Transfusion medicine









  • The decision to prescribe the transfusion of any blood product must be based on individualized indications and must take into account specific health problems.



  • Acute severe anemia (hemoglobin concentration <5 g/dL) increases the risk of death in critically ill patients.



  • There is no evidence that a red cell transfusion improves outcomes in stable critically ill children if their hemoglobin concentration is greater than 7.0 g/dL. A hemoglobin concentration greater than 7.0 g/dL may be required in unstable critically ill children and in pediatric intensive care patients with heart disease, particularly those with cyanotic heart disease, but the best threshold is unknown in such patients.



  • Plasma can be useful to treat severe coagulopathy in bleeding patients.



  • Platelets can be useful to treat bleeding caused by low platelet counts and/or platelet dysfunction.



  • In pediatric intensive care units, most transfusion-related adverse events are linked to immune-mediated effects of blood products rather than to transfusion-transmitted infectious diseases.



This chapter reviews the rationale for the transfusion of red blood cells (RBCs), plasma, platelets, whole blood, and cryoprecipitate in pediatric intensive care units (PICUs).


Red blood cells


Native RBCs contain hemoglobin (Hb), which carries oxygen (O 2 ) to cells, thus facilitating efficient adenosine triphosphate (ATP) production and cell survival. Because energy expenditure is high in critically ill patients, an adequate Hb is necessary to deliver enough oxygen to meet metabolic demand. Anemia is observed in up to 74% of critically ill children. The transfusion of RBC units is the only effective way to rapidly increase the Hb level. However, the efficacy and safety of transfused RBCs has been questioned for decades. Infections transmitted by blood products were the most important concern in the 1980s. In the 1990s, noninfectious serious hazards of transfusion (NISHOT), such as transfusion-related immune modulation (TRIM), transfusion-related acute lung injury (TRALI), transfusion-associated circulatory overload (TACO), and multiple-organ dysfunction syndrome (MODS) have become significant concerns.


This section discusses anemia, O 2 delivery (D o 2 ), and O 2 consumption (V o 2 ); reviews evidence on the effectiveness of RBC transfusion in the intensive care setting; discusses the recommendations found in guidelines on RBC transfusion in critically ill children; and reviews the most frequent transfusion reactions and transfusion-related complications.


Red blood cell transfusion: Why and why not


Anemia and oxygen delivery


Oxygen delivery in the critically ill


Variables related to oxygen delivery and consumption, and hemodynamic adaptive mechanisms related to anemia, are explained in more detail in Chapters 23 , 28 , and 34 .


Anemia decreases the capacity of blood to deliver O 2 because of lower Hb content. Systemic (global) D o 2 is dependent on cardiac output and the arterial concentration of O 2 (Ca o 2 ):


<SPAN role=presentation tabIndex=0 id=MathJax-Element-1-Frame class=MathJax style="POSITION: relative" data-mathml='Do2=cardiac output (stroke volume×heart rate)×Cao2′>Do2=cardiac output (stroke volume×heart rate)×Cao2Do2=cardiac output (stroke volume×heart rate)×Cao2
Do2=cardiac output (stroke volume×heart rate)×Cao2


Ca o 2 (mL O 2 /100 mL) is defined by the formula:


<SPAN role=presentation tabIndex=0 id=MathJax-Element-2-Frame class=MathJax style="POSITION: relative" data-mathml='(Hb×Sao2×1.34)+(0.003×Pao2)’>(Hb×Sao2×1.34)+(0.003×Pao2)(Hb×Sao2×1.34)+(0.003×Pao2)
(Hb×Sao2×1.34)+(0.003×Pao2)


where Sa o 2 is arterial oxygen saturation and Pa o 2 is partial pressure of arterial oxygen. Because systemic D o 2 is directly linked to Hb concentration, the most rapid and effective way of increasing D o 2 (within minutes) in anemic patients is by giving an RBC transfusion.


The ultimate goal of RBC transfusion is to improve cellular energy production, that is, ATP production, which translates into V o 2 . An adequate O 2 delivery to cells does not imply necessarily that V o 2 is adequate and that cells produce enough energy. V o 2 depends on substrate availability and metabolic demands; it can be amplified by increasing cellular O 2 extraction rate (O 2 ER), or by increasing D o 2 if there is V o 2 /D o 2 dependence.


The relationship between O 2 delivery and consumption is characterized by two phases: a directly linear relationship between V o 2 and D o 2 up to a “critical threshold” (often referred to as the critical D o 2 ) and a flat section above this threshold ( eFig. 91.1 ). Above this threshold, a fall in D o 2 does not cause a drop in V o 2 because it is compensated by mechanisms such as an increase in O 2 ER. These mechanisms are limited, though, and there is a critical threshold of D o 2 under which O 2 ER cannot increase any further, and below which V o 2 diminishes if D o 2 decreases.




• eFig. 91.1


Relationship between oxygen delivery (D o 2 ) and O 2 consumption (V o 2 ) in normal patients (black line) , between D o 2 and blood lactate level in normal patients (blue dashed line) , and between D o 2 and V o 2 in septic critically ill patients (red dotted line) . Under the critical threshold D o 2 (arrow) , V o 2 diminishes and blood lactate level increases. Over this threshold, a fall in D o 2 does not cause a drop in V o 2 because it is compensated for by an increase in O 2 extraction and/or an increased cardiac index. In critically ill patients, V o 2 is frequently higher than normal. Then, the level of the D o 2 /V o 2 curve (red dotted line) is shifted upward; the critical threshold may also be shifted to the right, which may explain why pathologic O 2 supply dependence is reported in some critically ill patients.


The stress of critical illness increases metabolic rate and V o 2 , which shifts the critical level of D o 2 to the right and upward. However, compensatory mechanisms are limited in anemic critically ill patients.


Adaptive mechanisms with anemia


Anemia significantly decreases blood O 2 -carrying capacity. However, in the normal host, the amount of O 2 delivered to tissues exceeds resting O 2 requirements by a twofold to fourfold factor. While the Hb concentration is dropping, several adaptive processes maintain V o 2 . These processes include (1) increased O 2 ER; (2) increased heart rate and stroke volume, which increase cardiac output; (3) redistribution of blood flow from nonvital organs toward the heart and brain at the expense of O 2 delivery to less important vascular beds, such as the splanchnic vasculature; and (4) left shift of the oxy-Hb-dissociation curve, which decreases affinity between Hb and O 2 , thereby increasing the amount of O 2 released to cells. Moreover, anemia decreases blood viscosity, which, in turn, decreases cardiac afterload and increases cardiac output.


Increasing O 2 ER is an important way to adapt to anemia. The upper range of normal O 2 ER is 30%; O 2 ER increases if O 2 requirements are not met. Higher O 2 ER is frequently observed in severely ill patients, which translates into low central venous O 2 saturation (S cvo 2 ). When maximal O 2 ER is attained and other adaptive mechanisms are overwhelmed, V o 2 /D o 2 dependence appears and may result in O 2 debt, which is associated with mortality. O 2 requirements are increased in patients with sepsis and MODS. Impaired left ventricular function and abnormal regulation of vascular tone can restrict D o 2 and disturb redistribution of regional blood flow. Moreover, the mitochondria of patients with severe sepsis and MODS are dysfunctional and cannot adequately produce ATP. A severe cellular energy crisis may result.


A number of host characteristics specific to children and infants may also impair their adaptive mechanisms. While an increase in cardiac output generally compensates for anemia, this may not occur in the first few weeks of life due to lower myocardial compliance during this period and a significant impairment in diastolic filling, which limits stroke volume increase. In addition, an elevated resting heart rate in newborns also limits their ability to increase cardiac output. Greater energy requirements in young infants and children mostly attributable to growth imply a greater need for substrates, including O 2 . Issues affecting O 2 transport and release in children also include a higher proportion of fetal Hb during the first months of life, which causes a left shift of the oxy-Hb saturation curve, and a physiologic anemia.


Oxygen kinetics in the critically ill


Tissue hypoxia from low D o 2 may be due to low Hb concentration (anemic hypoxia), low cardiac output (stagnant hypoxia), or low Hb saturation (hypoxic hypoxia) or some intoxications (e.g., toxic hypoxia caused by carbon monoxide). RBC transfusions are typically administered to increase D o 2 in critically ill children. RBC transfusion increases D o 2 in the central circulation, but it does not always increase tissue D o 2 and global V o 2 in ICU patients. , Several mechanisms may explain this. Mitochondrial dysfunction can impede O 2 utilization in critically ill patients. Moreover, peripheral D o 2 is impaired in ICU patients, and there is some evidence that RBC transfusion may worsen this problem.


Regulation by red blood cells of oxygen delivery to tissue


While RBC transfusion certainly increases systemic D o 2 in the central circulation, this does not imply that regional D o 2 is improved. There is evidence that RBC transfusion may disturb local D o 2 . However, data on regional D o 2 are inconsistent, their clinical significance remains to be determined, and the underlying mechanisms, which may involve blood viscosity, microcirculatory flow, local D o 2 , and cellular respiration, are not well characterized.


Resistance to blood flow increases rapidly if the hematocrit level exceeds 0.45, which corresponds to an Hb of 15 g/dL. RBC transfusion increases blood viscosity, which can lead to microcirculatory stasis and impaired D o 2 to tissues.


Activation of white blood cells (WBCs) in RBC units and cytokine generation in the supernatant of transfused RBC units may also have a microcirculatory effect: some cytokines can mediate vasoconstriction or thrombosis of small vessels and cause local ischemia. Most RBC units are now prestorage leukocyte reduced, which significantly decreases cytokine levels in the supernatant. The clinical impact of cytokines on regional D o 2 remains to be determined.


There is evidence that RBC transfusion can cause vasoconstriction of microvasculature via an interaction between intracellular Hb and uptake by RBCs of nitric oxide released by blood vessels. With local tissue hypoxia, Hb in the microvasculature releases nitric oxide and triggers local vasodilation. Conversely, if there is sufficient O 2 in the microvasculature, Hb traps nitric oxide and causes vasoconstriction. This regulatory mechanism is almost immediately lost once RBCs are stored: it has been shown that in vitro exposure of blood vessels to RBC units stored 3 hours or more causes vasoconstriction. The clinical significance of these observations is not clear; nonetheless, they suggest that regional D o 2 can be disturbed by RBC transfusion (see also Chapter 87 ).


RBC units undergo several changes during storage, which are generally referred to as storage lesions. , For example, the concentration of 2,3-diphosphoglycerate (2,3-DPG) in stored RBCs decreases over time and can induce a left shift in the oxy-Hb dissociation curve, which impedes O 2 release to tissues even if systemic D o 2 is increased. In addition, RBC deformability decreases after 2 or 3 weeks of storage, which may alter their capacity to pass through the capillary bed. Furthermore, hemolysis in older RBC units releases substantial amounts of free Hb, ranging from 0.5 mg/dL in 1-day-old RBC units to 250 mg/dL in 25-day-old units. Moreover, microparticles are released by RBCs stored more than 2 or 3 weeks. Free intravascular Hb and microparticles tightly bind nitric oxide and therefore cause vasoconstriction.


Thus, while RBC transfusions certainly increase systemic D o 2 , some evidence suggests that impaired microcirculatory flow and O 2 availability can occur, which may have adverse effects on tissue oxygenation and cellular respiration.


Risks of anemia


Severe anemia is associated with a higher risk of mortality. Carson et al. combined the data of two studies involving bloodless surgery in adults and reported that the unadjusted mortality rate was 0.9% when the nadir Hb level was 7 g/dL or greater, 9% if it was between 5 and 7 g/dL, 29.8% if it was between 4 and 5 g/dL, and 41% if it was less than 4 g/dL. Clearly, the risk of mortality increases when the Hb falls below 5 g/dL. It is less obvious when to transfuse if the nadir Hb level is between 5 and 7 g/dL. It is important to recognize that the relationship between anemia and outcomes does not mean that transfusion to treat anemia will necessarily improve outcomes.


There are less data on the relationship between anemia and adverse outcomes in children. Lackritz et al. followed 2433 anemic African children, 20% of whom received a whole-blood transfusion. Transfusion seemed beneficial if the Hb was below 4.7 g/dL and if the patient presented with respiratory distress. Given these data, local guidelines were implemented, recommending that whole blood should be transfused to hospitalized children if their Hb was less than 5 g/dL. Lackritz subsequently enrolled 303 children with a Hb less than 5 g/dL. Mostly because blood products were not available, 116 (38%) did not receive a transfusion. Each child with severe anemia was matched with the next child hospitalized with a Hb greater than 5 g/dL. Death rates were 19.5% in 303 patients with a Hb greater than 5 g/dL who were not transfused, 21.4% in 187 patients with a Hb less than 5 g/dL who were transfused, and 41.4% in 116 patients with a Hb less than 5 g/dL who were not transfused. English et al. completed a prospective cohort study of 1269 children with malaria hospitalized in Kenya. Whole-blood transfusion seemed to decrease mortality if anemia was severe (Hb <4 g/dL), or if a Hb less than 5 g/dL was associated with dyspnea. These studies suggest that the risk of mortality increases significantly among children ill enough to require hospitalization if their Hb is less than 5 g/dL, particularly if respiratory symptoms are present.


Transfusion of red blood cells: Indications (when)


Optimal RBC transfusion practice in PICU patients remains unclear. The number of RBC transfusions increased in children younger than 18 years between 1994 and 2014. It is estimated that 24.8 million RBC units were distributed by blood banks in the United States in 2015 and pediatric transfusions comprised 5.2% of all RBC transfusions. In the PICU, 17% to 76% of patients received RBC transfusions. , Decisions for all of these transfusions were made in the context of limited evidence to guide clinical practice. Despite this, some recommendations regarding RBC transfusion in PICU patients can be found in the medical literature.


A multidisciplinary panel of 42 international experts, the Pediatric Critical Care Transfusion and Anemia Expertise Initiative (TAXI), developed evidence-based and expert consensus-based recommendations on RBC transfusion for critically ill children, using the RAND/UCLA (University of California, Los Angeles) methodology. The experts focused on the following specific populations of PICU patients: life-threatening and non-life-threatening bleeding, respiratory failure, nonhemorrhagic shock, acute brain injury, acquired/congenital heart disease, sickle cell/oncology/transplant, extracorporeal membrane oxygenation (ECMO)/ventricular assist device (VAD)/renal replacement therapy, and other patients. Members of the TAXI consensus conference agreed on 97 recommendations and a decision tree. , The following recommendations refer to five nodes that were added to the decision tree in Fig. 91.2 .




  • Hemorrhagic shock or life-threatening bleeding (node #1): For life-threatening hemorrhage, RBC transfusions are clearly indicated, with plasma and platelets (ratio: 1/1/1 or 2/1/1) (see also Chapter 118 ). ,




• Fig. 91.2


Pediatric critical care Transfusion and Anemia Expertise Initiative (TAXI) red blood cell (RBC) transfusion clinical decision tree. Three different recommendations can be found at the end of each branch of the tree. (1) An RBC transfusion is probably useless (pink boxes); (2) an RBC transfusion can be useful or is mandatory (yellow or green boxes); (3) the right decision is unclear (white boxes). The pink circles show the numbers of five nodes. ECMO, Extracorporeal membrane oxygenation; Hb, hemoglobin; PARDS, pediatric acute respiratory distress syndrome; VAD, ventricular assist device.

(From Valentine SL, Bateman ST, Bembea MM, et al. Consensus recommendations for red blood cell transfusion practice in critically ill children from the Pediatric Critical Care Transfusion and Anemia Expertise Initiative [TAXI]. Pediatr Crit Care Med 2018;19[9]:884–898.)


Hemorrhagic shock or severe bleeding (node 1)


The first node of the decision tree is about hemorrhagic shock and life-threatening bleeding. In such instances, RBC transfusions are clearly indicated, with plasma and platelets (ratio: 1/1/1 or 2/1/1). , The decision to prescribe RBC transfusions should be based on physiologic state, estimated amount of blood loss (Canadian Blood Services recommends a transfusion if there is an acute blood loss >15% blood volume ) and risk of ongoing life-threatening hemorrhage, more than on Hb concentration, coagulation tests, and platelet counts. A systematic approach to treat hemorrhagic shock and to activate massive transfusion protocols is recommended.




  • Severe anemia in PICU patients without severe bleeding (node 2): The TAXI recommendation is that a RBC transfusion should be prescribed in PICU patients with Hb concentration less than 5 g/dL, and that practitioners prescribe RBC transfusion using their experience and clinical data, such as severity of illness and evidence of decreased O 2 consumption if the Hb concentration is between 5 and 7 g/dL. It is recommended to move to node 3 in critically ill children with Hb level of 7 g/dL or greater.



PICU patients with severe anemia without severe bleeding (node 2)


RBC transfusion is more questionable if there is no severe bleeding. Pediatric intensivists have stated in two surveys that their decision to prescribe an RBC transfusion would be based on reasons such as low D o 2 or V o 2 , cardiovascular insufficiency, respiratory failure or use of certain specific technologies such as ECMO, hemodialysis, hemofiltration, plasmapheresis, or exchange transfusion. Nonetheless, the most frequent reason to transfuse RBCs was reported to be a low Hb concentration, , which is the core of node 2 in the decision tree. The Hb level that should prompt a pediatric intensivist to prescribe RBC transfusion remains a matter of debate, but there is some evidence in the medical literature that can guide practitioners.


Hemoglobin less than 5 or between 5 and 7 g/dL (node 2)


There are data showing that the risk of mortality increases rapidly if the Hb concentration of hospitalized adults or children drops below 5 g/dL. There are strong data suggesting that whole-blood transfusion (with an equivalent of 1 RBC unit within it) might prevent death in severely ill children with an Hb level less than 5 g/dL. , The TAXI recommendation is that an RBC transfusion should be prescribed in PICU patients with Hb concentration less than 5 g/dL.


RBC transfusion in PICU patients is strongly associated with a low Hb level (multivariate odds ratio (OR) if Hb less than 7 g/dL: OR, 61.3; 95% CI, 27.75–134.7). However, what should drive RBC transfusion in critically ill children with Hb between 5 and 7 g/dL is unclear. There are no hard data on the association of Hb between 5 and 7 g/dL and adverse outcomes, and there are no data on the usefulness of RBC transfusion in this population. TAXI recommended that practitioners use their clinical judgment and prescribe RBC transfusion using their experience and clinical data, such as severity of illness and evidence of decreased O 2 consumption.




  • Hemodynamically unstable patients (node 3): The recommendation of TAXI is to use clinical judgment in hemodynamically unstable patients , and to move to node 4 in hemodynamically stable critically ill children (hemodynamically stable = mean arterial pressure not <2 standard deviations below normal mean for age and cardiovascular support—pressors/inotropes and fluids—not increased for at least 2 hours).



Hemodynamically unstable patients without serious bleeding (node 3)


In PICU patients without a life-threatening bleeding and with an Hb level of 7 g/dL or greater, members of TAXI suggested to check what is the hemodynamic status (node 3 in the decision tree). In the Transfusion Requirement in PICU (TRIPICU) trial, a child was considered hemodynamically stable if the mean arterial pressure was not less than 2 standard deviations below normal mean for age and if the cardiovascular support (vasopressors/inotropes and fluids) had not been increased in the last 2 hours before an RBC was considered. Members of TAXI considered a patient to be hemodynamically unstable if the patient does not meet this definition.


There are few hard data in hemodynamically unstable patients. However, a large randomized controlled trial and a meta-analysis of all available evidence reported that RBC transfusions do not improve the outcome of critically ill septic adults even if they are experiencing an uncontrolled shock state. , The recommendation of TAXI is to use clinical judgment in hemodynamically unstable patients. ,




  • Hemodynamically stable noncardiac patients (node 4): Specific recommendations for goal-directed transfusion therapy remain undetermined at the present time. , TAXI members concluded that an RBC transfusion should not be given in hemodynamically stable noncardiac PICU patients with an Hb level of 7 g/dL or greater if they belong to any of the following subpopulations: septic patients, noncardiac postsurgery, severely burned children, trauma patients, respiratory failure (excluding severe pediatric acute respiratory distress syndrome [ARDS]), nonhemorrhagic shock, non-life-threatening hemorrhage, and ECMO, VAD, or RRT. TAXI recommended to move to node 5 in cardiac patients.



Hemodynamically stable noncardiac patients (node 4)


What should prompt intensivists to prescribe an RBC transfusion in hemodynamically stable noncardiac PICU patients with an Hb level of 7 g/dL or greater? Some high-quality studies have addressed this question. Two randomized controlled trials evaluated RBC transfusion strategies in children. The first study was performed in 106 African children with malaria crisis (hematocrit 12%–17%). Whole-blood transfusion did not improve mortality rate (1 in 53 vs. 2 in 53) in patients without respiratory or cardiovascular compromise. In the TRIPICU study, a randomized controlled trial involving 637 stable critically ill children with an Hb level of 9.5 g/dL or less, 320 patients were allocated to an RBC transfusion threshold of 7 g/dL of Hb (restrictive group) and 317 to a threshold of 9.5 g/dL (liberal group). A statistically significant noninferiority was found: 38 and 39 patients, respectively, developed new or progressive MODS, and there were 14 deaths within 28 days postrandomization in both strategy groups. The conclusion was that a restrictive strategy is as safe as a liberal strategy in stable critically ill children. Moreover, 174 patients (54%) in the restrictive group received no RBC transfusion compared with 7 (2%) in the liberal group ( P < .0001), and patients in the restrictive group received 54% fewer RBC transfusions. These data and the primum non nocere principle support adopting a restrictive transfusion strategy for stabilized critically ill children.


Is there anything else than the Hb level that should guide RBC transfusion strategy in hemodynamically stable PICU patients with an Hb level of 7 g/dL or greater? Guidelines from many organizations emphasize that the decision to administer RBCs should not be based solely on Hb levels but should involve sound clinical judgment and good common sense. , ,


Should we use physiologic markers to guide RBC transfusion in PICU patients? Goal-directed transfusion therapy using such markers is frequently advocated. While it is theoretically rational to base the decision to transfuse RBCs on physiologic need, it is still a matter of debate as to what parameters best determine that need. It has been suggested that an RBC transfusion is indicated for patients with symptomatic anemia, but most critically ill children are unable to report these symptoms. Some have suggested that systemic markers of oxygenation deficit, such as systemic V o 2 , V o 2 /D o 2 dependence, blood lactate level, S cvo 2 , mixed venous O 2 saturation or O 2 ER, might be useful. Others have proposed that parameters reflecting regional oxygenation deficit—such as brain tissue O 2 pressure, gastric tonometry, tissue O 2 saturation measured by near-infrared spectroscopy, or digital O 2 ER measured by noninvasive devices—may be more reliable. Actually, it is presently not known which markers are best suited to guide RBC transfusion therapy and what cutoff values should be used to determine the need for RBC transfusion. The concept of goal-directed transfusion therapy is laudable but is presently vaguely defined and not supported by hard data. Specific recommendations for goal-directed transfusion therapy remain undetermined at the present time. ,


In practice, a low Hb concentration remains the most frequent and primary justification for pediatric intensivists to prescribe an RBC transfusion. Therefore, it makes sense that the Hb concentration be the first parameter assessed when an RBC transfusion is considered. Given the available evidence, critically ill children must receive an RBC transfusion if their Hb level is below 5 g/dL. In stable patients—including septic patients, patients having undergone noncardiac surgery, severely burned children, and trauma patients —it is suggested to consider RBC transfusion if the Hb concentration is less than 7 g/dL, but a transfusion is not recommended if the Hb concentration is above this threshold. However, determinants other than the Hb concentration must be considered, including age, severity of illness, or evidence of organ dysfunction or O 2 dependency, such as elevated blood lactate level or low S cvo 2 . For example, it would seem appropriate to consider a higher threshold and a more aggressive RBC transfusion strategy in unstable patients for whom the optimal and safe lower limit of the transfusion threshold has not been established. Moreover, any recommendations made must also factor in specific considerations for disorders such as sickle cell disease and cardiac conditions (node 5).


In summary, the TAXI recommendation in noncardiac PICU patients is “In critically ill children or those at risk for critical illness, who are hemodynamically stable and who have an Hb concentration ≥7 g/dL, we recommend not administering a RBC transfusion.” Can we apply this recommendation to all noncardiac PICU patients, whatever their basic disease?


TAXI members concluded that an RBC transfusion should not be given in hemodynamically stable noncardiac PICU patients with an Hb level of 7 g/dL or greater if they belong to any of the following subpopulations: noncardiac postsurgery, respiratory failure (excluding severe pediatric ARDS), nonhemorrhagic shock, non-life-threatening hemorrhage, ECMO, VAD, or RRT.


Should PICU patients with respiratory dysfunction receive more RBC transfusions? In the TRIPICU study, 234 and 246 patients with respiratory dysfunction were enrolled in the restrictive and liberal group (threshold Hb for RBC transfusion: 7.0 and 9.5 g/dL, respectively). The number of cases of new/progressive MODS was similar, 33 and 35, respectively. Duration of mechanical ventilation was also comparable in these 480 patients (6.4 ± 6.0 vs. 6.3 ± 5.3 days; absolute risk difference: –0.16; 95% CI, –1.2 to 0.9; P = .75), including 73 with an acute lung injury (ALI; 7.2 ± 6.5 vs. 7.1 ± 6.2 days; absolute risk difference: –0.12; 95% CI, –3.1 to 2.9; P = .94) and 48 with ARDS (10.5 ± 9.2 vs. 8.5 ± 7.2 days; absolute risk difference: –2.0; 95% CI, –6.8 to 2.8, P = .40). Thus, targeting higher Hb thresholds does not improve outcome in stable critically ill children with respiratory problems if their Hb level is greater than 7 g/dL.


There is little evidence supporting the belief that a higher Hb level is required in severely ill septic patients. The outcome of 137 septic children (34 in septic shock, 31 with severe sepsis) allocated in the TRIPICU study to a restrictive or liberal transfusion strategy was similar, with 13 of 69 versus 13 of 68, respectively, developing new/progressive MODS (absolute risk difference: 0.3%; 95% CI, –12% to 14%). Holst randomized 1005 adults with septic shock to receive an RBC transfusion only if their Hb level fell below 7 or 9 g/dL. Some of these patients were hemodynamically unstable. The 90-day mortality was similar (43.0% vs. 45.0%; relative risk: 0.94; 95% CI, 0.78–1.09; P = .44). Thus, a threshold Hb of 7 g/dL can probably be safely applied in critically ill children with sepsis, severe sepsis, and septic shock.


The optimal Hb level to trigger RBC transfusion in brain-injured patients and oncologic patients has not been defined yet. Members of TAXI recommended to consider giving an RBC transfusion in critically ill children with acute brain injury if the Hb level is 10 g/dL or less and in patients with oncologic diagnoses or with stem cell transplant if their Hb level is 8 g/dL or less.


TAXI suggests using clinical judgment in PICU patients with alloimmune or autoimmune anemia, with severe pediatric ARDS.


RBC exposure is very large in children under ECMO or VAD. This practice is not without risks. , TAXI suggests using clinical judgment in children under ECMO or VAD.




  • Cardiac patients (node 5): TAXI suggests a postoperative Hb level of 7 g/dL in pediatric cardiac patients when there is good postoperative cardiac function in the absence of persisting cyanotic heart disease and a postoperative Hb level of 9 g/dL in PICU patients with persisting cyanotic heart disease. TAXI suggests using clinical judgment in PICU patients with acquired myocardial dysfunction or pulmonary hypertension. The experts added, “there is no evidence that transfusion above 10 g/dL is beneficial.” ,



Cardiac patients (node 5)


Critically ill children with cardiac disease receive more RBC transfusions than other PICU patients. , There is no evidence indicating that this strategy improves outcomes. On the contrary, anecdotal experience with bloodless cardiac surgery for congenital heart disease in children whose families refuse transfusion suggests that a lower Hb level may be well tolerated. , eTable 91.1 summarizes the available data on this question.



eTABLE 91.1

Postoperative Mortality in Randomized Controlled Trials Comparing Restrictive and Liberal Red Blood Cell Transfusion Strategies in Children With Congenital Cardiopathy






























































Study Cardiac Physiology Threshold Hemoglobin (g/dL) Transfusion Strategy (Deaths/Patients) Mortality
Restrictive Liberal
Willems (2010) Biventricular 7.0/9.5 2/63 2/62 28-day
de Gast-Bakker (2013) Biventricular 8.0/10.8 0/53 0/54 In-hospital
Cholette Biventricular 7.0/9.5 1/53 1/52 In-hospital
TOTAL Biventricular 3/169 3/166
Cholette (2011) Univentricular 9.0/13.0 0/30 1/30 Pediatric intensive care unit
Cholette (2017) Univentricular 9.0/12.0 5/29 6/28 In-hospital
TOTAL Univentricular 5/59 7/58


Biventricular (noncyanotic) cardiac physiology


The results of three randomized controlled trials suggest that using a 7 g/dL threshold is safe in the postoperative care of noncyanotic congenital heart disease in stabilized patients older than 28 days. Willems et al. analyzed a subgroup of 125 postoperative cardiac patients enrolled in the TRIPICU study after cardiac surgery. No significant difference in the incidence of new or progressive MODS (12.7% vs. 6.5%; P = .36), PICU length of stay (7.0 ± 5.0 vs. 7.4 ± 6.4 days) or 28-day mortality (2 vs. 2 deaths) was found between the restrictive and liberal groups. De Gast-Bakker et al. compared outcome in pediatric cardiac surgery patients aged more than 6 weeks allocated to receive an RBC transfusion if their Hb level dropped below 8.0 g/dL (restrictive) or 10.8 g/dL (liberal group). Patients with cyanotic cardiac disease were excluded. Randomization occurred before surgery. With respect to RBC transfusion, the research protocol was initiated in the operating room and maintained up to PICU discharge. In the 107 patients enrolled and retained for analysis, duration of mechanical ventilation, PICU length of stay, and the incidence of adverse events were similar in both groups while hospital length of stay was shorter in the restrictive group (median = 8 and interquartile range = 7–11 vs. 9 and 7–14 days, P = .063). Cholette et al. enrolled 53 and 52 children with biventricular physiology after their cardiac surgery; 53 were allocated in a restrictive and 52 in a liberal RBC transfusion strategy. Despite lower-threshold Hb concentrations in the restrictive group (7.0 vs. 9.5 g/dL), lactate, arteriovenous O 2 difference, and clinical outcomes were similar. The British Society of Haematology and TAXI supports the acceptance of a postoperative Hb level of 7 g/dL in children when there is good postoperative cardiac function in the absence of persisting cyanotic heart disease.


Univentricular (cyanotic) physiology


Two randomized controlled trials have addressed RBC transfusion in cyanotic heart disease. Cholette et al. randomized 30 subjects to a restrictive and 30 subjects to a liberal RBC transfusion strategy (threshold Hb: 9.0 and 13.0 g/dL, respectively). No differences between groups in mean lactate were found (1.4 ± 0.5 vs. 1.4 ± 0.4 mmol/L) or peak (3.1 ± 1.5 vs. 3.2 ± 1.3 mmol/L). As well, no differences were found in C(a−v) o 2 , C(a−c) o 2 , or clinical outcome measures. Cholette et al. also randomized 57 children after bidirectional Glenn or Fontan procedures: 29 patients were allocated to a restrictive and 28 patients to a liberal transfusion strategy (respective threshold Hb: 9 and 12 g/dL). No differences were noted with regard to peak blood lactate level (3.0 ± 1.5 vs. 3.1 ± 1.3 mmol/L), ventilator support, duration of vasoactive agent administration, ICU or hospital length of stay, or survival. More data are required before a restrictive transfusion strategy can be safely implemented in patients with cyanotic heart disease. TAXI recommends a postoperative Hb level of 9 g/dL in PICU patients with persisting cyanotic heart disease.


Other cardiac patients


TAXI suggests using clinical judgment in PICU patients with acquired myocardial dysfunction or pulmonary hypertension. The experts added, “there is no evidence that transfusion above 10 g/dL is beneficial.” ,


Prevention of anemia and red blood cell transfusion


“Bloodless medicine” is a popular concept; it refers to all strategies that can be used to provide medical care without allogeneic RBC transfusion, including blood conservation. Many strategies can prevent or significantly decrease the number of RBC transfusions and exposure to transfusion. These strategies are integrated into the concept of a “patient blood management program,” which includes measures to prevent anemia and to limit the number of transfusions. , , , Adopting a restrictive RBC transfusion policy in stable critically ill children is one of these strategies. Other strategies could include raising the Hb concentration before elective surgery, using blood products only when necessary, limiting blood loss, and administering the patient’s own blood.


Pre–pediatric intensive care unit anemia


Bloodless medicine begins before surgery, when applicable. In the preoperative period, the use of erythropoietin and iron supplementation can be considered to optimize the preoperative Hb level. Collection of autologous donations can take place to minimize or prevent allogeneic transfusion. Medications that increase the risk of bleeding should be avoided, including herbal medicine (garlic, ginseng, ginger, etc.), and optimal control of any existing coagulation disorders should be attained just before surgery.


During surgery, maximal attention should be given to limiting blood loss and ensuring good hemostasis and rapid control of any bleeding. In some instances, desmopressin, fibrin sealants, or antifibrinolytic agents such as tranexamic acid may be used to prevent or control hemorrhage. Recombinant activated factor VII (rFVIIa) use is advocated by some practitioners, but it is associated with a significant risk of thrombosis. The cost/benefit ratio of rFVIIa in children is not well evaluated and its use should be limited to situations involving uncontrolled life-threatening bleeding. The use of prothrombin complex and fibrinogen concentrates are also potential agents that can improve hemostasis in children with significant bleeding, but the appropriate indications for these agents and their safety profile in children are not well established (see Chapter 89 ). The safety and cost-effectiveness of intraoperative blood conservation strategies—such as normovolemic hemodilution, autologous blood cell salvage modalities, intraoperative autotransfusion, and deliberate hypotension—remains to be determined.


Pediatric intensive care unit–associated anemia


Postoperative and PICU management of anemia and bleeding is also important. A restrictive transfusion strategy is in line with the concept of “permissive anemia” supported by the British Committee for Standards in Haematology Transfusion Task Force. A prospective study reported that 73% of blood loss in the PICU is attributable to blood draws. The number and frequency of blood tests must be limited, and the amount of blood collected reduced. Many devices can help to minimize blood loss, including the use of loop sampling, pediatric blood collection tubes, microanalysis techniques requiring small volumes of blood, and in-line measurement of parameters such as blood gases and Hb concentration. Some of these blood conservation methods have been demonstrated to be effective in reducing RBC transfusion in critically ill children.


Erythropoietin response to anemia is blunted and poorer than expected in critically ill patients. In spite of this, erythropoietin can prevent anemia in critically ill adults, in low-birth-weight preterm infants, and in the postoperative care of neonates. In critically ill children, there are no data to support the use of erythropoietin as a preventive measure because most RBC transfusions are administered within 2 or 3 days after PICU admission, , a period of time too short to allow for a response to erythropoietin, which generally occurs after several days. Standard use of erythropoietin is presently not recommended in the PICU.


Iron supplementation is also not indicated in critically ill patients: it had no discernible effect on iron-deficient erythropoiesis, Hb concentration, and RBC transfusion requirement, and can even be harmful.


Post–pediatric intensive care unit anemia


Given the results of many trials advocating for restrictive transfusion therapy, one can expect a high incidence of anemia at PICU discharge. Jutras et al. reported that 47% of 2073 PICU survivors who never received an RBC transfusion while in the PICU were found to be anemic prior to PICU discharge (anemia was defined as per the Canadian Blood Services diagnostic criteria). Demaret et al. reported a prevalence of 57.4% of anemia at PICU discharge in 679 PICU survivors. Such anemia is not without consequences. Yakymenko et al. reported that post-ICU anemia significantly decreased quality of life among critically ill adult survivors. Their randomized controlled trial also showed that RBC transfusion improved the quality of life of anemic post-ICU survivors. Consequences of anemia at PICU discharge, how it should be prevented, and how it should be handled post-PICU remain to be determined.


Types of red blood cell units


Standard red blood cell units


Storage of RBC units is made possible by refrigeration at about 2°C to 6°C and by storage in preservative anticoagulant solutions that contain dextrose, sodium citrate, citric acid, and sodium diphosphate. RBCs use dextrose and phosphate to generate ATP, which is essential for their survival. Citrate blocks coagulation by chelating calcium; it is also transformed into bicarbonate, which stabilizes the stored RBC unit pH above 6.4. RBCs in CPDA-1 (citrate-phosphate-dextrose-adenine) solution can be stored up to 35 days because the level of ATP is normal after 21 days of storage and is about 50% after 35 days. RBC units are prepared by removing 200 to 250 mL of plasma and platelets from one unit of whole blood after centrifugation. To support the nutrient needs of RBCs after plasma is removed, additive solutions were developed, such as AS-1 (Adsol), AS-3 (Nutricel) and AS-5 (Optisol), and saline-adenine-glucose (SAG) or SAG-mannitol (SAGM). These additive solutions further decrease RBC lysis and allow for storage up to 42 days.


The volume of a CPDA-1 unit is about 250 mL, including 63 mL of preservative solution; it must be diluted with 75 mL of NaCl 0.9% before transfusion (final volume: 325 mL). Volumes of an AS-3 unit and a SAGM unit are about 350 mL and 300 mL, respectively; both include approximately 100 mL of preservative solution and do not require any dilution.


Other types of red blood cell units


Many other types of RBC units are available: collected by apheresis, leukocyte-reduced, washed, irradiated, cytomegalovirus (CMV) negative, autologous, and directed.


Leukocyte-reduced red blood cell units


RBC units contain some nonviable platelets, small amounts of coagulation factors, and WBCs, which can release pro- and antiinflammatory mediators during storage. Prestorage leukocyte reduction is a standard procedure for all blood components in many countries, such as Australia, Canada, and the United Kingdom. It can decrease the number of WBCs in RBC units from 1 × 10 9 to less than 1 × 10 6 per product and reduce the concentration of cytokines in the supernatant as well as some T cell–regulated immunomodulation. In 2015, 90.9% of RBC units administered in US PICUs were leukocyte-reduced at collection. Transmission of intracellular viruses, such as CMV and herpes simplex virus, is less frequent if there are fewer leukocytes. Clinical trials in adults indicate that leukoreduction does not reduce mortality or ALI; however, it may decrease fever episodes and reduce infections and antibiotic use after RBC transfusion. Leukoreduction was associated with improvement in several clinical outcomes in premature infants (bronchopulmonary dysplasia, retinopathy of prematurity, necrotizing enterocolitis, and grade 3 or 4 intraventricular hemorrhage). Trials have not been performed in children to determine whether leukoreduction improves clinical outcomes.


Washed red blood cell units


RBCs can be washed with sterile saline; the process removes not only 98% of plasma but also up to 20% of RBCs. RBC washing increases the hematocrit to 0.8, but the procedure takes between 45 minutes and 2 hours; thus, it is impractical to use washed RBC units on an emergency basis unless they are prepared in advance. Washed units must be used within 24 hours after processing. Multiple wash cycles are required. Washing effectively reduces supernatant potassium, immunoglobulin A (IgA), cytokines, complement proteins and microparticles. The procedure may not completely remove all proteins and therefore does not prevent hypersensitivity reactions (e.g., hypersensitivity to IgA). The overall volume of a washed RBC unit is significantly decreased (from 350 to 200 mL for an AS-3 RBC unit), making it sometimes useful to limit the volume administered in some patients. Washed RBC units should not be considered leukocyte reduced.


A randomized controlled trial that included 161 children reported that washing RBCs transfused in cardiac PICU patients reduced postoperative inflammation and number of transfusions. A randomized trial of washed RBC and platelet transfusions in 43 adults with acute leukemia reported similar trends. However, the washing process is not without consequences; for example, it increases hemolysis and decreases RBC 2,3-DPG content. Further studies are required before washing of all RBC units can be recommended for PICU patients.


Washed RBC units can be considered for patients with severe, recurrent allergic reactions to blood, patients with anti-IgA deficiency, and some patients with a high risk for circulatory overload.


Irradiated red blood cell units


Some WBCs remain in RBC and platelet units, even in prestorage leukocyte-reduced units. The objective of irradiation is to induce enough DNA damage to prevent leukocyte proliferation. Irradiation destroys the ability of transfused lymphocytes to divide and therefore to respond to host foreign antigens, thereby decreasing the risk of developing transfusion-associated graft-versus-host disease (TA-GVHD) in susceptible recipients. However, irradiation is not without drawbacks. For example, it can damage the RBC membrane, causing the release of significant amounts of free Hb and potassium. Moreover, the shelf life of irradiated RBC units is reduced from 42 to 28 days.


In the United States, 20.6% of RBC units are irradiated before transfusion. Irradiated RBCs are indicated for all children with congenital or acquired cellular immunodeficiency (e.g., allogeneic stem cell transplant recipients, certain hematologic malignancies, myeloablative chemotherapy recipients) to prevent TA-GVHD. Because DiGeorge syndrome is not rare among infants with congenital heart disease undergoing cardiac surgery, these patients should receive irradiated units. Irradiation is also indicated for patients receiving directed donations from family members. Irradiation is not mandatory for most solid tumors, routine immunosuppressive therapy (e.g., corticosteroids), solid organ transplants, nonmyeloablative chemotherapy recipients, or humoral immunodeficiency.


Cytomegalovirus-negative red blood cell units


RBC transfusion can transmit CMV infection. A large proportion (30%–70%) of blood donors are CMV positive. Although it would be ideal to administer only CMV-negative RBC units to CMV-negative patients, the high prevalence of CMV infection among donors does not permit this. Nevertheless, prestorage leukocyte reduction of blood products decreases transmission of CMV to 1% to 2% (similar to the rate of infection following the transfusion of CMV-negative units) compared with standard products for which transmission is 13% to 37%.


Administration of a CMV-positive RBC unit is generally not an issue for immunocompetent patients. Established indications for CMV-negative units include CMV-negative recipients of organ or bone marrow transplants from CMV-negative donors, CMV-negative bone marrow transplant recipients, and intrauterine transfusions. Less well-established indications include CMV-negative patients who are potential candidates for autologous or allogeneic bone marrow transplant, CMV-negative patients undergoing splenectomy, potential seronegative donors for bone marrow transplant, and CMV-negative patients with HIV.


Directed red blood cell units


Directed blood is donated by family members or friends. Parents frequently believe that giving their own blood decreases the risks of transfusion, which, in practice, is not the case. A small increase of transfusion-transmitted infectious diseases has been reported. Moreover, the risk of contracting a TA-GVHD is increased even in immunocompetent patients. In spite of this, directed blood donation remains popular in the lay public; good clinical studies to better estimate the risk/benefit ratio of this practice are warranted. All directed RBC units must be irradiated pretransfusion.


Autologous red blood cell units


Older healthy children can donate their own blood a few weeks before elective surgery. It is frequently believed that autologous RBC units are free of risk, which is untrue. Moreover, autologous RBC units are not leukocyte reduced, at least in Canada. The risk/benefit ratio of autologous RBC units remains to be determined.


Transfusion of red blood cells: How


RBC transfusion is the best way to rapidly increase the Hb concentration. Once a practitioner decides that an RBC transfusion is warranted, several issues need to be considered, including the type of RBC unit (see previous section), blood type, volume and rate of transfusion, and the monitoring required. Many guidelines have been published on how transfusions should be given. , , ,


Blood types


Table 91.2 describes the compatibility of different blood products. A complete cross-match is mandatory before any transfusion is given, with few exceptions. Transfusion of group O Rh-negative RBCs can be life-saving, but this must be reserved for very severe and acute situations. It takes 15 to 20 minutes to type a patient for ABO and Rh. If there are no RBC antibodies, fully compatible blood or immediate spin cross-match may be issued quickly. In the case of RBC antibodies or other anomalies, a lengthier full serologic cross-match is required. The risk of severe reaction to typed, but not cross-matched, RBC units is about 1 in 1000 if the patient has never received a transfusion; the risk is decreased by 10-fold if a cross-match is performed. Repeat verification that the correct blood unit has been delivered to a given patient is essential because blood mismatch is the most important cause of severe transfusion reaction.



TABLE 91.2

Compatibility of Blood Products
























































Product Receiver Donor
RBC unit and whole blood A A, O
B B, O
O O
AB AB, A, B, O
Rh + Rh + or Rh
Rh Rh
Plasma or platelets A A, AB
B B, AB
O O, A, B, AB
AB AB
Platelets Rh + Rh + or Rh
Rh Rh or Rh + a

RBC, Red blood cells.

a Give an anti-D vaccine (Win Rho) if the receiver is Rh and the platelet concentrate is Rh + .



For high-risk elective procedures, type and cross-matching, completed prior to any bleeding, allows for compatible RBC units to be reserved. If an unforeseen emergency transfusion is required for an actively bleeding patient, it is impossible to deliver RBC units that are typed and cross-matched within a reasonable time frame, in which case group O Rh-negative RBC units can be administered (some hospitals also deliver group O Rh-positive RBC units for men and for women after menopause). STAT ordering a transfusion means that transfusion is required within a few minutes; this is not indicated unless the patient is actively bleeding.


Volume and number of red blood cell units


Prescribing the right volume of RBC units is important, as this prevents cardiac overload and limits exposure to several donors.


An easy, albeit simplistic, rule of thumb exists suggesting that administration of 10 mL/kg of RBC units increases the Hb level by 2 to 3 g/dL. If the volume prescribed is greater than the volume of 1 unit of RBCs, blood should be transfused 1 unit at a time to minimize exposure to multiple donors. Prior to the administration of additional RBCs, the Hb concentration should be measured after allowing at least 30 minutes posttransfusion for Hb and hematocrit values to equilibrate. The transfusion can be completed with another unit or partial unit if a reasonable Hb level is not attained.


If the volume of RBCs required is less than 1 unit, a partial unit can be given. Whole RBC units can be subdivided in half (standard division) or in small pediatric 75-mL transfer packs (Pedi-Pak). Partial units prepared nonsterilely expire 24 hours after preparation. On the other hand, partial units prepared sterilely can be kept as long as the original unit (up to 42 days for AS–3). A small volume of RBC unit placed in a syringe must always be administered within 24 hours.


Length of storage


Regulatory agencies and scientific societies, such as the US Food and Drug Administration (FDA) and the European counterpart to the FDA, mandate that RBC units can be stored up to 42 days based on the criterion that at least 75% of transfused RBCs will be alive 24 hours posttransfusion and that hemolysis will be less than 1%. However, a storage lesion, detailed previously, occurs over time, which raises many concerns. ,


There are safety concerns for transfusion of fresh RBCs as well, including the risk of transmission of some infectious diseases (malaria, Chagas disease, intracellular viruses), cell-free DNA, microchimerism, TA-GVHD, and risk of in-hospital mortality.


Five randomized controlled trials addressed the question of whether the transfusion of fresher rather than older RBC units improves the outcome of critically ill patients. All reported that fresher RBC units do not improve outcome; the three larger trials rather showed consistent, but not statistically significant, trends against the use of fresher blood. , , Many intensivists believe that children should receive fresher RBC units, especially for pediatric cardiac surgery patients. A large randomized controlled trial, the Age of Blood in Children in PICU study, enrolling 1538 transfused PICU patients to address this question. The use of fresh compared with standard-issue RBCs (median length of storage, 5 vs 18 days) did not reduce the incidence of new or progressive MODS (147 vs. 133) or PICU mortality (33 vs. 26). Presently, there is no justification to require fresh RBC units for critically ill children, even for cardiac surgery patients.


There are less data informing us on the effect on outcomes with the transfusion of old RBC units, such as those stored for 5 weeks or longer, but available evidence suggests that transfusion of blood stored for 35 days or longer has no effect on in-hospital mortality. ,


Perfusion, warming, and filtration


An RBC transfusion must be completed within 4 to 6 hours (4 hours in the United States) after the unit is delivered by the hospital blood bank. An RBC unit is usually given over 1 to 3 hours but can be given more slowly (up to 4 hours) or divided into two transfusions if there is some risk of cardiac overload. RBCs can also be given through a rapid infuser in a few minutes for life-threatening bleeding.


The viscosity of RBC units is high, which implies that it is preferable to use larger-bore needles to administer them. No drugs should be given in the line used for RBC unit perfusion. There is in theory a risk of hemolysis if RBCs are given with 0.2% saline or dextrose 10%, but the coinfusion of RBCs and hypotonic or hypertonic solution seems safe if it is short lasting (a few minutes). On the other hand, coinfusion of dextrose-containing fluids with RBC units is safe. It is inappropriate to mix RBC units with Ringer’s lactate (risk of coagulation) or calcium salts.


RBC units must be warmed before administration to diminish the viscosity of the blood product and to avoid hypothermia. Blood viscosity decreases by about 7% for each 1°C increase, reducing resistance and making it easier to administer blood products through catheters. RBC units are stored at 2°C to 6°C and could cause significant hypothermia if given to a patient without warming. Blood products are warmed to room temperature (about 20°C) before delivery to the bedside unless the RBC units are required on an emergency basis (e.g., in a case of hemorrhagic shock). Warming to body temperature (37°C) may be required for patients weighing less than 10 kg or if large volumes need to be given (>20%–30% of the circulating blood volume). RBC units can be warmed by ambient temperature or by active warming with a device. Standard blood-warming devices are used to raise the temperature of whole blood or RBC units and not microwave ovens, which can cause severe hemolysis.


Because all RBC units (even prestorage leukocyte-reduced RBC units) contain fibrin, platelets, and WBC aggregates, a filter (with 80-, 179- or 260-micropores) must always be used to filter these aggregates before they are administered. Although their cost-usefulness has not been determined, filters with 20- to 40-micropores are more effective, and some evidence suggests that they can prevent cases of TRALI. Filters with smaller micropores are not considered standard treatment.


Plasma


Types of plasma


Plasma is separated from RBCs after collection of whole blood or collected using an apheresis machine. It is then frozen for storage to preserve coagulation factors. The term fresh frozen plasma (FFP) is used if the unit is refrigerated within 8 hours of collection; the term frozen plasma (FP) is used if refrigerated within 24 hours of collection. There is a slight reduction in factor VIII levels in FP. In clinical practice, however, these two types of plasma are presumed to be essentially interchangeable. FP (frozen plasma) is used in this section to designate both types of plasma units.


FP units are collected from a single donor, while units of solvent detergent (SD) plasma (Octaplas, Octapharma) are constituted from a pool of FP collected from approximately 630 to 1520 donors and processed using SD for inactivation of lipid-enveloped viruses. In many countries, only FP may be available, but FFP is still available in some countries.


FP units are systematically leukocyte reduced by filtration before storage in many countries, but not in the United States. FP volume is about 200 to 250 mL/U, while the volume of SD plasma is about 200 mL/U. On average, FP contains 1 U/mL of all coagulation factors, but there is significant variability among individual units, which is attributable to biological variation in factor levels among individual donors and differences in processing, storage, and preparation for transfusion. The levels of coagulation factors in SD plasma are more standard with little variation among units, as it is a pooled plasma product. FP is stored at –18°C up to 1 year after collection. SD plasma can be stored for up to 48 months. Compared with FFP, SD plasma is practically cell free and contains significantly lower residual platelet concentrations and negligible amounts of microparticles due to its manufacturing process.


Multiple dried plasma products are currently used in a few European countries to allow for immediate availability of plasma for severely bleeding patients. Dried plasma products maintain hemostatic and endothelial repair properties similar to that of FP products. , The licensing of dried plasma for adults and children is anticipated soon in the United States.


Epidemiologic studies reported that adults transfused with plasma are more prone to contract ventilator-associated pneumonia, bloodstream infection, and septic shock (relative risk for all infections, 2.99; 95% CI, 2.28–3.93; P = .02). A prospective cohort study involving 831 critically ill children reported an increased incidence rate of morbidity and nosocomial infections. The adjusted odds ratios were 3.2 (95% CI, 1.6–6.6) and 2.3 (95% CI, 1.0–5.3), respectively. There was also a significant difference in the adjusted length of PICU stay. Plasma transfusions are associated with adverse clinical outcomes; this might be attributable to their immunomodulative properties or due to confounding by indication given that plasma is transfused to patients with increased severity of illness. It also might be that the association between plasma and outcomes is due to severity of illness and not due to the plasma.


In a secondary analysis of a prospective observational study that included 443 critically ill children who received plasma transfusions, FFP and SD plasma were associated with similar reductions in INR ( P = .80). However, ICU mortality was significantly lower with SD plasma compared with FFP (14.5% vs. 29.1%, P = .02).


Transfusion of plasma: Indications (when)


Generally, FP is transfused to correct multiple coagulation factor deficiencies (or single-factor deficiencies when no recombinant or plasma-derived coagulation factor concentrates are available) in patients with active bleeding or prior to invasive procedures when no alternative therapies are available or appropriate. Common coagulopathies for which FP may be given include liver disease and symptomatic disseminated intravascular coagulation (DIC). FP can be given for the emergency reversal of warfarin or vitamin K deficiency when prothrombin complex concentrates are not available. FP is also given to prevent bleeding in patients with abnormal coagulation tests. Guidelines suggest transfusing FP only when the INR, prothrombin time (PT), or activated partial thromboplastin time (aPTT) is more than 1.5 times normal, as coagulation factors are generally adequate for hemostasis below this level. However, recent data suggest that FP is not very effective at normalizing mild abnormalities of coagulation tests, and the potential clinical benefit of FP transfusion seems minimal when the INR is less than 2.5 or an aPTT is less than 60. Plasma is also indicated for plasma exchange therapies for certain conditions, such as thrombotic thrombocytopenic purpura.


FP should be administered empirically during massive transfusion for life-threatening bleeding. A randomized controlled trial in adults revealed that the early use of FP in the prehospital setting for severe traumatic bleeding improved 24- and 30-day survival. Some experts advocate early replacement of FP in a 1:1 or 1:2 ratio with RBC units in trauma patients with massive transfusion. , , According to two systematic reviews, application of a thromboelastography (TEG or rotational thromboelastometry [ROTEM])–guided transfusion strategy reduces the amount of bleeding in patients with massive transfusion, blood product utilization, and renal failure, and seems to improve morbidity and mortality. , However, these results were primarily based on trials of elective cardiac surgery involving cardiopulmonary bypass, and the level of evidence was low.


Many pediatric intensivists state that abnormal coagulation tests (INR, TEG, and so on) could prompt them to prescribe FP even if the patient is not bleeding. Actually, there is no evidence that FP should be given in prophylaxis to nonbleeding patients. For decades, there has been consensus that plasma should not be used as a volume expander; crystalloids, synthetic colloids, or purified human albumin solutions are preferred. This concept is currently being challenged by some who are considering reexamination of the use of plasma for volume resuscitation of patients who are in hypovolemic or septic shock. In vitro and animal data in septic models support the concept that plasma improves preload more effectively than crystalloids, can correct both hyper- and hypocoagulable states, repair endothelial injury, and is, in fact, less inflammatory than crystalloids. Until trials are performed in humans, plasma should not be routinely used for volume resuscitation even in patients with hypovolemic shock.


Epidemiologic studies reported that adults transfused with plasma are more prone to contract ventilator-associated pneumonia, bloodstream infection and septic shock (relative risk for all infections: 2.99; 95% CI, 2.28–3.93; P = .02). A prospective cohort study involving 831 critically ill children reported an increased incidence rate of morbidity, nosocomial infections, and duration of stay associated with FP transfusion.


In a secondary analysis of a prospective, observational study that included 443 critically ill children who received plasma transfusions, FFP and SD plasma were associated with similar reductions in INR ( P = 0.80). However, ICU mortality was significantly lower with SD plasma compared with FFP (14.5% vs. 29.1%, P = .02).


Transfusion of plasma: How


The typical dose of plasma in nonmassively bleeding children ranges between 10 and 15 mL/kg. The effectiveness of FP transfusion should be estimated by clinical judgment of ongoing bleeding and functional measurements of hemostasis when time allows. For life-threatening bleeding, FP is empirically transfused without guidance from laboratory measurements of hemostasis due to the inherent delay in reporting of these parameters. In patients with non–life-threatening bleeding, FP may be indicated by functional measures of hemostasis that indicate a hypocoagulable state. Karam et al. reported that a dose-response relationship was found in patients only with an INR greater than 2.5; patients with a lower INR had no significant change in coagulation tests. Normalization of coagulation tests often does not occur and therefore should not be used as the only guide for additional FP transfusions.


FP transfusions should be ABO compatible but do not require cross-matching. In contrast to RBCs, group AB–positive plasma is the universal plasma donor and can be given in emergency situations when a blood group is not available. When group AB plasma is not available, evidence data in adults indicate that it is safe to use group A plasma for emergency release for patients with life-threatening bleeding. , Cross-matching is not required, as FP units are screened for antibodies against non-ABO and Rh antibodies, which may cause hemolytic reactions.


FP must be thawed prior to transfusion (20–30 minutes with the water bath method). Thawing FP can be shortened to 7 minutes using microwave ovens specifically designed for this task. A thawed unit of FP is ideally transfused within 4 hours, but thawed plasma can be relabeled and stored for up to 5 days (American Association of Blood Banks [AABB] Technical Manual). The clinical indications for thawed plasma are similar to FP but there is some decrease in the labile coagulation factors (factor V and VIII), and reduction in thrombin generation capacity. An 80- or 170-micropore filter must be used.


Platelets


Types of platelet products


Standard platelet products


Different methods can be used to obtain platelet products: 13% are derived from whole blood, either by the platelet-rich plasma (United States and United Kingdom) or the buffy-coat method (Europe and Canada), and 87% are obtained by apheresis (single-donor) platelets. For platelets derived from whole blood, platelet concentrates are often pooled (4–6 units) for a single platelet transfusion.


Platelet units can be stored either warm (20°C–24°C) for 5 to 7 days or cold (2°C–6°C) for 3 days. The benefit of warm storage of platelets is increased circulation time, which is advantageous for patients who need platelets for prophylaxis of bleeding. The risk of warm storage is increased incidence of bacterial contamination. The benefit of cold storage of platelets is reduced bacterial contamination and increased hemostatic activity, which is advantageous for patients with active bleeding. The disadvantage of cold storage of platelets currently is the short shelf life of 3 days. Trials are in progress examining the effect of platelet storage temperature and duration of storage (up to 14 days) on hemostasis (NCT02495506; https://clinicaltrials.gov ).


Endogenous platelets also have immune functions and can influence both innate and adaptive immune response, which may affect outcomes in critically ill patients. The effect of platelet manufacturing methods (collection, processing, and storage) on hemostatic and immune function deserves thorough examination.


Special platelet concentrates


Leukocyte-reduced platelets


A platelet concentrate must contain less than 8.3 × 10 5 WBCs to be labeled leukocyte reduced. Prestorage leukocyte reduction is a standard procedure in many countries. Bedside leukocyte reduction filter should not be used when prestorage leukocyte reduction is done because it is useless and can decrease the number of platelets.


Irradiated platelets


The risk of TA-GVHD is increased in patients who receive human leukocyte antigen (HLA)–compatible platelets. Therefore, pretransfusion irradiation is mandatory for all HLA-compatible platelet concentrates. Irradiation is also recommended for intrauterine transfusion and infants at risk of TA-GVHD. In 2017, 80% of transfused platelets were irradiated.


Cytomegalovirus-negative platelets


Platelet concentrates can transmit CMV. The indications for CMV-negative platelets and RBCs are similar.


Cryopreserved platelets


While cryopreservation of platelets increases product shelf life to 2 years, the storage constraints (–80°C) and thawing protocols may not be conducive to use in all centers. , With respect to in vivo studies, there have been a limited number of trials assessing safety and efficacy of cryopreserved platelets.


Pathogen-reduced platelets


Globally, multiple technologies are licensed that prevent the infectious potential from parasites, bacteria, and enveloped viruses in platelet units. The increased safety of pathogen-reduced platelets may come at a cost of hemostatic efficacy, as has been reported in children. Large multicenter trials are ongoing that will assess efficacy and safety of pathogen-reduced platelets in critically ill children (NCT02549222: https://clinicaltrials.gov ).


Platelet transfusion (why)


Over 1.5 million platelet products are transfused in the United States each year. Between 3.3% and 7.1% of critically ill children receive at least one platelet transfusion while in PICU. , Approximately two-thirds of all platelet transfusions to PICU patients are for prophylaxis of bleeding in the presence of thrombocytopenia, with an average pretransfusion platelet count of 32 ± 27 × 10 9 /L (median, 21). The remaining one-third of platelet transfusions in critically ill children are for active bleeding, with an average pretransfusion platelet count of 76 ± 39 × 10 9 /L (median, 72).


Most platelet transfusions are administered to increase a low platelet count. The prevalence of thrombocytopenia, defined by a platelet count less than 150,000/uL (<150 × 10 9 /L), is 17.3% on admission into the PICU; 25.3% of children are thrombocytopenic at some point during their PICU stay. Thrombocytopenia arises from decreased platelet production, increased platelet destruction, and dilutional or distributional causes. In the PICU, most thrombocytopenia is caused by chemotherapy, sepsis, DIC, MODS, or hemolytic uremic syndrome. However, heparin-induced thrombocytopenia, massive transfusion, and reactive hemophagocytic syndrome are not so rare. In critically ill children, thrombocytopenia at PICU entry is associated with increased mortality (17.6% vs. 2.5%), bleeding complications, thrombosis, and prolonged PICU and hospital length of stay.


Pediatric subjects are at higher risk of bleeding over a wide range of platelet counts, indicating that their excess bleeding risk may be because of factors other than platelet counts. Platelet dysfunction might be the problem, at least in some instances. In PICU patients, platelet dysfunction is typically caused by specific treatments (cardiopulmonary bypass, hypothermia, pentastarch, and hetastarch) or antiplatelet drugs (low-dose aspirin, nonsteroidal antiinflammatory drugs) but it can also be due to hereditary disease (e.g., Bernard-Soulier disease).


Transfusion of platelets: Indications (when)


Platelet transfusions are indicated for the prevention or treatment of bleeding in patients with thrombocytopenia or platelet dysfunction. As platelet transfusions will result in only modest elevations for 1 to 3 days in patients with persistent thrombocytopenia, the purpose of platelet therapy is not to eliminate all bleeding but rather to prevent or stop major hemorrhagic events. Platelets are contraindicated in patients with thrombotic thrombocytopenic purpura and heparin-induced thrombocytopenia because of the increased thrombotic risk.


Therapeutic platelet transfusions are given to treat clinically significant bleeding associated with a low platelet count or platelet dysfunction. There is evidence that correction of thrombocytopenia reduces mortality of critically ill patients, , but a platelet transfusion should be considered only if the platelet count in an actively bleeding patient is less than 50,000/uL. However, caregivers must remember that “the principal treatment of ICU-associated thrombocytopenia is to treat the underlying disease.” For children with life-threatening bleeding, platelets should be transfused empirically in a ratio between 1:2 to 1:1 (ratio of whole blood–derived platelet unit to RBC unit) until the risk of bleeding resolves. When apheresis platelets are transfused, the range of the ratio transfused is 1:5 to 1:10 (apheresis platelet units to RBC units). This recommendation is based on randomized controlled trials in adults and expert opinion.


More than 50% of platelet transfusions in PICUs are given to prevent bleeding. There is insufficient evidence to support a particular threshold for prophylactic platelet transfusion in children. Most recommendations come from guidelines developed for adults, based on expert opinion. For patients with hypoproliferative thrombocytopenia (e.g., chemotherapy induced), a platelet transfusion threshold of 10,000/uL (10 × 10 9 /L) is recommended. This is based on two clinical trials in adults: Slichter reported no increases in bleeding rates when comparing platelet transfusion thresholds of 10,000 versus 20,000/uL; on the other hand, Stanworth et al. reported more bleeding in patients with hematologic cancer and a platelet count less than 10,000/uL who did not receive prophylactic platelet transfusion versus those who did. On the other hand, Curley et al. reported a significantly higher rate of death or major bleeding among preterm infants who received platelet transfusions at a platelet-count threshold of 50,000/uL rather than 25,000/uL.


In the ICU, higher platelet transfusion thresholds are usually employed. Intensivists generally prescribe platelets for patients on ECMO if their platelet count is less than 100,000/uL (10 × 10 9 /L). The same threshold is frequently used when a central nervous system procedure is undertaken. When the platelet count is less than 50,000/uL, platelets are usually given just prior to an invasive procedure (surgery, insertion of central venous catheter, and so on); platelet transfusion can also be considered in mechanically ventilated patients because the risk of pulmonary hemorrhage is significant. A threshold of 50,000/uL is recommended for lumbar puncture. Increasing the threshold count that would trigger a platelet transfusion may be appropriate if platelets are dysfunctional.


The capacity of platelets to stop a hemorrhage is not only related to their number but also to their function. Many tests can be used to estimate platelet function, such as TEG and devices or systems such as the Sonoclot coagulation analyzer (Sienco, Inc.), the Plateletworks analyzer (Helena Laboratories), the hemostatus platelet function test, the platelet function analyzer, and VerifyNow (Ultegra, Instrumental Laboratories) system. However, the results of these tests are not available on an emergency basis in most hospitals. Measurement of platelet mass may be a practical alternative. There is evidence that larger platelets exhibit increased hemostatic activity. The results of a randomized controlled trial conducted in a neonatal ICU suggested that using platelet mass (platelet count × mean platelet volume) rather than platelet count alone to trigger a platelet transfusion may reduce the number of transfusions. Nevertheless, it must be emphasized that the clinical usefulness of all of these tests remains undetermined in the PICU.


There is significant diversity in the stated practice pattern with respect to platelet transfusion. Most guidelines are based on expert opinion, not on quality data. More research is required to refine the indications for platelet transfusion in critically ill children.


Transfusion of platelets: How


There are about 3 × 10 11 platelets in a unit of apheresis platelets. A simple rule of thumb suggests giving 1 or 2 platelet units per 10 kg, but not more than 6 units per transfusion. For children weighing less than 10 kg, the platelet dose can be 5 to 10 mL/kg of pooled or apheresis platelet unit. Transfusion of one platelet concentrate per square meter (m 2 ) of body surface generally increases the platelet count by 7000 to 11,000/uL. By body weight, the administration of one unit per 10 kg should increase the platelet count by 30,000 to 50,000/uL. However, a clinical trial in which platelet dose was estimated by body surface area evaluated low-dose (1.1 × 10 12 platelets/m 2 ), medium-dose (2.2 × 10 12 /m 2 ), and high-dose (3.3 × 10 12 /m 2 ) prophylactic platelet transfusions and did not find any differences in bleeding among adult or pediatric patients. Platelets must be used within 4 hours after delivery from the blood bank, but there is some evidence that the platelet count increases more if the unit is given within 1 hour. A filter with 80- or 170-micropores must be used to remove aggregates that can form between harvesting and transfusion.


The volume of a whole blood–derived platelet unit is about 50 mL while that of an apheresis platelet unit ranges from 200 to 300 mL; 90% to 95% of this volume is plasma. It is important to note that this plasma is not an adequate source of coagulation factors because factor concentration drops rapidly during platelet storage at room temperature (20°C–24°C). Platelet units can be volume reduced (removal of plasma) prior to transfusion, but this process can decrease the platelet count by 15% to 20%, shortens the storage time to 4 hours, and delays platelet release from the blood bank by approximately 1 hour. Volume reduction should be considered only if there is a risk of severe circulatory overload but is not recommended as a standard procedure because it can activate platelets. In 2017, 8% of platelet transfusions were volume reduced.


Unlike RBC units, ABO compatibility is not mandatory with platelets. While it is possibly better to use ABO-compatible platelet units if inventory allows, no differences were seen in the incremental change in platelet count or in transfusion reactions when comparing major ABO-incompatible platelet transfusions with ABO-compatible transfusions in a large study of critically ill children. In addition, Rh compatibility is desired because all platelet units contain some RBCs. Transfusion of an Rh-positive unit to an Rh-negative patient can cause Rh alloimmunization; an anti-D immunoglobulin (e.g., Win Rho SDF, Saol Therapeutics) should be administered within 48 hours to prevent this complication when Rh-positive platelets are given to an Rh-negative patient, particularly if female.


The percent platelet increment (difference between post- and pretransfusion platelet count) should be higher than 20% if the dose is adequate and if a platelet count is performed 10 to 60 minutes posttransfusion; it should be higher than 10% if measured 18 to 24 hours posttransfusion. Platelet refractoriness can occur due to nonimmune factors such as DIC, in which platelet consumption can be high, acquired hemophagocytic syndrome, drugs such as amphotericin or heparin, or immune factors involving antiplatelet antibodies. Treatment of the underlying problem is mandatory in such instances (e.g., discontinuing all heparin administration in heparin-induced thrombocytopenia). Patients with anti-IgA antibodies should receive washed platelets or platelets collected from IgA-deficient donors.


Whole blood


Whole blood: Type of product


The volume of a typical whole blood unit is about 450 to 500 mL. It is stored at 2°C to 6°C, with storage duration between 21 and 35 days. Some centers limit the storage duration to 14 to 21 days to maximize platelet and RBC function when whole blood is used for only life-threatening hemorrhage. Whole blood can be leukoreduced prior to storage with a platelet-sparing filter. This filter has minimal effects on reducing platelet count and impairing hemostatic function. Whole blood can also be pathogen reduced, which has been shown to decrease rates of malaria transmission in a randomized controlled trial in Africa. Pathogen-reduced whole blood is not licensed in the United States, Canada, or Europe.


Whole blood: Indications


The use of whole blood has been advocated mostly for first-line therapy in hemorrhagic shock, as it contains a balanced proportion of RBCs, coagulation factors, and platelets. , Additional benefits of whole blood include platelets stored at 2°C to 6°C, which increases their hemostatic function and reduces the risk of bacterial infection, reduced donor exposures, and the logistic advantage of using one product compared to three. The use of whole blood should be considered only for patients with hemorrhagic shock. Approximately 15% of children’s hospitals in the United States provide whole blood to children, predominantly for cardiac surgery, liver transplantation, and massive transfusion protocols. Whole blood units are also used for some neonatal exchange transfusions.


Barriers to whole-blood availability have included the need to provide ABO-specific whole blood. This limited its availability due to the waste that would occur if each ABO type was maintained in the inventory of the hospital blood bank. To address this issue, the use of low-titer group O whole blood (LTOWB) has been adopted based on data showing that the benefits of immediate availability of LTOWB to a patient with severe bleeding outweighs any small risk in non–group O recipients. , The risk of transfusing group O whole blood to a non-O recipient is based on the concern that anti-A or anti-B antibodies will either cause hemolysis or endothelial injury. This risk is not different than when non-ABO-matched platelets are transfused or when type A plasma is used in group B or AB severely bleeding patients. In recognition of the data supporting the safety of LTOWB, the AABB has changed its standards to allow this product to be used for patients with life-threatening hemorrhage. Since this standard change, many trauma centers have implemented LTOWB for patients with severe bleeding. Pediatric centers have also begun using LTOWB for children with massive bleeding.


Whole blood: How


Whole blood can be transfused in a rapid infuser and blood warmer for life-threatening hemorrhage. It can be given via an intraosseous line and should be transfused with a standard micropore filter. The IgM titer of anti-A and anti-B for LTOWB can range from 1:50 to 1:256.


Cryoprecipitate


Cryoprecipitate is a concentrated source of fibrinogen, factor VIII, factor XIII, and von Willebrand factor. It is stored at a minimum of –18°C for a maximum of 1 year. Its indication is for fibrinogen replacement. It is used in patients with congenital or acquired hypofibrinogenemia and patients with coagulopathy and significant bleeding, mostly in the context of cardiac surgery and/or DIC in the setting of sepsis. It can also be used in patients with von Willebrand disease and hemophilia A if specific factor concentrates are not available. The dose is 1 or 2 U/10 kg (maximum, 12 U), which should increase fibrinogen level about 60 to 100 mg/dL. A blood filter (80 or 180 micropore) should be used. The recommended rate of delivery is 30 minutes (maximum, 4 hours).


Transfusions reactions and complications


Transfusion reactions


Transfusion of labile blood products (RBC, plasma, platelets) can cause immediate or delayed transfusion reactions and various complications. By definition, immediate reactions usually occur during the transfusion or within 6 hours after the end of the transfusion; delayed reactions can occur after a few days, weeks, or even months or years later. The incidences of transfusion reactions (adults and children) and transfusion-transmitted infectious diseases are reported in Tables 91.3 and 91.4 , respectively. Even though transfusion reactions are underdiagnosed in critically ill children, many large-scale studies reported that the risk of transfusion reactions signaled to hemovigilance systems is significantly higher in children than in adults. TACO and TRALI are the leading causes of transfusion-related morbidity and mortality in North America ( Table 91.3 ) and are discussed later. Epidemiology and proposed pathophysiology for other common transfusion reactions—including respiratory dysfunction associated with transfusion, hemolytic and nonhemolytic transfusion reactions, febrile and allergic reactions, bacterial contamination, and GVHD—are detailed later.


Jun 26, 2021 | Posted by in CRITICAL CARE | Comments Off on Transfusion medicine

Full access? Get Clinical Tree

Get Clinical Tree app for offline access