 1991). On the other hand, there is no consensus regarding the safety of additive RBC units in cardiac surgery (Luban et al, 1991; Transfusion Task Force, 2007). Additional considerations include provision of cytomegalovirus (CMV) safe blood products for low birth weight neonates and for thoracic organ transplant candidates who will be immunosuppressed (Friedman & Montenegro, 2004). If irradiation is required, units should be issued <24 h after irradiation to minimize the risk of hyperkalaemia. In the case of ECMO, extreme urgency may require compromise of the ideal blood product preparation as previously described (Friedman & Montenegro, 2004). While we generally prefer CPDA-1 RBC units for ECMO or cardiac surgery in neonates and infants younger than 4 months, we will dispense the freshest RBC units (<7 d old) available, even if these are additive units. RBCs are provided along with FFP that is used during the pump prime with a haemotocrit determined by the perfusionist and/or cardiovascular surgeon. If irradiation is clinically indicated, we irradiate blood products at time of issue.

Neonatal exchange transfusion

Stored whole blood, if available, or reconstituted whole blood is used for neonatal ET performed in the setting of neonatal hyperbillirubinaemia. For the reasons discussed earlier and to decrease the risk of hyperkalaemia, RBCs chosen for ET should be fresh (preferably <5–7 d) CPDA-1 units (Luban et al, 1991; Gibson et al, 2004). If only older CPDA units or additive units are available, the RBC units should be volume reduced or washed. All components should be CMV safe, irradiated <24 h before issue, and sickle-negative (Fasano & Luban, 2011). If the delivery of an infant with severe HDN is anticipated, then type-O Rh-negative blood cross-matched against the mother may be prepared prior to birth (Fasano & Luban, 2011). Blood prepared after delivery should be negative for the implicated antigen(s) and may be cross-matched against the infant (Fasano & Luban, 2011). In HDN due to ABO incompatibility, the blood must be type-O and Rh-negative or Rh compatible with the mother and infant (Fasano & Luban, 2011). In addition, the blood should have a low isoagglutinin titre, or be washed free of plasma (Fasano & Luban, 2011). Type-O RBCs are often used with AB plasma to ensure that no isoagglutinins are present, but this results in two-donor exposures per ET (Fasano & Luban, 2011). Blood prepared for ET for non-immune indications, such as non-immune hyperbilirubinaemia, drug overdose and sepsis, may be cross-matched against the infant only (Fasano & Luban, 2011). In our institution, for neonatal ET, we provide reconstituted CPDA-1 RBCs with plasma in which the desired haematocrit of the RBC unit is individually determined based on the infant's haematocrit and discussions between the neonatology and transfusion medicine services. We irradiate all cellular blood products intended for neonates weighing <1250 g at the time of issue to mitigate the risk of hyperkalaemia.

Massive transfusion-emerging strategies

Targeted haemostatic therapy based on global point-of care coagulation assays

Laboratory evaluation of massively bleeding patients can be quite challenging and there are no rapidly available and validated coagulation assays that can reliably identify coagulopathies in these patients. Conventional coagulation assays, such as prothrombin time (PT), partial thromboplastin time (PTT), platelet count and fibrinogen levels, are of limited value in this setting (Davenport & Khan, 2011). The results of most coagulation assays are not available in a real-time fashion (Kozek- Langenecker, 2010). Moreover, these assays provide partial information on clot formation and cannot quantify the relative contributions of pro-coagulants versus anticoagulants (Mann et al, 2003; Tripodi et al, 2009). Platelet dysfunction, hyperfibrinolysis and acquired factor XIII deficiency associated with MT are not detectable by any of these assays. Importantly, conventional coagulation screens are poor predictors of the need for MT and have limited ability to direct on-going blood component therapy (Dzik, 2004; Davenport et al, 2011). Point-of-care global haemostatic assays, specifically thromboelastography (TEG) and rotational thrombelastometry (ROTEM), have emerged as a potential alternative to better assess the complex haemostatic disturbances observed in patients requiring MT (Davenport & Khan, 2011). Both TEG and ROTEM assess the viscoelastic properties of blood samples under low shear conditions and the results are shown as a graphical representation of the entire haemostatic process from initiation to fibrinolysis (Fig 3). In addition, several parameters are measured or calculated by the software of the device, which provides a quantitative evaluation of the individual components of the haemostatic process (Fig 3, Table 3). The use of TEG/ROTEM in the setting of MT has multiple advantages. First, the results are available within a short time frame, making them relevant to clinical decision making (Bolliger et al, 2012). Second, hyperfibrinolysis, an important early haemostatic defect, not detected by conventional laboratory testing, can be reliably identified with TEG/ROTEM (Davenport & Khan, 2011). Third, the evaluation of the coagulation system in whole blood allows assessment of the combined influence of circulating plasmatic and cellular (platelets, RBCs, leucocytes) elements on clot formation, including platelet function and the contribution of factor XIII. Fourth, TEG/ROTEM can be performed at the patient's actual body temperature, making these assays more sensitive than conventional coagulation tests, which are performed at a constant temperature of 37°C for detection of haemostatic defects of hypothermia (Kettner et al, 1998). Hence, the use of TEG/ROTEM may provide timely accurate coagulation information to guide blood component therapy and direct resuscitation (Davenport & Khan, 2011). Haemostatic interventions and transfusion algorithms utilizing TEG/ROTEM (Table 3) have been suggested based on non-controlled studies in adults and single centre experiences (Rugeri et al, 2007; Carroll et al, 2009). Paediatric- but not neonatal-specific reference ranges for TEG/ROTEM have now been established (Chan et al, 2007; Oswald et al, 2010). Furthermore, several recent paediatric studies have suggested the potential utility of TEG/ROTEM for guiding transfusions in cardiac surgery and CPB (Moganasundram et al, 2010; Romlin et al, 2011). Nevertheless, the use of these assays in paediatric MT has not yet been examined; hence, the incorporation of these novel techniques in paediatric MTPs and other transfusion protocols requires further validation.

Table 3. Thromboelastography parameters, interpretation of abnormalities and interventions
ParameterDefinitionHaemostatic phaseCauses for abnormalitiesIntervention
  1. R, Reaction Time; CT, Clotting Time; K, Kinetics Time; CFT, Clot Formation Time; α, Alpha Angle; MA, Maximum Amplitude; MCF, Maximum Clot Firmness; LY, Lysis; ML, Maximum Lysis.

RCTTime from the initiation of test till the beginning of the clot formationInitiation of coagulation Prolonged R/CT Plasma (if prolonged R/CT)
- Factor deficiencies
- Anticoagulants
Short R/CT
Plasma hypercoagulability
KCFTTime from the start of the clot formation to the curve reaches amplitude of 20 mmAmplification of coagulation Prolonged K/CFT Cryoprecipitate
- Factor deficiencies
- Hypofibrinogenaemia
- Thrombocytopenia
- Platelet dysfunction
ααAngle between the baseline and the tangent to the curve through the starting point of coagulationPropagation of coagulation ‘Thrombin burst’ Low α Cryoprecipitate
- Factor deficiencies
- Hypofibrinogenaemia
- Thrombocytopenia
- Platelet dysfunction
MAMCFAmplitude measured at maximum curve widthPropagation of coagulation ‘Platelet–fibrin interaction’ Low MA/MCF Platelets
- Hypofibrinogenaemia(consider FXIII concentrate if ongoing bleeding and persistently low MA/MCF)
- Thromboyctopenia
- Platelet dysfunction
- FXIII deficiency
LYMLReduction in area under curve (LY) or in amplitude (ML) from the time MA/MCF is achieved until 30 or 60 min after MA/MCFFibrinolysis Increased LY/ML Antifibrinolytics
- Hyperfibrinolysis
Figure 3.

Thromboelastography tracing with commonly reported parameters in TEG/ROTEM. R, Reaction Time; CT, Clotting Time; K, Kinetics Time; CFT, Clot Formation Time; α, Alpha Angle; MA, Maximum Amplitude; MCF, Maximum Clot Firmness; LY, Lysis; ML, Maximum Lysis.

Novel adjuvant therapeutic interventions

These include antifibrinolytics, desmopressin, local haemostatics and clotting factor concentrates including fibrinogen, prothrombin complex concentrate, recombinant factor VIIa (rFVIIa) and factor XIII. The use of these adjuvant haemostatic agents has not been systematically studied in paediatric patients. Future use of these agents guided by TEG/ROTEM monitoring to avoid potential thromboembolic complications represents an attractive targeted therapeutic strategy.

Administration of systemic antifibrinolytics in children undergoing cardiac surgery, or scoliosis surgery reduces blood loss (Sethna et al, 2005; Pasquali et al, 2012) and in a recent multicentre randomized, placebo-controlled trial, tranexamic acid was found to reduce mortality in adult bleeding patients (Shakur et al, 2010). ROTEM-directed infusion of fibrinogen concentrate in nine children with dilutional coagulopathy from MT undergoing craniosynostosis repair surgery was reported to secure haemostasis without adverse effects in all nine patients (Haas et al, 2008). Administration of a 4-factor prothrombin complex concentrate was reported to result in cessation of intra-abdominal bleeding in a 5-month-old infant with abusive liver trauma that persisted after MT and multiple surgeries (Fuentes-Garcia et al, 2011). Lastly, rFVIIa can improve the coagulopathy in adult patients with post-traumatic haemorrhage and decrease blood transfusion requirements; however, there has been no evidence showing that administration of rFVIIa reduces mortality in this context. Uncontrolled paediatric data suggest that off-label use of rFVIIa in various dosing regimens ranging from 20 to 180 μg/kg could be useful in the management of refractory bleeding in massively transfused children after head trauma, during ECMO support and after cardiac surgery (Morenski et al, 2003; Uhrig et al, 2007; Niebler et al, 2010; Pychynska-Pokorska et al, 2011). While rFVIIa has been included in some MTPs, well-designed clinical trials are needed to better assess the timing, optimal dose, efficacy, and safety (particularly with respect to thrombosis risk) of rFVIIa therapy in the setting of MT.


Management of MT in neonates and children is extremely challenging. Much of what is known about the physiology and management of massively transfused paediatric patients has been extrapolated from adult data. Although the utility of MTPs has been well established in the adult patient population, the optimal strategy for paediatric MT remains unknown. Implementation of institution-specific MTPs may be feasible in large paediatric centres; however, large prospective multicentre studies are needed to evaluate outcomes associated with implementation of paediatric MTPs, especially if adjuvant haemostatic agents are included in the MTP. Individualized targeted haemostatic intervention guided by novel global haemostatic assays is emerging as an alternative promising future approach.


All authors contributed to the manuscript and approved the final version. Naomi L.C. Luban is supported through NIH/NHLBI R01 HL067229 grant. We wish to thank Ross M. Fasano, MD for his comments and revisions of the article and Valli R. Criss MT, ASCP, SBB for providing insight on current blood-bank practices with respect to MT and MTPs.