Hello and welcome. My name is Chris Watson, Pediatric Critical Care Faculty at the Medical College of Georgia. I'm excited to have the chance to join you today to speak on transfusion thresholds in critically ill children. I have no actual or potential conflicts of interest in relationship to this presentation today. This talk has three objectives. To outline the epidemiological patterns of blood product transfusion in critically ill children, to review existing guidelines for transfusion thresholds in these children, and to highlight key areas where there are gaps in our understanding and opportunities for future investigation. Let's begin here today with the first of two case studies for us to consider. That of a two-year-old child with septic shock who has received a volume resuscitation of 80 milliliters per kilogram thus far. The child has a heart rate of 150 beats per minute and a blood pressure of 65 over 32. A titrated epinephrine drip is being administered and the complete blood count results revealing that the hemoglobin is 8.5 grams per deciliter. The question for you to consider is whether you would transfuse red blood cells in this patient. As we pause to think about how we might answer this question, we may recognize that there is likely no correct answer as we don't have the full data to support either yes or no in this particular situation. Now let's consider this as our second case today. That of a two-year-old child with pediatric acute respiratory distress syndrome, or PARDS, who is receiving conventional mechanical ventilation at the settings shown here. This child's heart rate is 120 beats per minute and blood pressure is 80 over 42. No vasoactive support is currently being provided. The complete blood count results revealing a hemoglobin of 8.5 grams per deciliter. Again, would you transfuse red blood cells in this child? As we pause to consider this case, we may ultimately conclude that in this particular instance the answer is probably no. Even in the presence of what appears to be hypoxia, giving red blood cells may increase the oxygen carrying capacity, but it isn't going to increase the ability of oxygen absorption where it's needed in the tissues. In fact, it may actually hinder the ability to get oxygen into the tissues due to adverse effects on the pulmonary endothelium and epithelium. Now if we look at who receives transfusions, we see that it's relatively common that hospitalized children receive a blood product during admission. In this recent cohort of more than 61,000 hospitalized children, we see that 9% received at least one blood product, most commonly RBCs, followed by platelets, plasma, and cryoprecipitate, respectively. In these patients, the median pre-transfusion laboratory values were a hemoglobin of 7.9 grams per deciliter, a platelet count of 27 times 10 to the ninth cells per liter, and an INR of 1.6. Infants and children were more likely to be transfused single products in the ICU and general ward and multiple products in the operating room. Finally, children undergoing cardiac repair with cardiopulmonary bypass were more likely to be transfused. So who specifically receives red blood cell transfusions in the PICU? Well, to help us answer this question, we can refer to this study from the Pediatric Acute Lung Injury and Sepsis Investigators Network, or POLICI network, that's cited quite frequently. It is an observational study where almost 50% of kids who'd been admitted to the PICU for 48 hours or more received red blood cell transfusion. The most common indication was for anemia, and the mean pre-transfusion hemoglobin was about 8.2 grams per deciliter. These data were collected before the transfusion requirements in the Pediatric Intensive Care Unit Study, or TRI-PICU study was published, but they are still some of the best epidemiological data we have for red blood cell transfusions in the pediatric ICU. This 2018 study is a secondary analysis of a point prevalence study looking at prophylaxis of deep venous thrombosis. In the study, about 14% of children enrolled received red blood cell transfusion within the previous 24 hours. This point prevalence study was not looking at the entire ICU stay, it was just looking at one day. Sepsis, cancer, and renal disease were the diagnoses most frequently associated with transfusion. In addition, about 60% of children on ECMO received transfusions. So this study shows that even after the TRI-PICU study, patients in the Pediatric Intensive Care Unit still received transfusions fairly commonly. Few recommendations currently exist to guide transfusion practices during and after cardiac surgery in children. This recent 2021 secondary analysis of the REDS3 data describes transfusion practices in a multi-center cohort of nearly 900 children with congenital heart disease who underwent repair prior to the TRI-PICU study. Most children in this study with congenital heart disease requiring cardiopulmonary bypass were transfused with multiple products, both intra and post-operatively. Additionally, the majority of intraoperative transfusions were at seemingly normal or low normal pre-transfusion values. Finally, infants more commonly received red blood cell and cryoprecipitate transfusions for both single ventricle and biventricular repair as compared with older age groups. So why do we transfuse red blood cells in the Pediatric Intensive Care Unit? Well, if a patient has acute bleeding, depending on the severity of bleeding, certainly transfusion is appropriate. But what about patients who are not bleeding? The main thing is to consider oxygen delivery. The arterial oxygen content equation here reminds us that increasing hemoglobin is one of the most efficient ways to increase blood oxygen content and deliver that oxygen to end organs and tissues, at least mathematically. So it seemed that more hemoglobin would be good, right? Well, it turns out that more is not always better. This is a multicenter, randomized controlled trial of 637 PICU patients looking at both liberal and restrictive transfusion strategies. These were hemodynamically stabilized, critically ill children to include those with shock who were on stable doses of vasoactive infusions. Patients were randomized to two groups, a liberal transfusion strategy with the hemoglobin threshold of 9.5 grams per deciliter, and a restrictive transfusion strategy with the hemoglobin threshold of 7 grams per deciliter. At the end of the study, there was no difference in outcomes between the two groups. Thus, based on the results of this study, it was concluded that restrictive transfusion is absolutely safe for general PICU patients. At this point, we may be left asking ourselves, why isn't maintaining a higher hemoglobin better? Well, one reason may be that hemoglobin alone is a poor marker of who will benefit from a red blood cell transfusion. Here are the results of a recent observational single center study of 94 children with severe sepsis or septic shock. The investigators looked at hemoglobin alone as a marker to guide red blood cell transfusion. This plot illustrates the interaction between transfusion and clinical outcomes shown as predicted organ dysfunction days. The dashed lines represent transfusions, while the solid line represents patients who did not receive transfusions. On the left, the shock index is used as a marker for transfusion, and the circle on the left highlights the patients with low severity of shock. In this group, red blood cell transfusion was associated with worse outcomes. The circle on the right indicates where the two lines intersect. At this point, transfusion showed some benefit and better outcomes. The chart on the right looks solely at hemoglobin as a marker, and its lines run parallel. This indicates that in this population, hemoglobin is probably a poor predictor of which patients might benefit from red blood cell transfusion and which might be harmed. Although this is a small set of patients in a single center study, it suggests that the marker we normally use to guide transfusion might not always be the best marker to use. So now let's return to our question. Why isn't maintaining a higher hemoglobin necessarily better? Well, we've already argued that hemoglobin alone is a poor marker of who will benefit from transfusion. Another reason why more hemoglobin might not always be better is that even though hemoglobin is, in some patients, a good way to deliver oxygen to the organs, red blood cell transfusion might not be the most efficacious oxygen delivery method in your patient. So if we ask, is red blood cell transfusion always efficacious? It turns out that the results of studies on the hemodynamic effects of red blood cell transfusion are actually mixed. Some of these differences may be due to blood product-related factors like storage. That includes storage duration, the type of container used, the processing that occurs before transfusion, donor-related effects, and to some extent, the amount of blood that is transfused. Some of the differences may be due to patient-related factors such as the type of shock and whether the patient is septic or not. The third reason why increased hemoglobin might not always be better is that the risks of transfusion might outweigh the potential benefits. Some patients benefit from transfusion while the risks outweigh benefits for others. This could be the key concept. So what do we mean by that? Well, we tend to think of therapy having the same effect in a specific population, but within an intensive care population, a lot of differences can occur among patients. Those differences can influence the effect of therapy so that some patients experience a benefit, some experience harm, and some experience no effect at all. While the effect of therapy might generally be the same for the overall patient population, it might not be the same for every individual within that population. Maintaining a higher hemoglobin level isn't generally helpful, but that might not be true for the specific patient you're treating. What's needed is a way to personalize transfusion prescriptions to improve outcomes. Let's talk now about the risks of red blood cell transfusion. So what are the risks? Well, relatively uncommon problems include transfusion-related acute lung injury, or TRALI, transfusion-associated circulatory overload, or TACO, and infections. But studies also show that transfusion is independently associated with adverse outcomes in critical illness. There are probably several potential mechanisms, including transfusion-related immune modulation, transfusion-associated pulmonary disease, transfusion-related coagulopathy, and potentially endotheliopathy as well. Recommendations have been made for thresholds for red blood cell transfusion. This 2018 decision tree comes from the Pediatric Critical Care Transfusion and Anemia Expertise Initiative, or TAXI. It looks intimidating, but we'll go through it section by section. As we begin to go through the TAXI RBC transfusion recommendations, first we'll consider critically ill children or children at risk of critical illness. For the child that is in hemorrhagic shock, transfusion is recommended using RBC plasma and platelets in either a 2 to 1 to 1 or 1 to 1 to 1 ratio until bleeding is no longer life-threatening. If the child is not in hemorrhagic shock and the hemoglobin measure is less than 5 grams per deciliter, transfusion is indicated. If the hemoglobin measurement is between 5 and 7 grams per deciliter, we should consider transfusion based on our best clinical judgment. With the hemoglobin of 7 or greater in a child who is not hemodynamically stable, then the recommendation from TAXI is to use our clinical judgment. No evidence was found for a specific hemoglobin-based transfusion threshold. This slide shows the recommendation for hemodynamically stable children with hemoglobin measurements of 7 grams per deciliter or greater. We see that across multiple populations, including generally critically ill patients, post-operative patients, patients who have pediatric acute respiratory distress syndrome or PARDS that is not severe, patients with non-life-threatening bleeding, and patients requiring renal replacement therapy, that the recommendation is to not transfuse routinely. For patients with acute brain injury and oncologic patients, the hemoglobin ranges vary and pediatric studies from which to glean data are few. Finally, for those with hemolytic anemia or severe PARDS and patients requiring ECMO or a ventricular assist device, the recommendation is to use clinical judgment. These last two groups were excluded from the Tri-PICU study and available data were insufficient to recommend specific thresholds in these patients. So, in considering patients specifically on ECMO, what should our RBC transfusion threshold be? Well, currently, there are insufficient prior studies for recommended transfusion thresholds in children receiving this form of oxygenation. The Extracorporeal Life Support Organization, or ELSO, recommends maintaining a normal or near-normal hematocrit level, which would be a hemoglobin above 12 to 14 grams per deciliter. That's likely more than what's needed, but the optimum level is unknown. This is a survey of red blood cell transfusion practices in ECMO at Polisi network sites. In it, the investigators showed that for children with no congenital heart disease or a cyanotic congenital heart disease, most clinicians use a hemoglobin threshold of around 10 grams per deciliter. For kids with cyanotic congenital heart disease, greater variation was seen, ranging from 10 to 13 grams per deciliter. The final section of the TAXI decision tree pertains to children who are hemodynamically stable with a hemoglobin of 7 grams per deciliter or greater who have heart disease. For those with uncorrected congenital heart disease, the recommendation is to transfuse to maintain a hemoglobin of 7 to 9 grams per deciliter, depending on the patient's cardiac reserve. In cases of biventricular repair and single ventricle stage 1, 2, or 3 palliation, the recommendation is to generally avoid transfusion with the caveats shown here in red. The data for these three recommendations are from single-center trials. Finally, for those patients with congenital or acquired myocardial dysfunction or pulmonary hypertension, the recommendation is to use clinical judgment. In these cases, there is no evidence showing that transfusion above 10 grams per deciliter is beneficial. What we know right now about red blood cell transfusion in critically ill children is that for most children, less is probably more. The right transfusion strategy is probably not the same for every patient. In fact, it's unlikely that hemoglobin levels alone are the best way to decide when to transfuse a patient. Here's another case for us to consider. This is a two-year-old child with abdominal sepsis. There are no signs or symptoms of active bleeding, but the surgeon wants to take the child to the operating room for an exploratory laparotomy. Laboratory results show the INR is 1.9 and the platelet count is 70,000. So with this presentation, would you transfuse this child with plasma? Would you transfuse this child with platelets? How would your decision change if the INR was 3? So who receives platelets in the PICU? This is from the P3T study, an international point prevalence study that included 82 PICUs. Overall, more than 3% of patients received platelets regardless of why they were in the ICU. The pie chart shows a breakdown of who got platelets. Only 12% received platelets for major bleeding and 21% for minor bleeding, while 35% of patients received prophylactic transfusions either because of a perceived risk of bleeding from a device before a procedure or just to meet a threshold. And some of these thresholds exceeded platelet counts of more than 20,000. Prior to 2022, there were no specific guidelines for platelet transfusion or plasma transfusion in critically ill pediatric patients. Recommendations were largely extrapolated from adult studies and for platelets from several neonatal studies. The American Association of Blood Banks guidelines suggest giving platelets in four situations. In cases of hypoperliferative thrombocytopenia with a platelet count below 10,000, bleeding accompanied by a qualitative defect, in instances of unexplained significant bleeding during cardiopulmonary bypass, and with a platelet count below 100,000 while receiving ECMO. In the latter situation, ELSO guidelines actually specify a platelet count below 80,000. The last two situations are expert opinions, not evidence-based recommendations because of a lack of evidence. This set of guidelines comes from the American Society of Clinical Oncology. Under these guidelines, platelet transfusion is recommended when the platelet count is under 10,000 for patients with malignancy who are undergoing treatment and for those who have undergone stem cell transplant. If the platelet count is below 20,000 before central venous line placement, transfusion is indicated, as well as before major invasive procedures if the platelet count is below 40,000 to 50,000. Lastly, if the count is under 50,000 before performing a lumbar puncture, platelets should be given. For patients who have a chronic stable thrombocytopenia and who are not undergoing treatment, the recommendation is to use clinical observation to decide when and if platelet transfusion is needed. That recommendation is based on adult studies that generally showed no benefit for platelet transfusion in these patients. But remember, there are differences between adults and children with thrombocytopenia. This P3T subgroup analysis helps us to understand a bit better which critically ill oncologic patients are receiving platelet transfusions in the PICU. Of the original study cohort of 548, 43% had an underlying oncologic diagnosis. Overall, oncology patients were older, had less organ dysfunction, received less mechanical ventilation, milrinone, and were less frequently placed on ECMO. As compared to the cohort without an underlying oncologic diagnosis, those with an oncologic diagnosis more commonly received at least one platelet transfusion prior to study enrollment. Most received the transfusion to maintain a defined threshold, with nearly 60% receiving a count greater than 20,000. Finally, oncologic diagnosis was associated with a poor platelet increment response. New in 2022 are recommendations for thresholds for platelet and plasma transfusion in critically ill children from the Transfusion and Anemia Expertise Initiative Control Avoidance of Bleeding, or TAXICAB. The clinical decision tree to help guide platelet transfusion is shown here. Similar to TAXI, these recommendations conclude that the decision to transfuse either platelets or plasma should not be driven by commonly used coagulation parameters alone, but should incorporate monitoring hemostasis into the overall clinical context and balance by considerations of risks, benefits, and alternatives. The clinical decision tree initially branches based on severity of bleeding into three categories with clear definitions, none or minimal, moderate or severe, and massive hemorrhage. Similar to TAXI, for massive hemorrhage, a balanced transfusion strategy is recommended with a ratio of either 1 to 1 to 1 or 2 to 1 to 1 of RBCs, plasma, and platelets until bleeding is no longer life-threatening. Though the full TAXICAB recommendations are worthy of much greater review, some key expert consensus statements include considering not transfusing platelets in patients on ECMO, following cardiopulmonary bypass, or with severe TBI for counts greater than 100,000 in the absence of clinically significant bleeding, considering platelet transfusion following non-cardiac surgery for counts less than 20,000 with minimal or no bleeding and less than 50,000 with moderate to severe bleeding, and consider transfusing for invasive procedures outside of the operating room for counts less than 20,000. In pediatric oncology patients, stem cell patients, and those with sepsis and or DIC, transfusion may be considered when the count is less than 10,000 with no bleeding and less than 50,000 in patients with sepsis or DIC and moderate to severe bleeding. Finally, for patients with moderate to severe bleeding, much discretion is given to clinical judgment that incorporates hemostatic assessment into the overall clinical context given insufficient evidence to make specific recommendations at this time. Certainly, further review of these guidelines is worthwhile. Now let's talk about who receives plasma in the pediatric ICU. This study by the plasma TB investigators is another international point prevalence study. This time, more than 100 PICUs and more than 400 children were enrolled. In this study, 3% of patients received plasma. About a fifth received plasma for critical bleeding and slightly more for minor bleeding. A third of these patients got plasma when they were not bleeding and there was no procedure planned. Presumably, in these patients, transfusions were given to correct abnormal prerequisites. Again, prior to 2022, as with platelet guidelines, there were no plasma guidelines specific to transfusion in critically ill pediatric patients. With the recent publication of the TAXICAB guidelines, we now have a more comprehensive evidence-based and expert-based set of guidelines for plasma transfusion practices in PICU patients, which we'll return to shortly. Prior to returning to a discussion of the TAXICAB recommendations as they pertain to plasma transfusions, it's first worth highlighting this set of adult guidelines from the British Society of Hematology. These guidelines do recommend plasma as part of resuscitation for hemorrhagic shock, but they do not recommend routine prophylactic transfusion to correct abnormal coagulation test results in patients who are not bleeding. That includes transfusion before invasive procedures. These guidelines are based on two things that we know. First, routine coagulation test results poorly predict bleeding risk, and second, there is no evidence that plasma transfusion in critically ill patients corrects an INR of less than 2.5. So now if we go back to the new 2022 TAXICAB guidelines, this is the clinical decision tree to help guide plasma transfusion in critically ill children. As highlighted earlier, these recommendations encourage that transfusion decisions be driven by incorporating measures of hemostasis into the overall clinical context, balanced by consideration of risk, benefit, and alternatives. Some of the key expert consensus statements regarding plasma transfusion include utilizing balanced transfusion ratios during massive hemorrhage resuscitation, considering transfusion in children following non-cardiac surgery with moderate bleeding with a PT, INR, or APTT greater than two times the reference values, and to consider prophylactic plasma transfusion in patients undergoing invasive procedures outside the OR without bleeding if the INR is greater than 2.5. Beyond these subpopulations and in line with the previously mentioned British adult guidelines, plasma transfusion is generally not recommended prophylactically in the absence of bleeding, though in the case of ECMO, following cardiopulmonary bypass, and in severe trauma, clinical discretion remains. Again, further individual review of these guidelines is worthwhile and suggested. So what do we know about transfusion in the pediatric intensive care unit? Well, we know that blood product transfusion is commonly used in the care of critically ill children. The TAXI and TAXICAB guidelines are extremely helpful, but we also know that many questions remain, and the optimum strategies are unclear for the transfusion of red blood cells, plasma, and platelets in critically ill children. And with that, I'll conclude, and I thank you again for your time and your attention.

critically ill children

transfusion thresholds

TAXI and TAXICAB guidelines

personalized transfusion

blood product transfusion

PICU transfusion risks