Blood transfusions in neonatal cardiac surgery and ECMO

Approximately 6–8 of 1000 newborns have a congenital heart malformation often requiring surgery. Extracorporeal membrane oxygenation (ECMO) provides life support to patients with respiratory and/or cardiac failure. After cardiac surgery ECMO is indicated in 0.5–3.5% of infants [1]. Currently, 1–2% of critically ill newborns are treated by ECMO. The most frequent indications for ECMO in newborns are meconium aspiration syndrome (MAS) and congenital diaphragmatic hernia (CDH), together accounting for more than 50% of ECMO procedures. Other indications are persistent pulmonary hypertension associated with cardiac or respiratory failure (PPHN), respiratory distress syndrome (RDS), patients failing to come off from cardiopulmonary bypass after cardiac surgery and sepsis. An indication for primary ECMO generally emerges on the second day after birth and the median duration of treatment is approximately 8 days. The survival rate is, depending on the indication, 60% to over 85% as compared with an estimated mortality rate of 40–70% should ECMO not have been applied. Long-term morbidity, after intra-cranial haemorrhage (ICH) or chronic lung disease is however frequent. Based on a randomized study, the costs associated with ECMO were estimated approximately 75 000 pounds (level 1998) per infant without disability at 1 year follow-up [2]. Although exact figures are lacking, in high income countries, infants below the age of 1 year use less than 2% of the blood supply [3,4]. In neonates, approximately 20% of red blood cell (RBC) and most of the plasma products are used for surgical procedures, of which cardiac surgery is the most common indication. The major reason for transfusion in cardiac surgery is because the extracorporeal circuit (ECC) needs to be primed with blood. Ten to 25% of neonates admitted to a neonatal intensive care unit (NICU) receive one or more platelet transfusions (PT). As compared with critically ill infants not receiving PT, the mortality rate was reported 10 times higher. In infants receiving 5 or more PT this risk increased almost 30 times [5]. Infants on ECMO receive almost daily RBC and PT transfusions. A retrospective survey in the period 2002–2007 among 12 329 neonates referred to several American NICU's, identified 45 infants who had received more than 20 units of PT. Almost half of these, 21 infants, because of ECMO [6], underscoring that infants on ECMO obviously belong to a multi-transfused group. Both cardiac surgery and ECMO are associated with exposure of the infants to multiple donors. The volume of the ECC in relation to the infant's circulating blood volume often requires priming of the system with blood to avoid unnecessary haemodilution. For cardiac surgery, devices requiring lower priming volumes are becoming available. Reducing the extracorporeal volume from over 500 ml to less than 300 ml. and even as low as 130 ml, enabled even in small children of 5–10 kg bloodless cardiac surgery in 55% of the cases [7,8]. Further downsizing is in progress. A decrease of the extracorporeal volume will also benefit the extent of inflammation and of the systemic inflammatory response syndrome (SIRS) following open heart surgery [8,9]. All ECCs, using a roller pump or a centrifugal pump, cause haemolysis [10]. Haemolysis leads to an increase of plasma-free haemoglobin (PfHb) and decrease of the haptoglobin scavenger. Free RBC constituents increase the systemic and pulmonary vascular resistance, induce platelet dysfunction and renal tubular damage. The use of special volume adapted pump systems and coating of tubing and extracorporeal surfaces show promising effects reducing haemolysis and thrombus formation [11]. However, increased degree of haemolysis and inflammation-induced capillary leakage can still be expected in young children, extensive surgery, ventricular assist devices (VAD) and ECMO because blood is exposed to a relatively large ECC surface [12]. For infants less than 3 months of age and for premature infants it will remain difficult to circumvent transfusions. Besides haemolysis, a fall of approximately 40–50% of the platelet count is an inevitable consequence of ECC. Both in cardiac surgery and in ECMO this happens immediately after connection to the circuit. After ECMO a nadir is reached after 3 days [13]. Besides, there is dysfunction of platelets, which become activated and at the same time show impaired aggregation to agonist. Platelet function tests such as TEG are disturbed after cardiac surgery and during ECMO [14]. In combination with heparinization to prevent clotting in the ECC circuit, this may enhance the risk for ICH, a severe but frequent complication during ECMO [15]. In addition to cellular activation and damage, ECMO also causes biochemical changes. A frequent problem at initiation of ECMO is hypercalcaemia (serum Ca > 11 mg/dl; > 2.74 mmol/l) affecting approximately one-third of neonates and associated with longer ECMO support and more platelet transfusions [16]. On average, ECMO is applied for 6–8 days, with a wide range of 2–10 days. The daily need for RBC is 0.5–1 unit and 0.5–2 units of PT, besides plasma and other fluids. (ELSO: ExtraCorporealLifeSupport Organization registry). The use of PT in relation to the vascular access technique was studied in 234 infants treated for 7.9 ± 6.3 days [17]. Veno-venous (VV)-ECMO was applied in 81 infants requiring daily 1.06 PT, whereas veno-arterial (VA, 138 infants) or the use of both techniques (15 infants) required an average number of 1.57 PT/daily. Recently, a case comparison study suggested that addition of continuous haemofiltration to ECMO may result in less days requiring ECMO (98 h compared with 126 h in historical controls) and fewer RBC transfusions (0.9 units/day as compared with 1.8 U/day), while PT (0.7 unit/day vs. 0.9 unit/day), fresh frozen plasma (FFP) usage and mortality (16%) was equal between the two groups [18]. Several questions related to transfusion supportive care are regularly put forward, but few are definitely answered. Cases of bloodless open-heart surgery in children > 3 months by priming of the circuit with crystalloids have been published in Jehova's Witness. In cyanotic congenital heart diseases, in which the haematocrit is high, autologous blood has been withdrawn (12 ml/kg) after induction of anaesthesia prior to the ECC to avoid haemolysis. This could avoid blood transfusion in 70–80% of cases of repair surgery [19]. A retrospective survey compared the use of fresh whole blood with packed red blood cells for priming. For infants < 5 kg body weight, lower transfusion needs the first 12 h after surgery (64% needed 1 unit) were observed after priming with fresh whole blood as compared with 85% of infants needing > 2 units, if the pump was primed with packed RBC [20]. Clinical outcome was similar in both groups. There are a few prospective studies comparing priming of the ECC with fresh whole blood or with reconstituted blood consisting of fresh or stored RBC and FFP. A prospective randomized study in 200 infants compared fresh (< 48 h old) whole blood with reconstituted packed RBC with FFP for pump priming. No difference in blood loss, in transfusion needs and clinical outcome was found. The hospital stay was even shorter when reconstituted blood (the storage interval of RBC in this study was 6 days) was used. In this study, 40% of the infants were < 28 days old [21]. In a prospective study in 30 infants, priming with reconstituted blood, containing fresh (stored < 5 days) or old (> 5 days) packed RBC in mannitol and adenine solution, was compared. Despite there was a storage-dependent increase in potassium (from 5.4 to 18.4 mEq/l) in the packed RBC, already 20 min after mixing and circulation with other priming constituents this difference had disappeared and during surgery and upon arrival at the NICU no differences in pH, lactate, potassium, and glucose were observed [22]. Contradictory to the above-mentioned studies came a randomized study from Canada [23]. In 64 infants, less than 1 month of age, the group transfused with reconstituted fresh whole blood throughout the whole procedure (priming, surgery and up to 24 h postoperative transfusions) had less chest tube blood loss (7.7 vs. 11.8 ml/kg) and improved clinical outcome with respect to ventilation (119 vs. 164 h) and hospital stay (12 vs. 18 days) as compared with the group transfused with stored reconstituted blood. In adults, a Cochrane analysis summarized over 200 RCTs that recruited 20 781 patients. The results showed a small advantage of aprotinin, a serine protease inhibitor, reducing operative blood loss, number of transfusions and need for re-exploration as compared with the lysine analogues, epsilon aminocaproic acid (EACA) or tranexamic acid (TXA). In this analysis, the incidence of myocardial infarction, stroke and renal failure were not higher in the aprotinin-treated patients [24]. However, in newborns, the coagulation system, in particular fibrinolysis, differs from older children and adults [25]. On the use of aprotinin, EACA and TXA in newborns and infants to reduce blood loss in cardiac surgery, few prospective studies comparing these drugs have been performed. In a RCT, comprising 320 patients, aprotinin was compared with EACA or a combination in congenital cyanotic heart diseases. The results showed no difference between the two drugs with respect to bleeding, transfusion needs, incidence of re-exploration and clinical outcome [26]. The same group also compared EACA with TXA or placebo in 150 cyanotic infants and found both drugs evenly superior to placebo [27]. The three drugs were compared in 100 infants with cyanotic heart disease equally divided in four groups of 25 patients. No advantage for a combination of aprotinin (2 × 30 000 KIU/kg) with TXA (three times 100 mg/kg) was observed. TXA alone or aprotinin alone were equally effective as the combination of both drugs [28]. In a review on this subject, no difference in amount of bleeding was reported between aprotinin, EACA and TXA. Almost all studies reported less bleeding with the use of antifibrinolytics as compared with placebo. However, the studies did not allow a conclusion on safety [29]. Increased renal failure using aprotinin could not be confirmed in 200 neonates and, as shown in other studies, the cardiopulmonary bypass time (if longer than 100 min) has the highest association with renal failure [30]. Studies examining rVIIa in neonates are limited and randomized studies are not available. In most published cases, rVIIa was used for uncontrolled bleeding as compassionate need. Combined case series together comprising 44 infants below 1 year of age, reported cessation of bleeding in all cases except one case with surgical bleeding [31–34]. In an open label study in six infants with postoperative bleeding, re-operation because of bleeding was prevented [35]. Some series observed no adverse effects, however, altogether at least five thrombotic events occurred in these 44 cases, in particular, in combination with ECMO. Shortly after bypass termination it is important to limit bleeding and optimize systemic oxygen delivery. Heparin, hypothermia and inadequate protamine correction are important factors that contribute to bleeding and should be adequately treated [36]. Guidelines recommend to aim at a platelet count > 100 × 10 E9/l. Platelets of young infants (< 2 months of age) with congenital heart disease are less activated by cardiopulmonary bypass compared with platelets from children older than 12 months. The clinical significance, for instance regarding the trigger for platelet transfusion for the very small infants is unknown [37]. Because platelet number and function are impaired after cardiac surgery, there is concern about volume replacement with solutes affecting platelet function. In adults after open heart surgery, the use of hydroxyl ethyl starch (HES 15 ml/kg) as compared with albumin did not result in obvious more chest tube drainage blood loss, despite HES prolongs clot formation in trombelastography (TEG) [38]. In a randomized clinical trial, 42 children undergoing cardiac surgery were assigned to receive HES (130/0.4) or FFP, both in a volume of 10 ml/kg after termination of cardiopulmonary bypass (CPB). Endpoints were APTT, INR, blood loss and blood product usage until the first postoperative day. Although the INR was significantly longer in the HES group, postoperative blood loss and transfusion of blood products were not different between the two groups [39]. Regarding red cell transfusions it is a general assumption that after cardiac surgery there is a lower margin of safety for low Hb levels, because stroke volume, heartbeat and coronary blood flow are major compensators to maintain tissue oxygenation. In neonates, oxygen delivery is at near maximal levels and in case of a high oxygen demand there is a risk of tissue hypoxia. However, although the relationship may not be causal, despite improvement of oxygen transport by RBC transfusions, these are also associated with increased postoperative complications, in particular infections [40]. More than 90% of neonatal RBCs contain HbF, having impaired oxygen delivery by a left shift of the oxygen dissociation curve [41]. After cardiac surgery, often more than 50% of the RBCs carry HbA. The consequences of this shift from HbF to HbA have however not been considered in setting transfusion targets in newborns. Given the limited number of studies, firm conclusions for priming and transfusion in neonatal cardiac surgery can not be made, but: There is no high level evidence to prime the CPB circuit with fresh whole blood. There is still controversy on the limitation of storage time of erythrocytes for reconstituted RBC and FFP for priming the CPB circuit. Antifibrinolyic treatment reduces blood loss, number of transfusions and need for re-exploration because of bleeding. Lysine analogues (TXA, EACA) seem as effective as aprotinin to reduce bleeding. CPB bypass time (>100 min) is the major factor contributing to blood loss, transfusion needs and postoperative renal failure. rVIIa for compassionate need to be balanced in individual cases between benefit (prevention of re-exploration) and possible adverse thrombotic effects. Rapid degradable hydroxyethylstarch impairs platelet function tests but is not associated with more blood loss or transfusions after cardiac surgery in infants. Postoperative haemoglobin level may not be an appropriate transfusion trigger if HbF/HbA ratio's are not taken into account. Bleeding and thrombosis are major problems in ECMO treatment and functional coagulation tests, such as thrombelastography (TEG), shows the whole range between severe coagulopathies (DIC) to hypercoagulability [42]. Intracranial haemorrhage is a major cause of neurodevelopmental dysfunction and mortality in children using ECMO and is reported to occur in 10–52% of the patients [43]. Risk factors for ICH are low pH, bradycardia < 80 min, hypotension (MBP < 30 mmHg), difficulties to maintain the activated clotting time (ACT) between 190–210 s, lower platelet counts and the need of more transfusions [15]. ICH may also be associated with larger intravascular volume administration in the first 24 h after starting ECMO [44]. In comparison to a historical control group, 42 newborns receiving EACA showed significant less bleeding and none developed ICH [45]. However, in a double-blinded randomized study between EACA and placebo in 29 neonates, five cases of ICH occurred (17.2%), without a significant difference (EACA 23% and placebo 12.5%) between groups [46]. As compared with slightly older infants subjected to open heart surgery, the newborn has an increased risk for bleeding and thrombosis the first week(s) of life [47]. Despite AECA, TXA and activated factor VIIa were reported to have life-saving effects in individual cases and effective to reduce bleeding and transfusion needs, a lower incidence of ICH during ECMO has not been demonstrated and increased thrombosis at other sites has been reported after the use of rVIIa. In a retrospective unmatched case-control study of 12 patients treated with rVIIa for severe bleeding after cardiac surgery and placed on ECMO, 25% of patients developed thrombosis of which two with major thrombotic complications in the ECMO circuit [31]. Four patients surviving after rVIIa, showed occlusion of the oxygenator in two cases, but this was not at variance with controls [48]. Another small report on four patients with ongoing bleeding despite aprotinin, rVIIa caused cessation of bleeding without excessive thrombosis [49]. Platelet transfusions still are the major treatment option to treat and prevent bleeding during ECMO. Trigger levels used for platelet transfusions in ECMO are 80–110 × 10 E9/l [17,44]. Infants on ECMO receive PT in 30% of the cases for bleeding symptoms and in 70% for prophylaxis [6]. The mechanism of platelet dysfunction induced by ECMO is incompletely elucidated. Besides thrombocytopenia apparently due to consumption, platelets show a decreased in vitro aggregability with collagen, and in vivo increased levels of soluble P-selectin and metalloprotease MMP-2. MMP-2 is released from activated platelets and mediates a new, nonthromboxane, non-ADP-dependent pathway of platelet aggregation. It has been shown that exposure to a large artificial surface and oxidative stress during ECMO induce release of MMP-2. This MMP-2 is purely derived from platelets through contact with the artificial surface without signs of endothelial activation as concluded from absence of soluble E-selectin or vascular generated nitric oxide. [50]. Because transfused platelets during ECMO acquire the same bypass-induced platelet dysfunctions as the patient's circulating platelets, the value of prophylactic transfusions to prevent bleeding has been questioned [6,51]. Comparing 1600 thrombocytopenic children (not only infants on ECMO) in the NICU who –at the same level of thrombocytopenia-received or not received platelet transfusions the authors concluded that platelet transfusions themselves were likely responsible for some fraction of increased mortality in PT receivers [6,51]. The mortality of patients receiving >20 PT is over 50% and most patients do not die from bleeding [6]. This poses a crucial dilemma for future studies. Concluding (intracranial) haemorrhage of the infant and thrombosis of the ECMO circuit pose major challenges for the balance between antifibrinolytic drugs, use of rVIIa and transfusion support. However, clinical research is limited and evidence-based conclusions are not possible: Antifibrinolytics during ECMO may not reduce the incidence of ICH. rVIIa may lead to more thrombotic complications in the ECMO circuit. The appropriateness of prophylactic platelet transfusions is questioned, because transfused platelets acquire platelet dysfunction. When blood products are transfused as relatively large volumes and at a fast speed, special complications should be anticipated: High dose of citrate anticoagulant & preservative solution Hypothermia due to refrigerated blood Hyperkalaemia Depletion of 2,3 DPG in stored blood Many clinicians routinely infuse calcium gluconate in case of massive transfusions. Hypothermia enhances cardiac dysrhythmia/asystole. Transfusion of cold blood has been associated with apnea, hypotension, and hypoglycaemia. Moreover, the platelet function and coagulation processes may be impaired at temperatures below 34 °C. Hyperkalaemia is related to the shelf-life of RBC and with manipulations (centrifugation, irradiation) of RBC. During storage of RBC, potassium leaks from the cell and at the end of shelf-life (35–42 days) the potassium concentration can be as high as 50–70 mmol/l. Arythmia's and cardiac arrest have been reported after bolus transfusion of stored blood in at least six case reports (reviewed by [52]). For this reason, restriction of the storage time of RBC for 5–12 days is often recommended for (partial) exchange transfusions for infants and neonates [53]. However, such serious events are presumably very rare, given the high potassium load often present in stored RBC. Parshuram & Joffe measured a potassium concentration above 25 mmol/l in one-third of the RBC products stored 3–41 days (median 10 days). When measured after an average interval of 75 min after transfusion in only in three children the K+-concentration reached > 5 < 6 mmol/l without clinical symptoms [52]. Six of 54 transfusions in this study had been given as bolus of > 5 ml/kg/10 min without events. A case of cardiac arrest in a 2-month-old infant undergoing a Blalock-Tausing surgery is worth mentioning. She received 120 ml over 10 min through a central line in the inferior vena cava. The RBC had been stored for 6 days, but had been 48 h prior to The potassium concentration in the unit was mmol/l and in the withdrawn after with 6.3 mmol/l enhances potassium starting > 3 h after this and a after 3 days. It is to limit the storage time after as as An is to the and it prior to transfusion with FFP or However, although the potassium level by of RBC, this to cause more haemolysis in the ECMO circuit A transfusion of 5 the for transfusions, causes no in levels, the shelf-life of blood Several studies showed that infants below the age of months red cells and of infants after multiple transfusions could be 5 months after birth and none showed RBC found no 3 months after transfusion of an average units of RBC to infants have been An undergoing cardiac surgery at of age after 2 PT from case reports in newborns reported on at the age of 18 days and 11 and one at the age of 12 formation in transfused infants and the comprising > 150 patients, no formation Because of the risk of disease of the newborn of it is if PT are given to infants to A dose of is to five PT over a period of 6 For FFP and it is not to the factor because of of of the small of the formation of newborns is considered as result from massive transfusions, can cause adverse effects, such as haemolysis and lung A review cases of severe haemolysis due to in PT reported The age of the from 8 months to 18 and three children from the reported over a period of 10 six cases of severe in blood group children receiving from haemolysis and one developed a tests with In the published an to the for neonates and older children to avoid platelets for possible and of group platelets, to or should for high of is in neonates and small infants. A review found approximately 10 published cases of of which only one in an infant of months of age after cardiac surgery The authors the use of blood because of the that the that may cause In the analysis of 20 cases have been but the was 2 prevent other adverse transfusion effects for which neonates are such as and it is recommended to use RBC and PT and to cellular blood products with 25 for newborn with a age < [53]. transfusion studies in children are available. The system in the severe of transfusion reported for a period of 10 these occurred in children < 18 and in infants < 12 months of blood product transfused a severe transfusion occurred in in adults, in children < 18 and a three times higher incidence in infants < 12 months. to FFP and PT are approximately times more frequent as after RBC products In a prospective survey in Canada in a intensive care of the blood products caused an adverse these, blood products had been given to cardiac surgery patients in products a transfusion some on low levels of can be platelet transfusions should not be given to if not platelet with low should be platelet transfusions if to infants should be with In case RBC transfusions are at a > 5 an infant be for may be in young infants. of transfusion effects to a may to transfusion for newborns and small infants. In and prospective studies may small improvement of transfusions) The of on postoperative complications is unknown and a for future studies. on reducing the extracorporeal volume and activation of

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.