Age is inversely related with the risk of bleeding. Neonates and particularly premature babies are the highest-risk group for postoperative bleeding. Neonates do not a have a completely developed coagulation system at birth, but there is a balance between the low levels of endogenous procoagulants and anticoagulant systems (Arnold 2014). In addition neonates do not regulate calcium hemostasis well, which is essential in the coagulation process (Jain et al. 2010). Traditional coagulation testing such as prothrombin time (PT), thrombin time (TT), and activated partial thromboplastin time (APTT) are prolonged in neonates (Long et al. 2011). Refer to Chap. 10 for extensive discussion regarding preoperative testing.

Comorbidities like congenital coagulopathies (e.g., von Willebrand disease) or acquired coagulopathies secondary to diabetes and liver or kidney dysfunction also increase the bleeding risk.

Single-ventricle patient in the pre-Fontan stage has coagulation anomalies characterized by lower levels of protein C, protein S, antithrombin III, and factors II, V, VII, and X and longer prothrombin times probably due to chronic passive congestion of the liver. These factor anomalies tend to correct post-Fontan surgery probably due to improvement in systemic oxygenation and overall perfusion (Cheung et al. 2005). Independent from the primary lesion, cyanotic heart disease is associated with secondary erythrocytosis as a compensatory mechanism for hypoxemia. This compensatory increase in the red cell mass reduces the plasma volume with the consecutive reduction in coagulating factors, fibrinogen, and platelet count increasing the postoperative bleeding risk. The effect on the coagulation is directly related to the amount of hypoxia and polycythemia (Zabala and Guzzetta 2015). Cyanotic patients with multiple aortopulmonary collateral vessels have an increased venous return to the heart affecting surgical visualization and increasing postoperative bleeding due to poor surgical hemostasis (Donmez and Yurdakok 2014).

Antiplatelet agent and anticoagulants are commonly used in pediatric patients with congenital heart disease and have been implicated in increasing surgical bleeding. Of the antiplatelets, aspirin has the lower risk of bleeding, ADP-receptor antagonist has an intermediate risk, and the GPIIb/GPIIIa receptor antagonists have the highest risk. Aspirin causes an irreversible inhibition of the platelet cyclooxygenase 1 (COX1) which is responsible of formation thromboxane A₂ essential for platelet activation and aggregation. Aspirin duration of action is related to the platelet turnover (about 10 days) because the inhibition is not reversible. About 10 % of the platelet COX1 activity recovers per day due to platelet turnover, and only 20 % of the platelet COX1 activity is needed to achieve a normal hemostasis (Awtry and Loscalzo 2000). ADP-receptor antagonist such as clopidogrel affects the geometry of the platelets making them spherical and unable to aggregate (Wijeyeratne and Heptinstall 2011). GPIIb/GPIIIa receptor antagonists are used infrequently in pediatrics, but they will also increase the risk of bleeding due to the profound capacity to prevent platelet aggregation, thrombus formation, and distal thromboembolism. Alternatively early withdrawal of aspirin and/or clopidogrel is not feasible in shunt-dependent patients at life-threatening risk for thrombosis. Similarly in emergency cases like heart transplant with harvesting or left ventricular assist device (VAD), patient undergoes surgery under full anti-aggregation and anticoagulation. Preoperatively the platelet count in these patients is normal since the production is not affected. Platelet functional studies such as PFA-100 and multiplate platelet aggregometer, which are described at length in Chap. 10, can detect platelet inhibition but are not used routinely in the preoperative period. Platelet aggregation studies showed conflicting data in terms of predicting not postoperative blood loss and more research is needed (Hofer et al. 2011; Orlov et al. 2014). Recently in adults undergoing heart surgery, low platelet activity predicted 30-day mortality bringing up the question when to discontinue the antiplatelet agents (Kuliczkowski et al. 2016).

Anticoagulants are another group of medications strongly associated with postoperative bleeding. Unfractioned heparin (UH), which is a mixture of polymers of sulfated glycosaminoglycans (molecular weight 5–30 kd), potentiates the anticoagulant effects of antithrombin III. Heparin has a short half-life since is it cleared from the circulation by endothelial cells (saturable mechanism) and by kidney excretion. Due to the extensive metabolism, UH is administered as a bolus and followed by an infusion. Target APTT is 1.5–2.5 normal values. Heparin is usually stopped 6 h before surgery, and residual effect can be checked by the activated clotting time (ACT) in the operating room. Normal ACT values are 80–160 s. Heparin concentration measurement in pediatrics has not correlated well with antifactor Xa activity and is not commonly used to detect residual UH effect (Gruenwald et al. 2000).

Low molecular weight heparin (LMWH) is a depolymerized molecule with an average weight of 5 kd which its effect is mediated by the inhibition of factor Xa. Due to lack of monitoring needed, longer half-life and predictable effect LMWH have gained popularity in the pediatric population. LMWH should be held 12 h before surgery, and if needed its residual effect can be checked by the antifactor Xa levels (0.5–1.0 U/mL therapeutic, 0.1–0.3 U/mL prophylactic).

Direct factor X inhibitors have limited indications in pediatrics, and its use has been restricted to heparin-induced thrombocytopenia. Due to its long half-life and lack of an antidote, these drugs are not ideal to be used in the preoperative period (Young 2008).

Vitamin-K antagonists like Coumadin inhibits the production of vitamin-K-dependent coagulation factors (II, VII, IX, and X). In addition Coumadin also inhibits the production of physiologic anticoagulant proteins C and S. Its effect is monitored by the international normalized ratio (INR). Therapeutic INR values depend on the indication (INR 2–3 thromboembolism, INR 2.5–3.5 mechanical valve). Vitamin-K antagonist’s long half-life makes it impractical for the perioperative period and is usually held for 3–5 days and transitioned to UH. If emergency cardiac surgery is needed in a patient on Coumadin, its effect can be reversed with prothrombin complex concentrate (PCC) which is going to be later in the chapter.

The intraoperative risk factors for bleeding are related to the procedure and to the cardiopulmonary bypass. Complex procedures by RACHS-1 score such as neonatal repairs (e.g., Norwood, arterial switch, truncus repair, and total anomalous pulmonary vein repair), single-ventricle palliation (Glenn Shunt and Fontan), redo surgeries, and aortic surgeries are the highest risk for perioperative bleeding (Guay and Rivard 1996; Guzzetta et al. 2015). Duration of surgery has been related with postoperative bleeding measured by chest tube output and ROTEM trace abnormalities (Hayashi et al. 2011).

Ultimately high risks for postoperative bleeding are patients on mechanical circulation with either extracorporeal membrane oxygenator (ECMO) or ventricular assist device (VAD). ECMO patients are kept fully anticoagulated on UH infusion to avoid thrombosis triggered by the contact of patients’ blood with ECMO circuit. In addition to coagulation activation, there is a dilutional effect on coagulation factors due to the ECMO prime. The target values for anticoagulation while on ECMO are ACT of 180–220 s, antifactor Xa levels of 0.3–0.7 IU/mL, and/or APTTs of 1.5–2.5 times the normal. Currently most of the ECMO circuits are heparin coated decreasing the amount of UH needed. Low levels of anticoagulation could lead into ECMO circuit thrombosis, but on the other hand excessive anticoagulation could lead to bleeding. Neurological injury due central nervous system (CNS) bleeding is the most feared complication of ECMO and frequent cause of withdrawing support.

VAD patients do not need full intravenous anticoagulation, but they are still at risk from thrombosis and emboli from the device, cannulas, and valves (e.g., Berlin Heart). Most institutions follow the Edmonton triple antithrombotic protocol entailing aspirin, dipyridamole, and either warfarin ≥12 months or enoxaparin <12 months. During the ECMO wean or VAD harvesting coagulation point of care (POC), testing is crucial to minimize the bleeding risk and conduct a coagulation goal-directed therapy (Annich and Adachi 2013; Esper et al. 2014; Seibel et al. 2008).

Cardiopulmonary bypass (CPB) causes massive physiologic changes in children characterized by hemodilution, coagulation activation, and hyperfibrinolysis (Sniecinski and Chandler 2011). The hemodilution effect is inversely proportional to size, neonates and infants being the ones affected the most (Table 36.2). The artificial surface of the CPB circuit activates platelets and the kallikrein-kinin system promoting thrombosis (Fig. 36.1). Heparin use even though effective to avoid circuit thrombosis and thrombin formation does not inhibit completely platelet and coagulation activation. During the CPB run, there is also platelet sequestration, downregulation of GPIIb/GPIIIa receptor, and destruction due to thrombogenic bypass circuit surfaces. Due to these facts platelet function and fibrinogen concentration are affected the most post CPB in pediatric cardiac surgery (up to 50 % of baseline values). An early study by Miller et al. showed that platelets and cryoprecipitate (rich in fibrinogen) restore hemostasis in the initial post CPB period (Miller et al. 1997, 2000). Infant CPB is conducted under some degree of hypothermia for most procedures and in some cases (e.g., aortic reconstruction – Norwood) under deep hypothermia circulatory arrest (DHCA). Mossad et al. showed that when comparing with adult cardiac surgery, pediatric patients have a higher incidence of DHCA use and blood transfusion requirements in the perioperative period (Mossad et al. 2007). Coagulation and inflammation activation is caused by the stress caused by the trauma of the circulating blood components with the artificial surface of the CPB circuit. Due to immaturity of coagulation and immune system, this activation is more profound in neonates and infants. Modulation of the stress response with intravenous steroid is common practice even though there is doubtful evidence for its use. Steroid will decrease the inflammatory mediators interleukin-1, interleukin-6, interleukin-8, tumor necrosis factor, leukotrienes, and endotoxin, but its effects on coagulation and postoperative bleeding are not well defined (Augoustides 2012). The priming solution varies with patient size and weight. Usually prime solution for patients <18 kg includes packed red blood cells (PRBC), crystalloids (e.g., PlasmaLyte), colloid (e.g., albumin), and/or fresh frozen plasma (FFP), trying to keep the solution as physiological as possible. In addition to the prime solution heparin, buffer solution (e.g., sodium bicarbonate), mannitol, and steroids are added. The use of FFP for CPB prime is debatable. Traditionally FFP is added to blood prime in patients <18 kg, but there is no evidence that this practice will improve outcomes and decrease postoperative bleeding. Desborough et al. in a Cochrane database of systematic reviews showed that in patients without coagulopathy, the addition of FFP did not improve the outcome (Desborough et al. 2015). Miao et al. showed that adding FFP to the CPB in a population of cyanotic patient (6 months–3 years) undergoing cardiac surgery did not decrease postoperative bleeding. Preoperative fibrinogen was an independent predictor of postoperative blood loss (Miao et al. 2014).

Cardiac surgery exposes the subendothelium, which is rich in thromboplastin, triggering platelet activation and aggregation binding to the von Willebrand factor and collagen forming the initial vascular plug. The coagulation system through the factors IX (FIXa) and factor X (FXa) is activated by the binding of the wound tissue factor (TF) to active factor VII (FVIIa) transforming prothrombin to thrombin. The activated platelets, factors V, VIII, and XI, work as catalyst accelerating the coagulation process. Next the clot stabilizes with fibrinogen and factor XIII. Finally, once the clot is formed and stable, the fibrinolytic system avoids further thrombus formation (Fig. 36.2).

Uncontrolled surgical bleeding can lead to coagulation activation by exposure of the subendothelium, coagulation factor loss, and thrombocytopenia. Blood product replacement should be balanced (e.g., packed red blood cells, coagulation factors, fibrinogen, and platelets) and POC targeted. Isolated PRBC replacement further dilutes coagulation factors, fibrinogen, and platelets perpetuating the vicious circle of coagulopathy and further bleeding. Hypothermia, acidosis, low-ionized calcium, and hyperfibrinolysis should also be tackled since they are major contributors of maintaining the postoperative bleeding cycle. CPB rewarming strategies are crucial since infants are prone to hypothermia due to widespread use of hypothermia and DHCA in pediatric cardiac surgery. Neonates and infants due to limited fat stores, inability to shiver, and larger ratio of body surface area to weight are especially susceptible to hypothermia. Hypothermia affects coagulation factors and platelet function perpetrating postoperative bleeding. Furthermore uncorrected hypothermia will increase oxygen consumption causing metabolic acidosis impairing hemostasis even further. Hypocalcemia is common in infants due to sarcoplasmic reticulum underdevelopment and reduced calcium storages. Furthermore massive use of citrated blood product will decrease ionized calcium even further (Kozek-Langenecker 2014). Residual heparin effect due to reheparinization and/or excess protamine has also been associated to postoperative bleeding. Lastly hyperfibrinolysis in pediatric cardiac surgery is a potential cause of postoperative bleeding. Miller et al. were the first group to described hyperfibrinolysis using thromboelastogram (TEG) in the post-protamine period. In his series hyperfibrinolysis was uncommon (2 out 32 patients, 6.25 %) and only present in the bigger patients (>8 kg cohort) (Miller et al. 1997, 2000).

A detailed description of the value of POC in postoperative bleeding is available in Chap. 10. The value of ROTEM, TEG, and traditional preoperative testing in predicting bleeding after pediatric cardiac surgery is under investigation. POC algorithms have been used for the stepwise approach of postoperative bleeding which is also presented in Chap. 10. The aim of POC algorithms is a targeted treatment of postoperative bleeding minimizing blood transfusion while improving surgical outcomes. Romlin et al. demonstrated that ROTEM could be used early during the rewarming period of CPB before hemoconcentration accelerating the analysis by running the HEPTEM/FIBTEM receiving information of clot firmness after just 10 min (Romlin et al. 2013). The same Swedish group studied a pediatric cardiac surgery population using TEG as POC testing showing decreased transfusion rate of PRBC and FFP while receiving more platelets and fibrinogen (Romlin et al. 2011). Other studies like the one by Lee et al. could not show that ROTEM predicted chest tube output after cardiac surgery (Lee et al. 2012).

Currently at Texas Children’s in high-risk patients, we use ROTEM as POC running HEPTEM and FIBTEM upon rewarming of CPB following Romlin et al.’s approach. The more prevailing finding is the decrease of maximum clot firmness (MCF) in both tests (HEPTEM MCF <50 mm, FIBTEM <9 mm). We utilize a modified version of the Clinic Cologne–Merheim algorithm (Vorweg et al. 2001) (Fig. 36.3) (See section “Case Vignette”). Espinosa et al. showed that ROTEM and TEG parameters correlated well with post CPB hemostasis changes and plasma fibrinogen and helped to guide fibrinogen replacement (Espinosa et al. 2014). Nakayama in pediatric cardiac surgical population validated the use of thromboelastometry-based algorithm reducing postoperative bleeding and decreasing the intensive care unit stay versus conventional treatment (ACT and platelet count transfusion guided) (Nakayama et al. 2015).