The Circuit and Cannulas

Unfortunately the circuit size cannot be reduced proportionately to the patient’s size; this disproportion commonly leads to hemodilution and dilutional coagulopathies in children. The surgical procedures require extremes of temperature, hemodilution, and changes in flow rates. Because of the smaller size cannulas and higher flow rates (150 to 200 mL/kg/min) in infants and children, cannulas assume a much more important management issue than in adults. Shear stress is significant in small cannulas and is several-fold greater than needed for activation of blood cells and platelets, leading to a disproportionately exaggerated systemic inflammatory response syndrome (SIRS).

Cardiopulmonary Bypass Pumps

The two pumps used most commonly for CPB are roller pumps and centrifugal pumps. Roller pumps have the advantages of simplicity, low cost, ease and reliability of flow calculation, and the ability to pump against high resistance without reducing flow.⁸ Disadvantages include the need to assess occlusiveness, spallation of the inner tubing surface (potentially producing particulate arterial emboli), capability for pumping large volumes of air, and ability to create large positive and negative pressures. Compared with roller pumps, centrifugal pumps offer the advantages of less air pumping capabilities, less ability to create large positive and negative pressures, less blood trauma, and virtually no spallation. Disadvantages of centrifugal pumps include higher cost, the lack of occlusiveness (creating the possibility of accidental patient exsanguination), and afterload-dependent flow that requires constant flow measurement. In the setting of short-term CPB for cardiac surgery, it remains uncertain whether the selection of a roller pump over a centrifugal pump, or of any specific centrifugal pump over another, has clinical importance. Pulsatile perfusion may prove to be beneficial in the future, but further outcome data and technical improvements are needed.⁹

Cardiopulmonary Bypass Prime

The optimal priming fluid in cardiac surgery is a topic of enduring debate. Crystalloid solutions, colloids, and mixtures of both are used. Children appear to benefit from a colloid prime. If crystalloid is used for priming, it should not contain lactate or dextrose because CPB induces a metabolic acidosis¹⁰ that has been shown to be iatrogenic and not splanchnic in origin.¹¹ The addition of lactate to the prime increases postoperative serum lactate concentrations and should be avoided.¹² Hyperchloremic metabolic acidosis is the second contributing component of a metabolic acidosis on CPB. This is often only detected by measuring the strong ion difference via the Stewart approach to the acid-base homeostasis.¹³ Both acidifying events are attenuated by the dilutional hypoalbuminemia induced by the administration of the pump prime. Because a hyperchloremic acidosis of a mild degree seems to be well tolerated and not associated with a poor outcome, no intervention seems necessary. Understanding the nature of CPB-associated acidosis, however, is likely to prevent unnecessary investigations or interventions.

The avoidance of dextrose is especially important during complex repairs using deep hypothermic cardiac arrest in which the risk of neurologic injury is substantive. The additives in banked blood, namely, glucose in citrate-phosphate-dextrose (CPD) storage solutions, also need to be considered as a source of glucose (together with the increased plasma concentrations of potassium in stored blood). We use a balanced electrolyte solution, such as Plasmalyte, for the crystalloid component of our prime.

The proportionally large volume of the bypass circuit compared with the child’s blood volume has a significant impact on the coagulation factors and cellular components. Platelet count decreases and coagulation factors, including fibrinogen, are diluted after bypass; these factors may contribute to a coagulopathy. The fibrinogen concentration at the end of bypass has been shown to correlate with the 24-hour chest drainage in children weighing less than 8 kg.¹⁴ This is seen more frequently in infants and neonates in whom an average decrease in plasma concentrations of hemostatic proteins by 56% immediately on initiation of bypass can be observed.¹⁵ Overall, younger age represents the single most important risk factor for coagulopathy and bleeding complications.¹⁶

One approach to the just-mentioned problems is the addition of whole blood to the circuit prime. Proponents cite two advantages: (1) improved hemostasis and (2) a decreased SIRS with less edema formation and less organ dysfunction. One study has disproved these perceived advantages; the investigators found that the use of fresh whole blood increased perioperative fluid requirements, leading to a longer duration of mechanical ventilation and ICU stay than in the single component group.¹⁷ The only advantage found by their study was the lesser number of donor exposures, a problem we try to overcome by matching packed red blood cells (PRBCs) and fresh frozen plasma (FFP) from the same donor.¹⁸ In addition, whole blood is frequently not available. An alternative approach is the use of FFP in the prime.¹⁹ Other investigators found that the use of FFP led to greater fibrinogen concentrations at the end of surgery. On average, children in the FFP group needed 1.3 fewer donor exposures and tended to need fewer PRBCs. The lower donor exposure was primarily because of fewer transfusions of cryoprecipitate.¹⁹ FFP may be safely substituted by 5% albumin in the prime in children with less complex repairs and acyanotic lesions.²⁰ Whenever possible, we prefer fresh blood less than 5 days old. Fresh PRBCs are presumably more balanced metabolically than stored PRBCs; the former contain less potassium, a greater concentrations of glucose, reduced concentrations of lactate, and a greater pH.²¹ Also postoperative morbidity increases with increasing age of red blood cells.²² Pulmonary complications, acute renal failure, and increased infection rates were among the main complications associated with increased red blood cell storage time. As far as potassium levels and acid-base balance are concerned, PRBC priming can be safely performed with stored PRBCs if the priming solution is circulated for 20 minutes before the initiation of CPB.²³

Depending on the size and age of the child, and the complexity of the repair, a target hematocrit is chosen. Based on the child’s blood volume and the prime volume, homologous blood is added using the following calculation:

The average prime volume of the circuits we use is shown in Table 17-2. Other prime additives are heparin, antifibrinolytics, antiinflammatory agents (corticosteroids), antibiotics, vasodilators, and, sometimes, diuretics (mannitol, furosemide). At the end of the case and before separation from bypass, blood gas analysis is done to ensure that the electrolytes (including calcium and magnesium ions), glucose, and hematocrit are within a desired range. Acid-base changes and sodium concentration are corrected with sodium bicarbonate, and residual lactate is washed out with the help of the hemofiltration.

Antifibrinolytic Agents (See Chapter 18)

Heparin-Induced Thrombocytopenia

The use of unfractionated heparin for anticoagulation for CPB in adults produces antiheparin antibodies in 25% to 50% of patients within 10 days postoperatively. In a small minority of these patients, high-titer IgG platelet-activating antibodies form and make immune complexes with heparin and platelet factor 4 (PF4).⁴⁷ This results in activation of platelets (via their Fc receptors) and formation of procoagulant platelet microparticles, leading to thrombin generation and thrombosis. Thus, the major problem in heparin-induced thrombocytopenia (HIT) is thrombocytopenia several days after heparin exposure accompanied by thrombosis, often in major vessels or structures. HIT appears to be less common, of milder course, and probably underrecognized in neonates and children. About 1% of children exposed to CPB have PF4 antibodies when tested before their second CPB, and actual HIT is much less common.⁴⁸ When HIT is suspected, either PF4 enzyme-linked immunosorbent assay or a functional assay for HIT can be used to make the diagnosis; if positive, no further heparin should be given. If CPB is necessary, alternatives, such as the direct thrombin inhibitors argatroban, lepirudin, and bivalirudin, may be used. None of these agents is approved for use in children for anticoagulation for CPB, but case reports and small series have documented their successful use when HIT is diagnosed.^49–51 The partial thromboplastin time (PTT), activated clotting time (ACT), and a specialized clotting time called the ecarin clotting time can be used to follow anticoagulation with these agents, but there is no reversal agent for them. Thus, treatment of post-CPB bleeding involves administration of blood products and coagulation factors as well as recombinant factor VIIa (see later discussion).

Antithrombin III Deficiency

Heparin produces anticoagulation by combining in a 1 : 1 ratio with antithrombin III (ATIII), which then binds to and inhibits thrombin, leading to anticoagulation. Of adult patients, 4% to 13% have a resistance to normal doses of heparin for CPB; most instances occur because of a partial deficiency of ATIII, rendering heparin less effective at producing anticoagulation.⁵² In children this is often unknown, and the first suspicion of ATIII deficiency may occur when the standard heparin dose of 300 to 400 units/kg fails to adequately anticoagulate before CPB, that is, the ACT remains less than 300 seconds. The usual response is to apply another dose of heparin from a different vial and remeasure the ACT, but if the ACT is still not adequately prolonged, a diagnosis of ATIII deficiency may be suspected. Infants less than 6 months of age and children with congenital heart disease have decreased ATIII concentrations.⁵³ Therefore, heparin may not achieve adequate anticoagulation, and disorders in hemostasis and thrombosis and an exaggerated inflammatory response may occur. In this case, blood can be sent for ATIII levels, but to proceed with CPB, the ATIII must be increased. This can be accomplished in two ways: (1) by supplementing ATIII with recombinant ATIII, 75 units/kg, and ensuring that the ACT is adequately prolonged before proceeding with CPB, or (2) by adding FFP (which has ample levels of ATIII) to the CPB prime or administering it to the child.^52,54

Recombinant Factor VIIa for Massive Hemorrhage

Recombinant factor VIIa (rFVIIa) was originally approved for use in patients with hemophilia who possess inhibitors to factors VIII or IX, and was shown to be effective at treating bleeding in these patients at doses of 90 µg/kg (see Chapter 10).⁵⁵ Endogenous factor VII circulates at low concentrations in the plasma. At a site of tissue or blood vessel injury, tissue factor (TF) is exposed, and the extrinsic coagulation pathway is activated by the binding of factor VII to TF, resulting in the activation of factor X to factor Xa, leading to the generation of thrombin from prothrombin, with further activation of platelets and the coagulation cascade.⁵⁶ High concentrations of rFVIIa result in major activation of the extrinsic pathway at the site of injury, theoretically without resulting in systemic hypercoagulability. However, thrombotic complications are increased after its use. rFVIIa also activates platelets, adding to the potential benefit of this agent in significant hemorrhage. Thus, this therapy seems attractive for the treatment of surgical bleeding. A recent analysis of factor VII concentrations during pediatric cardiac surgery recommended that it be used to treat postoperative coagulopathies that were resistant to conventional therapy, when no identifiable surgical cause of bleeding could be determined.⁵⁷

Because of its high cost, and the paucity of data in children undergoing cardiac surgery, rFVIIa should be reserved for life-threatening hemorrhage unresponsive to other measures. A dose of 45 to 90 µg/kg, repeated every 2 hours, can be used. rFVIIa cannot produce hemostasis alone and should only be administered after the transfusion of sufficient amounts of platelets, plasma, and fibrinogen to form the substrate for hemostasis.

Sickle Cell Disease

Sickle cell disease (SCD), one of the most common hemoglobinopathies among patients of African-American or West Indian origin (with a prevalence of 0.2% to 0.3% in that population), results from the substitution of valine for glutamic acid in position 6 of the β-hemoglobin chain. Normal adult hemoglobin is referred to as HbA, whereas hemoglobin containing the mutant β-hemoglobin chains is referred to as HbS. SCD is represented by a homozygous genotype (HbSS) with fractional concentrations of HbS in the range from 70% to 90%. Sickle cell trait, on the other hand, is a heterozygous manifestation (HbAS) with a prevalence of 8% to 10% in the same population. The definitive diagnosis of any sickle cell hemoglobinopathy is confirmed by hemoglobin electrophoresis (see Chapter 9).

Children with SCD are at a particular risk for perioperative complications.^58,59 Sickling can be triggered by hypoxia, dehydration, acidosis,⁶⁰ hypothermia, stress, and infections. Hypoxia induces opening of a Ca²⁺-activated K⁺ channel (Gardos channel) that causes intracellular dehydration.⁶¹ Chain formation occurs and leads to increased blood viscosity with vaso-occlusion. Opening of the Gardos channel is an important mechanism of sickle cell dehydration, which is temperature dependent, with greater potassium efflux at lower temperatures.⁶² Shrinkage of sickle erythrocytes may also result from activation of a K⁺/Cl⁻ cotransport pathway under acidotic conditions.⁶³ Activation of this pathway can be blocked by increasing the abnormally low level of intracellular magnesium in sickle erythrocytes. The use of magnesium and hydroxyurea in the perioperative period therefore seems to be beneficial.⁶⁴

CPB, particularly for more complex surgical procedures, may involve periods of low flow or even circulatory arrest, as well as hypothermia with consequent local vasoconstriction, hypoxemia, and acidosis. There is some evidence that CPB can be safely undertaken in SCD.⁶⁵ Flow conditions are an important determinant of sickle erythrocyte adherence to endothelium. Under low-flow conditions sickle cell adhesion to endothelium increases with contact time in the absence of endothelium activation or adhesive proteins, whereas under venular flow conditions sickle cell adhesion occurs only after endothelial activation. During CPB, both low-flow conditions and endothelial activation may occur. Multiple triggers of sickling are likely to occur during CPB, and close attention should be paid to the conduct of all aspects of bypass.

In the past, routine exchange transfusion has been recommended to prevent these complications.⁶⁶ More recent experience provides evidence that not all children require an exchange transfusion.⁶⁷ The growing evidence of the harmful effects of blood transfusion adds to the need to carefully reconsider routine exchange transfusion.⁶⁸ For uncomplicated bypass surgery without periods of cardiac arrest, the omission of exchange transfusion has led to good outcomes.

Guidelines have been proposed for the perioperative management of children with sickle cell disorders.⁶⁷ It is essential to avoid hypothermia using tepid or warm CPB in its stead; blood transfusion only for a decrease in hematocrit to less than 20%; maintenance of intravascular volume and body temperature while on CPB; the avoidance of vasopressors; the use of postoperative multimodal pain therapy; and early incentive spirometry to prevent pulmonary complications.⁶⁹ In our practice, we utilize cerebral near-infrared spectroscopy (NIRS) to help determine an acceptable hematocrit for the individual child.

For children undergoing hypothermia, successful management with⁷⁰ and without⁷¹ partial or complete exchange transfusion on bypass has been reported. Exchange transfusion can be performed preoperatively or on initiation of CPB.⁷² For exchange transfusion during CPB, the extracorporeal circuit is primed with blood and the usual components. When CPB is commenced, the child’s blood volume is drained into storage bags and separated. The platelet-rich plasma is reinfused at the end of CPB, and the concentrated sickle cells are discarded. Platelet and plasma sequestration in conjunction with exchange transfusion reduces the need for postoperative transfusion and protects the platelets from the negative effects of CPB.⁷³

There seems to be no consensus as to a suitable target HbS level. Reducing the absolute level of HbS may be of greater benefit than achieving a particular ratio of HbA to HbS because the remaining sickle-prone cells are still at risk for sickling.⁷⁴ In SCD, exchange transfusion has been shown to favorably affect cerebral tissue oxygenation.⁷⁵ Exchange transfusion will decrease both the proportion and absolute amount of HbS, but it does not remove every cell that may sickle; it may also have favorable effects on hypoxic pulmonary vasoconstriction.⁷⁵ In this context, these children may benefit from continuous hemofiltration to reduce inflammatory mediators and improve pulmonary recovery.⁷⁶ Inhaled nitric oxide also has been suggested as an adjunct for the prevention of sickle cell crisis. It may improve the binding of oxygen, thereby reducing the formation of sickle cells; reduce pulmonary hypertension; and improve pulmonary function without adverse effects on normal hemoglobin.⁷⁷

Pre-Bypass Period

This phase begins with surgical incision and lasts through initial dissection and preparation for cannulation. During this period transesophageal echocardiography (TEE) is performed to confirm the diagnosis and establish a basis for post-bypass comparison.

Cannulation and Initiation of Bypass

After sternotomy and mediastinal dissection, the aorta is cannulated, along with either the right atrium, if single venous drainage is planned, or the superior and inferior venae cavae for bicaval venous drainage. A large dose of heparin (300 to 400 units/kg) is administered intravenously, and the adequacy of anticoagulation is measured using the ACT before initiating CPB. The target ACT is usually 480 seconds. High ACTs are maintained during CPB with the addition of heparin to the prime as needed, because larger doses of heparin lead to a reduced degree of consumptive coagulopathy, which translates into reduced blood product therapy requirements.⁹⁶ Other methods of measuring anticoagulation include the Hepcon system (a plasma heparin concentration assay), which may allow for more accurate titration of heparin and protamine dosages.¹⁰⁶ The thromboelastogram may also be used as a baseline measure of the coagulation system and then may be repeated during bypass, with heparinase added to more objectively assess each child’s anticipated need for coagulation products.¹⁰⁷ An improved preservation of the hemostatic system with subsequent reduction of blood loss and a reduction in transfusion requirements has been demonstrated after maintenance of high heparin levels during CPB.¹⁰⁸ The additional maintenance of high ATIII concentrations may further contribute to a reduction of hemostatic activation.¹⁰⁹

In most centers, bicaval cannulation is used for all but the smallest children (less than 2 kg) to prevent venous return from interfering with the surgical field. A gradual transition to full CPB is then performed to minimize myocardial stress, using a prime that has essentially the same composition as the child’s blood with regard to temperature, pH, calcium, potassium, and hematocrit. CPB flows of 150 mL/kg/min are used for infants weighing less than 10 kg, and 2.4 L/min/m² is used for children weighing more than 10 kg. Flow rates may be reduced during periods of hypothermia (see later), although many centers now prefer to maintain greater flows throughout the bypass period. Misplaced cannulas can lead to significant morbidity. Obstruction of the inferior vena cava (IVC) by a misplaced IVC cannula can lead to increased venous pressure, which causes ascites and decreased perfusion pressure in mesenteric, hepatic, and renal vascular beds. Misplacement of the cannula in the superior vena cava can result in cerebral edema from inadequate venous drainage and a subsequent reduction in cerebral blood flow, potentially resulting in ischemia. Arterial cannula misplacement can also occur. If the cannula inadvertently slips beyond the takeoff of the right innominate artery, preferential perfusion to the left side of the brain can be observed. This can be detected on the NIRS monitor, which may be an important monitor, particularly in pediatric cardiac surgery.¹¹⁰

The presence of any anomalous systemic-to-pulmonary shunts can lead to shunting of blood away from the systemic circulation, through the pulmonary circuit, and then through the venous cannula to the CPB machine. Thus, the systemic perfusion is shunted away from the body in a futile circuit back to the CPB machine. Anatomic lesions where such shunting can occur include an unrecognized patent ductus arteriosus and large aortopulmonary collaterals, as found in pulmonary atresia. Bypass flow needs to be increased to compensate for these shunts.

Cooling Phase

Systemic cooling is used for nearly every case. Hypothermia is classified as mild (30° C to 36° C), moderate (22° C to 30° C), or deep (17° C to 22° C). In general, lower temperatures are used for more complex operations that carry a greater potential for requiring periods of low-flow bypass or circulatory arrest. Cooling is primarily achieved extracorporeally through the heat exchanger in the bypass circuit, although some surgeons also request that ice be applied to the head to prevent rewarming during the circulatory arrest.

Aortic Cross Clamping and Intracardiac Repair Phase

The aorta is cross clamped, with the heart then rendered asystolic after infusion of cardioplegia solution into the aortic root.

Deep Hypothermic Circulatory Arrest or Selective Cerebral Perfusion Phase

If circulatory arrest is to be used, it is initiated after a slow cooling period of at least 20 minutes, and an attempt is made to limit the total duration of deep hypothermic circulatory arrest (DHCA) to less than 40 minutes. Special bypass techniques (see later) have been developed to avoid the necessity of using DHCA and may also be performed during this time.

Removal of Aortic Cross Clamp and Rewarming Phase

After completion of the intracardiac repair and de-airing of the heart, the aortic cross clamp is removed, allowing reperfusion of the myocardium. Optimally, normal sinus rhythm and myocardial contractility are restored during this time, while the child is slowly rewarmed. During rewarming, surgery is completed, inotropic and vasoactive agents are started, and ventilation begins. Hemofiltration and blood transfusion are used to achieve the desired hematocrit. Left atrial and/or pulmonary artery monitoring lines, if indicated, are placed at this time, as are temporary atrial and ventricular pacing wires. If the child is incompletely rewarmed before separation from CPB, a significant afterdrop with precipitous post-bypass reduction in core body temperature can occur. This can lead to vasoconstriction, shivering, increased oxygen consumption, and acidosis. However, postischemic hyperthermia can lead to delayed neuronal cell death.¹¹¹ Mild degrees of hypothermia and certainly the avoidance of hyperthermia are essential in the perioperative period.¹¹² In children, rectal temperature mostly reflects peripheral temperature. One study showed that the temperature of the foot was more sensitive than the temperature of the hand.¹¹³ Another study revealed that for anatomic or physiologic reasons, temperature gradients in the toes develop more readily than those in the fingers.¹¹⁴ Several endpoints have been proposed, such as nasopharyngeal temperatures greater than 35.0° C, bladder temperature greater than 36.2° C, or skin temperatures greater than 30° C^115,116; we use an endpoint of 35.5° C rectal temperature.¹¹⁷

Separation From Bypass

The child’s core body temperature, hematocrit, and metabolic parameters should be optimized before attempting separation from CPB. Careful observation for left-sided air, confirmation with the TEE, and concurrent inspection of electrocardiogram (ECG) changes should continue throughout the weaning process, with the child in the Trendelenburg position and the aortic root vented. While fluid volume is gradually added to the child by reducing the outflow to the venous reservoir until optimal filling pressures are achieved, CPB flow is then slowly reduced to zero. If inotropic support is anticipated to separate from CPB, infusions should be prepared before beginning separation.

Post-Bypass Period

This phase lasts until chest closure and transfer to the ICU have been accomplished. During this time, modified ultrafiltration (MUF) may be performed for 10 to 15 minutes after cessation of CPB. Cardiac function and the quality of the surgical repair are assessed via TEE, and, if found to be satisfactory, protamine is administered to neutralize residual heparin. The usual dose of protamine is 1.0 to 1.3 mg/100 units of heparin given at the onset of bypass. Limiting protamine to this dose prevents an overdose with its associated effects on platelet function (reduction of the interaction of glycoprotein Ib receptor interaction with von Willebrand factor).¹¹⁸ If the ACT remains increased or prime blood is given back to the child, an additional 25% of the initial dose of protamine is added and the ACT is rechecked. However, particularly in infants, the administration of protamine and the persistent treatment of a suspected incomplete heparin reversal should not distract and delay the treatment of other commonly associated post-bypass coagulopathies, such as thrombocytopenia, platelet dysfunction, and other coagulation factor deficiencies.

Protamine reactions occur much less frequently in children younger than 16 years of age, approximately 1.76% to 2.88%.¹¹⁹ Independent risk factors are a female gender, a larger protamine dose, and smaller heparin doses. Type I reactions or effects during administration are rare and adding calcium does not change the hemodynamic consequences of injection.¹²⁰ Fortunately, severe anaphylactic reactions (type II) or catastrophic pulmonary vasoconstriction (type III) are rare but have been observed by us and others.¹²¹ Administering the protamine over no less than 5 minutes reduces the severity and precipitous nature of any protamine reaction.

Unstable neonates and small infants may have their sternums temporarily left open, with surgical closure planned 24 to 72 hours later when cardiac function has improved and myocardial edema diminished.

Because CPB can have a multitude of adverse physiologic effects, attempts are made to minimize both the duration of CPB and ischemic (aortic cross clamp) time; thus, as much of the surgery as possible is performed outside of these phases. In general, physiologic responses to bypass are more extreme with decreasing age and size of the child. The neonate experiences a greater degree of hemodilution on bypass and colder temperatures on bypass and frequently requires longer aortic cross clamp times, all of which can result in a greater inflammatory response. Table 17-4 summarizes clinical management issues during the major phases of CPB.

TABLE 17-4 Checklist for Bypass Management

CPB, Cardiopulmonary bypass; ECG, electrocardiogram; TEE, transesophageal echocardiography.

pH-Stat Versus α-Stat Management

Some degree of hypothermia is used for nearly every cardiac operation to slow the metabolism and oxygen consumption of all organs, particularly the brain and heart.¹²² During cooling, the carbon dioxide contained in blood becomes more soluble and its partial pressure decreases. The Paco₂ sensed by the body decreases as body temperature decreases, with the result that at a core temperature of 17° C to 18° C, if pH and Paco₂ have not been corrected for temperature, the body experiences a pH of about 7.6 and Paco₂ of 15 to 18 mm Hg (Fig. 17-2).¹²³ This very low Paco₂ causes cerebral vasoconstriction, particularly during the cooling phase of bypass, which in turn leads to less cerebral blood flow, less efficient brain cooling, and less cerebral protection at a given temperature.¹²⁴ Because blood samples are normally heated to 37° C before measurement of pH, Paco₂, and Pao₂, the use of pH-stat management indicates that blood gases are being corrected for the child’s actual body temperature by increasing the Paco₂ during bypass, as it is measured at 37° C, so that the body experiences a Paco₂ of approximately 40 mm Hg and a pH of 7.4 at all temperatures. Conversely, α-stat management means not correcting the blood gases for temperature, as if the child’s blood was always at 37° C, with the goal of pH 7.4 and Paco₂ 40 mm Hg. In the early days of CPB, pH-stat was used to preserve cerebral blood flow at all ages.¹²³ Subsequently, in the 1970s and 1980s, randomized controlled studies in adults undergoing CPB confirmed that acute, post-CPB neurologic problems were worse with the use of pH-stat management.¹²⁵ α-Stat management was, therefore, adopted for both adult and pediatric CPB. However, studies in a neonatal pig model have challenged this conclusion, proving that neurologic outcomes, both behavioral and neuropathologic, are significantly worse when α-stat management is used in infants.^124,126

FIGURE 17-2 pH and Pco₂ changes when blood temperature is varied between 17° C and 37° C. Point A is the starting point, with pH 7.4 and Pco₂ 40 mm Hg at 37° C. Points B, C, D, and E are the conditions the brain experiences at 17° C with various blood gas management strategies. pH-stat management (correcting the pH and Pco₂ for temperature) results in an acid-base environment that is neutral, whereas α-stat management (not correcting for temperature) results in a very alkalotic environment at 17° C. Warming the blood sample (as is done for blood gas measurement) results in very high Pco₂ values when pH-stat is used. The pH of blood becomes slightly more alkalotic with cooling, owing to the decreased dissociation of hydrogen ions.

(From Jonas RA. Carbon dioxide, pH, and oxygen management. In: Jonas RA, DiNardo J, Lawson PC, et al, editors. Comprehensive surgical management of congenital heart disease. London: Arnold Publishers; 2004, p. 151-60.)

Advantages of pH-stat CPB have been shown to include:

In cyanotic infants with aortopulmonary collaterals, pH-stat management results in significantly improved brain oxygenation as measured by near-infrared cerebral oximetry.¹³⁴ A retrospective study of 16 infants revealed worse neurodevelopmental outcomes with α-stat management.¹³⁵ In a randomized prospective trial of pH- versus α-stat management in 182 infants younger than 9 months of age, there was a strong trend toward improved outcomes with pH-stat management, including earlier return of electroencephalographic activity, fewer seizures, and improved psychomotor development index.¹³⁶ One study examined the effects of α-stat and pH-stat on developmental and neurologic outcomes after deep hypothermic CPB in infants.¹³⁷ Psychomotor Development Index scores of 110 patients did not differ significantly between the groups (P = .97). The results of the Mental Development Index scores were dependent on diagnosis. In all but the ventricular septal defect subgroup, the pH-stat group did not have statistically greater Mental Development Index scores. Abnormalities on the electroencephalogram (P = .77) and neurologic examination (P = .70) were similar with the two methods of blood gas management. The authors concluded that the use of α-stat or pH-stat strategy is not consistently associated with improved or impaired early neurodevelopmental outcomes in infants undergoing deep hypothermic CPB.¹³⁷ One reason for the differing results between pediatric and adult studies is that cerebral emboli are more common in the adult population because of the presence of atherosclerotic plaques, which cause microembolic infarcts, and the increased cerebral blood flow produced by pH-stat management, which leads to a greater number of cerebral emboli in adults. Emboli occur much less frequently in children, and the primary cause of neurologic injury from CPB in children is hypoxic-ischemic.¹³⁸ Thus, the increased cerebral blood flow observed on CPB with pH-stat management lessens this risk in children. Interestingly, this putative mechanism has been recently challenged by a study involving a controlled microembolic load and DHCA in pigs that revealed that pH-stat was still associated with improved outcomes when compared with α-stat.¹³⁹ pH-stat also improves oxygen delivery by counteracting the pH- and hypothermia-associated leftward shift in the oxyhemoglobin dissociation curve. Studies have also revealed a decrease in peak postoperative troponin levels, reduced ventilator dependence, and reduced ICU stays with pH-stat versus α-stat.¹⁴⁰ Most programs specializing in surgery for congenital heart defects currently use pH-stat management. This necessitates careful attention to Paco₂ during all phases of bypass, and possibly reducing the sweep gas flow into the CPB oxygenator (to decrease the efficiency of CO₂ removal), and often adding inspired CO₂ to the sweep gas of the bypass circuit, particularly in small infants.

Hematocrit on Bypass

The relatively small total blood volume in children, along with the volume required to prime the CPB circuit, means that adding blood to the CPB prime is mandatory for small infants. Practice is institution specific; but in many centers adding either whole blood, PRBCs with FFP (for children less than 8 kg), or PRBCs alone (for children less than 12 to 15 kg) is necessary to ensure that the hematocrit on bypass is not less than 20%. As a result of increased transfusion-related concerns from bloodborne viral disease transmission during the 1980s and 1990s, and given that a low hematocrit is thought to be necessary to ensure adequate blood flow through capillary beds (because the blood viscosity increases at low temperatures), hematocrits of 20% or less on CPB with deep hypothermia were frequently tolerated.¹⁴¹ There is increasing evidence that the practice of extreme hemodilution is detrimental to neurologic outcome in children. In a piglet model, one group of investigators determined that the incidence and degree of hypoxic-ischemic brain injury after a period of DHCA was significantly greater with a hematocrit of 20% versus one of 30%, regardless of whether pH- or α-stat strategy was used.¹⁴² In another piglet model, using intravital microscopy of pial capillaries during deep hypothermic CPB, a hematocrit of 30% did not impair cerebral microcirculation when compared with a hematocrit of 20%.¹⁴³ Finally, in a prospective randomized trial of CPB hematocrit of 20% versus 30% at the Children’s Hospital Boston, children in the lower hematocrit group demonstrated significantly reduced psychomotor development index scores 1 year after surgery.⁸³ In a follow-up study of hematocrit 25% vs. 35%, the same group did not observe a difference in neurodevelopmental outcomes.¹⁴⁴ However, when they combined all children from both hematocrit trials, they found that a hematocrit less than 24% was associated with lower psychomotor index scores 1 year after surgery.¹⁴⁵ The hypoxic-ischemic damage most likely occurs during the cooling and rewarming phases of bypass, when cerebral oxygen metabolism is not suppressed, yet hematocrit and, thus, oxygen delivery, is low. Therefore, many centers are now maintaining higher hematocrits on CPB (at least 25%), which either means using more donor blood products or using hemofiltration to raise the hematocrit during bypass. The current estimate by the American Red Cross of the risk of viral transmission from a single unit of blood is 1 in 205,000 for hepatitis B, 1 in 1.9 million for hepatitis C, and 1 in 2.1 million for human immunodeficiency virus.¹⁴⁶ The risk–benefit ratios therefore favor the greater hematocrit approach, a definitive change from previous practice patterns. Balanced against this practice of a greater hematocrit on CPB is the finding that greater transfusion of blood products in the intraoperative, and early postoperative periods, is associated with longer duration of mechanical ventilation in infants undergoing two-ventricle repairs.¹⁴⁷ Additional studies are required to optimize strategies for the use of blood products in infants and children undergoing CPB.

Flow Rates on Bypass

The traditional practice in many institutions has been to decrease CPB flows, particularly during hypothermia, to reduce the volume of blood returning to the surgical field and allow more efficient completion of the surgery, particularly in small infants. This concept has been questioned in recent years owing to the inability to determine the safe low-flow bypass rate in the individual child. One report studied 28 neonates who underwent arterial switch operation with α-stat blood gas management during CPB.¹⁴⁸ At 14° C to 15° C, bypass flow was sequentially reduced from 150 mL/kg/min to 50 mL/kg/min, and then further decreased in increments of 10 mL/kg/min until circulatory arrest was begun (to 0 mL/kg/min). All neonates had detectable cerebral blood flow by transcranial Doppler at CPB flows above 20 mL/kg/min, but one had no detectable perfusion at 20 mL/kg/min, and eight had none at 10 mL/kg/min, leading the authors to conclude that 30 mL/kg/min was the minimum acceptable flow in this population. A neonatal pig model determined that, at normothermia, bypass flows of at least 150 to 175 mL/kg/min were necessary to ensure full oxygenation of all end organs and tissues.¹⁴⁹ Clinical studies of a high-flow bypass strategy, which included flows of 150 mL/kg/min at all phases of bypass except during DHCA, minimal use of DHCA, and α-adrenergic receptor blockade with phenoxybenzamine to produce long-duration systemic vasodilation, demonstrated excellent short- and long-term clinical and neurodevelopmental outcomes, with no child scoring outside normal ranges for testing performed at a mean age of 9 years.¹⁵⁰ This strategy also has led to excellent early results for the Norwood operation, with an early perioperative survival of 83% for cases carried out from 1993 to 1999.¹⁵¹ During the same era, one report documented that 26.7% of arterial switch children had neurologic abnormalities and 55% had at least one abnormal area on neurodevelopmental testing (performed at a mean of 10 years) when DHCA and low-flow bypass had been used.¹⁵¹

Vasoconstriction and increased vascular resistance, resulting in uneven regional organ perfusion, are among the undesired side effects of CPB. Endogenous catecholamine production and the alkaline α-stat CPB technique, if used, are responsible for these effects. To be able to run full flow during hypothermic CPB without significant hypertension, vasodilators are often used. Agents currently used to provide systemic vasodilation and more even cooling and rewarming include phentolamine, nitroprusside, or nitroglycerin. Phenoxybenzamine, which is no longer available, was used as part of a treatment strategy after stage 1 palliation for hypoplastic left heart syndrome, and has been associated with improved outcome.^152,153 Phenoxybenzamine was more effective than sodium nitroprusside in improving peripheral circulation, as shown by temperature gradients intraoperatively.¹⁵⁴ Greater CPB flows are associated with an improved oxygen delivery, which can improve patient outcome.¹⁵⁵

Phentolamine is a nonselective competitive α₁ and α₂ catecholamine receptor blocker. It has a half-life of 19 minutes and is eliminated mainly by the kidneys. Through postsynaptic α₁ and α₂ receptor inhibition it has a vasodilating and hypotensive effect that can improve cardiovascular parameters and metabolic acidosis during CPB management.¹⁵⁶ In children receiving phentolamine, increasing lactate concentrations at the end of the CPB period show a steady state toward the end of the surgery, whereas lactate continues to rise in patients who did not receive phentolamine.¹⁵⁶ These findings suggest that the use of phentolamine limits lactic acid production during the hypothermic period and aids the disposal of lactic acid from tissues. Seelye and associates called the physiologic state after hypothermia the “oxygen debt repayment” period in infants.¹⁵⁷ Although it has a beneficial effect on CPB management, the potential harmful effects of phentolamine, especially on the brain, have still not been fully elucidated. One study provided evidence that phentolamine increases S100B protein and a parameter indicative of altered cerebrovascular resistance, the pulsatility index in the middle cerebral artery, in infants given phentolamine during open-heart surgery.¹⁵⁸

Nitroprusside has been used as an easily titratable agent with α-adrenergic receptor–blocking capacity. One study examined the effect of perioperative sodium nitroprusside application in 25 neonates undergoing an arterial switch operation for transposition of the great arteries.¹⁵⁹ In comparison to the pre-bypass values, a similar increase in the concentration of S100B protein was found 2 hours after the termination of CPB in the sodium nitroprusside-treated and nontreated neonates, which decreased over the subsequent 48 postoperative hours. However, significantly reduced post-bypass serum levels of S100B protein were found in the sodium nitroprusside-treated group after 24 and 48 hours of treatment.

Nitroglycerin has been used with the same success. The only proven benefit over other agents is its nitric oxide donation capacity.¹⁶⁰ In Japan, high-dose chlorpromazine has been used as part of a low-resistance strategy during CPB for the Norwood procedure.¹⁶¹

We routinely use phentolamine, 0.1 to 0.2 mg/kg, to provide normal CPB flow and mean arterial pressure in the range of the diastolic pressure. If hypotension develops during bypass, the flow should be increased up to 150% of predicted; also one should examine the acid-base status in conjunction with cerebral oxygenation and mixed venous saturations. Often, severe hemodilution with oxygen debt is the cause and should be treated as such. After exclusion, we treat the hypotension carefully with vasoconstrictors, knowing that normal systemic pressures will not restore splanchnic hypoperfusion¹⁶² and that vasoconstrictors will often lead to a greater base excess. Excessive α-adrenergic receptor blockade can be antagonized by vasopressin.¹⁶³ One study demonstrated that vasoconstrictor treatment results in more sodium bicarbonate to treat the acidosis and is associated with a later time to extubation and return of bowel function.¹⁶⁴ In conclusion, α-adrenergic receptor blockade during bypass should be considered because of its benefits for tissue perfusion, but carefully executed and balanced against potential drawbacks afterwards.

Conventional Ultrafiltration and Modified Ultrafiltration

Ultrafiltration involves placing a hemofilter (similar to those used for continuous arteriovenous or venovenous hemofiltration in the ICU) in the CPB circuit and has become the standard of care for nearly all programs that specialize in surgery for congenital heart defects.¹⁶⁵ Conventional ultrafiltration (CUF) is performed during CPB, with the filter placed between the arterial and venous sides of the CPB circuit. The hemofilter has thousands of fibers with pores, which allow water, electrolytes, and small molecules to be filtered out of the blood. Suction is applied to the hemofilter on CPB, and an ultrafiltrate of plasma is produced. Advantages of ultrafiltration include the ability to increase the hematocrit, fibrinogen, plasma proteins, and platelet count,^166,167 without necessitating further blood transfusion, the ability to remove excess free water and sodium (which contribute to excess intravascular volume, tissue edema, pulmonary and myocardial edema), as well as the ability to correct acid-base and electrolyte imbalances, and to remove small molecules, such as interleukins and tumor necrosis factor-α (TNF-α) in particular,¹⁶⁸ which are involved in the post-bypass inflammatory process.^169,170 This improves systolic and diastolic function of the myocardium and reduces endothelial dysfunction in the systemic and pulmonary vasculature.^170,171 Pulmonary function is better preserved, probably owing to a slight reduction in interleukin 6 (IL-6) and thromboxane B₂,¹⁷² even though this is not a consistent finding in the literature.^173,174 Endothelin-1, another mediator of pulmonary damage and hypertension, was not reduced by any filtration method.¹⁷⁴ Clinically, however, any ultrafiltration method seems to benefit children, especially those undergoing complex repairs, neonates, and children with preexisting pulmonary hypertension.¹⁷³

Modified ultrafiltration (MUF) is performed for 10 to 15 minutes immediately after the conclusion of CPB. It can be performed in an arteriovenous manner with a hemofilter placed between the aortic cannula and the IVC cannula, or in a venovenous fashion using bicaval cannulation or an internal jugular venous catheter.¹⁷⁵ It was developed in 1991¹⁷⁶ as an alternative method to reduce the side effects of CPB. CUF during bypass is often limited by the minimal venous reservoir levels and requires the addition of crystalloid or colloid to be able to continuously remove cytokines during ultrafiltration. During MUF, blood passes out of the aorta, through the hemofilter, and is returned through the IVC cannula. The theoretical advantage of MUF over CUF is that only the child’s blood volume is filtered, yielding a more efficient system for achieving the goals just outlined. The disadvantages are that the child remains heparinized, and body temperature may decrease during the process (unless the circuit is modified to include the heat exchanger).¹⁷⁷ It also requires extra time, an aortic cannula is needed that can obstruct the aorta in small infants, and acute intravascular volume shifts may occur at a time when the child is prone to hemodynamic instability. Opposite to the expected effects of fluid removal, MUF actually increases arterial pressures despite decreasing filling pressures and improving myocardial performance.¹⁷⁸

There is increasing evidence that the use of ultrafiltration reduces bypass-related postoperative morbidity. Outcome studies have demonstrated that ultrafiltration improves myocardial and pulmonary function, lessens tissue edema, allows faster weaning from mechanical ventilation, and decreases the need for inotropic support.¹⁷⁹ In that aspect it may be as efficient as the perioperative application of steroids.¹⁸⁰ The reduction of inflammatory transmitters is only temporary, because the levels of cytokines are similar after 24 hours.¹⁸¹

Although each method has its proponents, and some centers perform both techniques in the same children, controlled comparative studies revealed no difference in outcome between MUF and CUF.^179,182 We routinely use a balanced ultrafiltration technique for all cases on CPB because it removes fluids and cytokines, as well as reduces lactate, which can aggravate reperfusion injury.¹⁸³

Changes in Pharmacokinetics

The initiation of CPB introduces additional volume to the intravascular space (hemodilution). This greatly affects drug distribution, plasma concentrations, and elimination. The major factors responsible for this are hemodilution and altered plasma protein binding,¹⁸⁴ hypotension, hypothermia,¹⁸⁵ pulsatility,¹⁸⁶ isolation of the lungs from the circulation, and uptake of anesthetic drugs by the bypass circuit.^187,188 Drugs in the blood exist in the free (unbound and therefore the active form) or plasma bound (inactive form bound to protein, e.g., albumin) forms and therefore are subject to marked changes with alterations in plasma protein levels. CPB alters all these factors, which makes description of pharmacokinetic parameters during CPB problematic. The greatest changes occur within 5 minutes of initiation of CPB. The addition of the prime volume immediately reduces the protein concentration, and the ratio of bound-to-free drug in the circulation changes. A reduction in red blood cell concentration occurs, and this reduces the free drug concentrations. This will reduce the amount of drug available for interaction with the receptors. Most studies show a reduction in total drug concentration in plasma with little change in unbound drug concentration over time, whereas on CPB other than the transient (less than 5 minutes) reduction at initiation of CPB¹⁸⁹ it would appear that the greatest risk for unwanted “lightening of anesthesia” is within this time frame, and additional doses of fentanyl, muscle relaxant, and midazolam are generally administered just before or with the onset of CPB. The explanation for why unbound drug concentrations are sustained during CPB is that the volume of distribution for most anesthetic agents is large relative to the volume of the CPB prime and serves as a huge reservoir for drug after intravenous administration. A decrease in the plasma concentrations of medications as a result of hemodilution shifts drugs down their concentration gradient from tissue to plasma. Hypothermia contributes to the changes in plasma concentrations primarily by depressing enzyme function and slowing the metabolism of medications. Drug metabolism is diminished during hypothermia; enzyme activity is approximately halved for every 10° C reduction in temperature. This may increase the free drug available for binding. When normothermia is reestablished, reperfusion of tissues might lead to washout of drug sequestered during the hypothermic CPB period. This may explain the secondary increases in plasma concentrations of opioids reported during the rewarming phase. pH-stat management also affects the degree of ionization and protein binding of certain medications, leading to increased unbound drug. During CPB, the lungs are out of circuit and medications that are taken up by the lungs (e.g., opioids) are sequestered during CPB. These medications are released when systemic reperfusion is established and concentrations are transiently increased. The volume of distribution of many drugs is expanded because of the priming volume of the bypass circuit, especially with neonates and small infants, where the priming volume is often greater than the child’s blood volume. Finally, medications may be taken up by various components of the CPB circuit itself.

Changes in Pharmacodynamics

The pharmacodynamic effects of anesthetic agents are affected primarily via the central nervous system, which undergoes major changes during CPB. For example, hypothermia during CPB reduces anesthetic requirements. Hypothermia causes a host of other effects, including decreases in receptor affinity (e.g., decreased opioid receptor affinity¹⁹⁰ and nicotinic acetylcholine receptor sensitivity¹⁹¹), enhanced effects of neuromuscular receptor blocking drugs at the neuromuscular junction,^192,193 and alterations in tissue blood flow that may affect the response to catecholamines.¹⁹⁴

CPB also affects the degree of ionization and protein binding (hence free or unbound drug concentrations) of weak acids and bases, as well as the electrolyte balance achieved by the blood gas management strategy used during CPB. Plasma concentrations of calcium, magnesium, and potassium decrease during CPB,^195,196 and these changes may lead to muscle weakness, dysrhythmias, and digitalis toxicity. The number of receptors available for interaction with a ligand will determine the subsequent magnitude of a drug effect. A reduction in the number of cardiac receptors has been observed in congestive heart failure, and defects in receptor transduction, as well as impairment of synthesis and reuptake of norepinephrine occur.

Administration of β-adrenergic agonists in this condition has been associated with further reductions in β-receptor numbers, with diminished pharmacologic effect. Removal of β-adrenergic blockade may lead to β-adrenergic receptor upregulation and increased adrenergic responsiveness.¹⁹⁷ Changes in receptor density and function may occur very quickly and have been observed to occur during cardiac surgery. Many perfusionists, under the direction of the anesthesiologist, can also administer inhalation agents via a separate vaporizer mounted on the bypass machine. Anesthetic requirements decrease with systemic hypothermia,¹⁹⁸ but as rewarming is initiated, additional anesthetic drugs, including a benzodiazepine, are added to the pump to ensure that amnesia is maintained. Further work is required to elucidate the mechanisms and clinical implications of these acute changes in receptor density and function.

Management of Deep Hypothermic Circulatory Arrest

In the early days of cardiac surgery and CPB, hypothermia was used to improve intracardiac surgical exposure. In 1950, Bigelow and colleagues were the first to show that hypothermia decreases the metabolic rate.¹⁹⁹ Since then, we have discovered other advantages of hypothermia, including decreases in the inflammatory response of CPB,²⁰⁰ decreases in blood loss,²⁰¹ myocardial protection,²⁰² and neuroprotection.²⁰³ The last effect relates primarily to the decrease in the metabolic rate (by approximately 64%) that is achieved by cooling from 37° C to 27° C. The disadvantages of DHCA include a prolongation of CPB and a greater tendency toward postoperative bleeding.²⁰⁴ Postoperative recovery, however, is not prolonged by hypothermia.²⁰⁵ The rate of wound infection is uninfluenced by hypothermic bypass.²⁰⁶

Hypothermia during cardiac surgery gained widespread acceptance only after the development of a heat exchanger that could be integrated into the CPB machine in 1959.²⁰⁷ Deep hypothermic circulatory arrest involves cooling the child’s body temperature during CPB to 17° C to 18° C, stopping the bypass machine, draining the blood from the child into a venous reservoir, and removing the cannulas from the heart. After the first reports of DHCA in the 1960s, this technique gained popularity in the 1970s and 1980s because the bloodless field it provided, thus facilitating complex intracardiac and aortic repairs in neonates and small infants,²⁰⁸ as well as reducing myocardial edema. However, it soon became evident that DHCA was associated with neurologic morbidity. Choreoathetosis, seizures, coma, and hemiparesis were all noted, especially with prolonged (>60 min) DHCA. The incidence of these acute morbidities seemed to increase when α-stat management became the accepted standard in many centers. Long-term adverse neurodevelopmental outcomes have also been associated with long periods of DHCA, including abnormalities in mental development and in fine and gross motor skills.²⁰⁹ The Boston Circulatory Arrest Study is a remarkable achievement in which 155 neonates undergoing the arterial switch operation from 1988 to 1992 were studied, with follow-up complete to 8 years of age.²¹⁰ The CPB protocol in those years included α-stat management, routine hemodilution to a hematocrit of 20%, and the absence of an arterial filter on the CPB circuit. A DHCA time of greater than 40 minutes was associated with a significant increase in adverse long-term neurologic outcomes (Fig. 17-3). Although the 40-minute cutoff is now well-accepted in surgery for congenital heart defects, a number of changes have subsequently been made to bypass protocols. Results from animal experiments using a neonatal pig model of DHCA, as well as data from the Boston Circulatory Arrest Study, led to the following recommendations for increasing the child’s safety margin when using DHCA:

FIGURE 17-3 Safe duration of circulatory arrest. From the Boston Circulatory Arrest Study, 155 8-year-olds underwent arterial switch operation for d-transposition of the great arteries as neonates. Bypass protocol used α-stat pH management, with hematocrit of 20% and temperature of 18° C. Above 40 minutes DHCA, both mental and physical performance test scores decreased significantly.

(From Wypij D, Newburger JW, Rappaport LA, et al. The effect of duration of deep hypothermic circulatory arrest in infant heart surgery on late neurodevelopment: the Boston Circulatory Arrest Trial. J Thorac Cardiovasc Surg 2003;126:1397-403.)

FIGURE 17-4 Cortical oxygen saturation (So₂) during deep hypothermic circulatory arrest in pH-stat and α-stat groups. The cortical So₂ half-life during arrest was significantly greater in the pH-stat than in the α-stat group. Mean ± SD, eight animals per group. *P < .05 between groups.

(From Kurth CD, O’Rourke MM, O’Hara IB. Comparison of pH-stat and alpha-stat cardiopulmonary bypass on cerebral oxygenation and blood flow in relation to hypothermic circulatory arrest in piglets. Anesthesiology 1998;89:110-8.)

Neurologic monitoring (see later discussion) may be useful in the individual child to aid in determining the safe duration of DHCA.^124,216

Although there are situations in which DHCA must be used, many surgeons are avoiding it whenever possible, minimizing its duration and dividing the periods of its use, or using alternate methods, such as selective cerebral perfusion (see the following discussion).

Regional Cerebral Perfusion

To avoid the use of DHCA, several novel CPB techniques have been developed. The purpose of these techniques is to allow perfusion of the brain during critical periods of surgery, such as aortic reconstruction during the Norwood operation.^217,218 These techniques are collectively referred to as selective cerebral perfusion. Regional cerebral perfusion (RCP) is one variation in which a small Gore-Tex graft of 3 to 4 mm is sewn onto the innominate artery before initiation of CPB and is then used as the aortic cannula during CPB (Fig. 17-5). During aortic reconstruction, snares are placed around the brachiocephalic vessels and CPB flow is decreased, with only the brain receiving perfusion via the right carotid artery during this period. In this way, a bloodless operative field is achieved, just as if DHCA was being performed, yet the brain is still receiving blood flow and oxygen, theoretically increasing protection from hypoxic ischemic brain injury. Another potential advantage of this technique occurs in neonates, who frequently have extensive arterial collaterals between the proximal branches of the aorta and the lower body via the internal mammary and long thoracic arteries. In this instance, the use of selective cerebral perfusion also provides some blood flow to the lower body, protecting renal, hepatic, and gastrointestinal systems from hypoxic damage as well.²¹⁹ This protection is, however, incomplete and RCP at 25° C is no more protective than DHCA.²²⁰ Also, the ongoing perfusion prolongs the effective bypass time, leading to more cytokine release and capillary leakage, with worse pulmonary function, more weight gain, and decreased right ventricular function.²²¹

FIGURE 17-5 Selective cerebral perfusion for the Norwood stage I palliation for hypoplastic left heart syndrome. Arterial inflow for bypass is provided by a small polytetrafluoroethylene graft sewn to the right innominate artery. Instead of deep hypothermic circulatory arrest, flow is provided to the brain at low rates, while the brachiocephalic vessels and descending thoracic aorta are snared, providing a bloodless operating field.

(From Pigula FA, Nemoto EM, Griffith BP, Siewers RD: Regional low-flow perfusion provides cerebral circulatory support during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 2000; 119:331-339.)

Despite the theoretical advantages of selective cerebral perfusion and a study demonstrating that selective cerebral perfusion does provide oxygenated blood flow to both cerebral hemispheres,²²² no long-term outcome studies have been performed that prove it is superior to standard techniques. This may in part be related to the novelty of this procedure. Neurologic monitoring has been used to determine the flow rate that is necessary during RCP.^222,223 We use 40% to 50% of full flow and adjust it according to brain saturation or Doppler measurements, maintaining baseline saturation before the onset of RCP. If a left radial arterial line or a femoral arterial line (or umbilical arterial line) is in place, an abdominal perfusion pressure of about 12 mm Hg, which correlates with radial artery pressures of 25 to 30 mm Hg, is the goal.²²²

Cardiac Effects

In addition to myocardial ischemic injury secondary to aortic cross clamping, several other factors can contribute to perioperative myocardial dysfunction. The first is entrainment of air into the coronary arteries, which frequently occurs during weaning from bypass.²²⁴ Despite meticulous de-airing of the heart, air may enter the right coronary artery, producing ischemia that is heralded by a pale myocardium, poor contractility, and ST-segment elevation of the ECG. Should this occur, appropriate management involves remaining on CPB, increasing perfusion pressure, and “milking” the air through the coronary arteries, allowing time for recovery of the ECG and ventricular function before attempting to wean from bypass. Surgical factors, such as reimplantation of coronary arteries with possible resultant ischemia or residual surgical defects, can also occasionally contribute to myocardial dysfunction.

The inflammatory response to CPB (see later discussion) has important implications for cardiac function.²²⁵ This systemic response results in a capillary leak syndrome, which in turn leads to accumulation of edema fluid in interstitial and extravascular spaces, including the myocardium.²²⁶ Myocardial edema can contribute to post-CPB myocardial dysfunction by impairing diastolic function and causing mechanical limitation of cardiac filling and outflow in small infants whose sternums have been closed. Additionally, myocardial edema has been implicated as a causative factor in the frequent decline in myocardial function that occurs 6 to 12 hours after conclusion of CPB. Inflammatory mediators also affect the responsiveness of the myocardium to catecholamines by interfering with their binding to the cell surface receptors,²²⁷ rendering exogenously administered drugs, such as dopamine and epinephrine, as well as the child’s endogenous catecholamines, less effective at increasing cardiac output in the perioperative period.

Mechanisms for prevention and treatment of myocardial dysfunction include the use of ultrafiltration and antiinflammatory drugs, such as corticosteroids and aprotinin.^228,229 The prophylactic use of noncatecholamine inotropic agents, such as milrinone, has also been shown to prevent low cardiac output syndrome in infants, even if cardiac function is adequate in the immediate postoperative period.²³⁰

Systemic and Pulmonary Vasculature Effects

The inflammatory response to CPB often produces mediators that directly increase pulmonary and systemic vascular resistance. These include interleukins, leukotrienes, and endothelin.²³¹ Indeed, when pulmonary artery pressure is measured directly, it is often significantly increased immediately after bypass, even if surgical results are optimal. This increase can be extremely detrimental in children with large left-to-right shunts, those undergoing cardiac transplantation secondary to dilated cardiomyopathy, and those undergoing bidirectional cavopulmonary anastomosis, where right ventricular output depends on maintaining low PVR. Prevention and treatment of increases in PVR include maintaining an adequate depth of anesthesia, ventilating with 100% oxygen, and judicious use of hyperventilation. Milrinone will increase right-sided heart output via its actions as both an inotropic agent and a pulmonary vasodilator. When PVR is significantly increased, inhaled nitric oxide is often used to assist in the early postoperative period.²³² Although effective, its cost is not inconsequential, and because PVR almost always decreases with time, nitric oxide is generally reserved for selected cases of pulmonary hypertension. Other simpler, less expensive treatments include oral or intravenous sildenafil^233,234 and inhaled nebulized prostacyclin.²³⁵

Pulmonary Effects

The lungs are not ventilated during CPB and are usually totally collapsed by intention, with the ventilator circuit disconnected, especially in small infants. This leads to significant atelectasis. The lungs are also at least partially ischemic during the bypass period, resulting in decreased production and alveolar levels of surfactant after CPB.²³⁶ In addition, reperfusion injury (pulmonary edema or hemorrhage after a sudden increase in pulmonary flow) can also occur after creation of a systemic-to-pulmonary artery shunt or pulmonary artery unifocalization. Inflammatory mediators liberated by the bypass run, also predispose to increases in smooth muscle tone and resistance, and can result in bronchospasm.²³⁷

In addition to complement, endotoxins and certain cytokines can also activate neutrophils and attract them toward sites of inflammation.²³⁸ In animal studies, endotoxin-induced lung injury can lead to rapid (within 45 minutes) accumulation of neutrophils within lung capillaries. Activation of neutrophils, with upregulation of adhesion molecules, neutrophil adhesion to the endothelium of lung vessels, and endothelial damage through proteases, appears to be the main step of the underlying pathophysiologic mechanism (Fig. 17-6). Macrophages play an important role in the evolution of the inflammatory acute lung injury through the secretion of cytokines, cytotoxic metabolites, and chemoattractants for leukocytes. At the clinical level, acute respiratory distress syndrome (ARDS) is often only one part of multiorgan failure, and lung injury should be seen as part of a more general state of systemic inflammation. The reported prevalence of ARDS after CPB in adults is 0.5% to 1.7%; the incidence in children is unknown. Interestingly, general hypothermia at 28° C failed to prevent the loss of ATP and the accumulation of lactate in lungs.²³⁹ Other methods that aim to protect the lungs during CPB, such as continuous lung perfusion, pneumoplegia, and nitric oxide ventilation at lung reperfusion, prevent more severe hemodynamic deterioration and preserve reactivity of the pulmonary vasculature, but fail to prevent pulmonary dysfunction.

FIGURE 17-6 Leukocytes, endothelial cells (EC), and humoral inflammatory mediators have been shown to play an important role in the cardiopulmonary bypass-induced lung injury. Complement (C) activation and complement-independent mechanical injury activates leukocytes, which, in their turn, secrete several inflammatory mediators, such as proteases and cytokines. Complement, cytokines, and ischemia-reperfusion also activate endothelial cells. Endotoxin, probably released from intestinal bacteria, exerts similar effects on leukocytes and endothelium. This process leads to disruption of endothelial and epithelial integrity and allows albumin, plasma, and activated leukocytes to enter the interstitial and alveolar space, causing tissue edema and reducing pulmonary compliance and blood oxygenation. AAM, Arachidonic acid metabolites; ICAM-1, intercellular adhesion molecule 1; IL, interleukin; IS, interstitial space; M, monocyte; Mg, macrophage; MPO, myeloperoxidase; N, neutrophil; NO, nitric oxide; O₂-DFR, oxygen-derived free radicals; P, platelet; PAF, platelet-activating factor; PDGF, platelet-derived growth factor; Pn, pneumocyte; TGF, tumor growth factor; TNF, tumor necrosis factor.

(From Asimakopoulos G, Smith PL, Ratnatunga CP, Taylor KM. Lung injury and acute respiratory distress syndrome after cardiopulmonary bypass. Ann Thorac Surg 1999;1107-15.)

The severity of pulmonary dysfunction after CPB can be measured via changes in the alveolar-arterial oxygenation gradient, intrapulmonary shunt, degree of pulmonary edema, pulmonary compliance, and PVR. Treatment of pulmonary atelectasis includes careful reinflation of the lungs when weaning from bypass (by administering several vital capacity breaths), gentle but thorough suctioning of the tracheal tube, and prophylactic use of inhaled bronchodilators before separation from CPB. Using these measures, pulmonary function has been shown to improve immediately in most children with large left-to-right shunts, with the duration of CPB seemingly having little effect on pulmonary outcomes.²⁴⁰ Thus, CPB itself has little effect on pulmonary function in most children. There is still an occasional child, however, who experiences classic “pump lung” ARDS, caused by the factors noted earlier. Treatment is supportive as for anyone with ARDS.

Neurologic Monitoring and Effects of Cardiopulmonary Bypass on the Brain

Cerebral monitoring can help to detect those children who are at risk for neurologic sequelae after bypass, promptly recognize and treat changes in cerebral blood flow/oxygenation, evaluate the effect of therapeutic interventions on cerebral physiology, optimize brain protection during the vulnerable periods of CPB, and potentially improve short- and long-term neurologic outcomes.²⁴¹

The cerebral NIRS monitor measures brain tissue oxygenation. This device noninvasively measures the cerebral tissue oxygen saturation and displays a numerical value for the cerebral regional oxygen saturation (rSo₂), the ratio of oxyhemoglobin to total hemoglobin in the light path. rSo₂ is a measure of local microcirculatory oxygen supply-and-demand balance and is reported on a scale from 15% to 95%. It has been assumed from anatomic models that 75% of the cerebral blood volume in the light path is venous and 25% is arterial. One study verified this in children with congenital heart disease by directly measuring the jugular venous bulb and arterial oxygen saturations and comparing these with the cerebral oxygen saturation measured with NIRS.²⁴² The actual ratio in children varied widely, but on average the venous to arterial ratio was 85 : 15. All devices measure both the arterial and venous blood oxygen saturations. Accordingly, this device does not provide a measure of the jugular venous bulb oxygen saturation (Sjvo₂). A corollary of this is that maneuvers that increase arterial oxygen saturation (e.g., increasing Fio₂) increase cerebral oxygenation as measured by these devices, although the Sjvo₂ may remain unchanged. In a study of 40 infants and children with congenital heart disease who were undergoing cardiac surgery or catheterization, NIRS correlated poorly with Sjvo₂ measurements, except in infants younger than 1 year of age.²⁴³ In contrast, in a study of 30 children undergoing cardiac catheterization, NIRS correlated very well with Sjvo₂ (r = .93).²⁴⁴ These data suggest that NIRS is a useful indicator of trends in cerebral oxygenation in individual infants and children. NIRS values also correlate with long-term neurodevelopmental outcomes in infant heart surgery. In a prospective study of 104 two-ventricle repairs, low rSo₂ in the intraoperative period did not correlate with death or major morbidity.²⁴⁵ However, when these children underwent neurodevelopmental testing at 1 year of age, lower average and minimum rSo₂ in the 60 minute period immediately after CPB correlated with worse psychomotor development index scores. Low rSo₂ also correlated with remote ischemic changes on brain MRI at 1 year of age.²⁴⁶

Systemic Inflammatory Response Syndrome

In cardiac surgery, SIRS is thought to result from four main sources of injury: (1) contact of the blood components with the artificial surface of the bypass circuit, (2) ischemia-reperfusion injury, (3) endotoxemia, and (4) operative trauma. Inflammatory cytokines, together with endothelial activation and endothelial-leukocyte interactions, appear to play an important role in the induction of this systemic inflammatory response.

Exposure of blood to the artificial materials in the bypass circuit—plastics, polypropylene oxygenator fibers, and metal suction devices—initiates a cascade of inflammatory responses, including activation of the complement system, the kallikrein system, and the coagulation system.²²⁶ As a result, interleukins, tumor necrosis factor, endotoxin, heat shock protein, and many other inflammatory mediators are released into the circulation. Leukocyte activation also results in secretion of inflammatory mediators, such as proteases and cytokines (e.g., TNF-α and IL-1), which are secreted early in the evolution of the inflammatory process. This chemokine-mediated increased leukocyte activation constitutes an important link in the chain of the propagation of the inflammatory response (see Fig. 17-6).

This inflammatory response is counterbalanced by a complex system of inhibitors, such as IL-10 and soluble cytokine receptors.²⁵¹ Also, the inflammatory response of the neonate may be more exaggerated than that of the infant or older child,²⁵² justifying a more aggressive approach to its modulation in the neonate (see later discussion).

A number of novel treatments have been studied, including monoclonal antibodies for inflammatory products, such as complement, endotoxin, and tumor necrosis factor. Although theoretically attractive, no clinical difference has been noted with any of these treatments.

Effective treatments used every day in the operating room and ICU include:

Corticosteroids interrupt the inflammatory response at several levels by entering cell nuclei and changing the rate of transcription of inflammatory molecules. Increasing evidence suggests that glucocorticoids act by regulating transcription or translation of antiinflammatory cytokines, such as IL-10, and altering expression of other proteins, such as endothelin-1 and inhibitor NF-κβ.^255,256 Because these processes take time to develop, the effects of corticosteroids are not immediate, taking up to several hours.²⁵⁷ Thus, the common practice of adding corticosteroids to the CPB prime will not fully prevent the inflammatory response²⁵⁸; to be effective, corticosteroids may need to be administered 4 or more hours before the onset of CPB.²⁵⁹

Despite these theoretical advantages of using corticosteroids to modulate the inflammatory response, a recent large discharge database review of over 46,000 infants and children, in which 54% did receive corticosteroids, demonstrated no difference in mortality. Using propensity score matching the authors concluded that corticosteroids were associated with greater length of hospital stay, greater rate of infection, and greater use of insulin. There was no difference in duration of ventilation. Steroids conferred no significant benefit; conversely in the simpler surgery categories, there was increased morbidity with these drugs.²⁶⁰ Although inflammatory activation from CPB definitely occurs, and it would seem intuitive that this would lead to worse outcomes in those patients with excessive inflammation, in contemporary practice, the correlation of the magnitude of this response and length of ICU stay and blood product administration, is statistically significant, but clinically modest, accounting for only 4% to 9% of the difference in these variables, in a large study of infants undergoing two-ventricle repairs.²⁶¹

Coagulation Effects

Blood coagulation is frequently abnormal after CPB for several reasons. The inflammatory cascade activates the coagulation system, resulting in factor consumption and fibrinolysis, which, in turn, breaks down existing blood clots, leading to increased bleeding.¹⁶ Treatment is adequate heparinization, reversal with protamine, and the use of an antifibrinolytic to inhibit fibrinolysis and improve platelet function.³³ In addition, the smaller the child, the greater the dilution of clotting factors by the bypass prime, and the greater the risk for low concentrations of clotting proteins and fibrinogen postoperatively. Platelets are also degranulated and consumed by the CPB circuit, leading both to low platelet counts and nonfunctioning platelets.¹⁶ The smaller the infant, the greater the duration of bypass, and the more complicated the surgery, the greater the incidence of coagulopathy after bypass. Efforts to minimize the post-bypass coagulopathy in infants includes priming the CPB circuit with fresh whole blood for small infants, if available, or packed cells plus fresh frozen plasma (FFP) if fresh whole blood cannot be obtained.^19,262 Treatment involves administration of platelets to small infants as the first line of therapy, followed by the replacement of fibrinogen and FFP to replace clotting factors. If these factors are not effective after correcting coagulation parameters, such as platelet count, prothrombin/PTT, fibrinogen, and thromboelastogram, then surgical bleeding may be the cause and surgical reexploration may be warranted.²⁶³ Factor VIIa has also been used as a last resort in children who have significant post-bypass bleeding unresponsive to standard measures.⁵⁶

Hepatic, Renal, and Gastrointestinal Effects

The liver, kidneys, and gastrointestinal tract, like the brain and heart, may be rendered ischemic by prolonged CPB, DHCA, or low cardiac output syndrome. Renal function is compromised on CPB. This is manifested by the appearance of proteinuria and impaired tubular cellular function immediately after CPB. Renal dysfunction from ischemia is also common. Low urine output may occur secondary to secretion of antidiuretic hormone, a response to surgical stress. However, the latter appears to be transitory and usually resolves spontaneously.²⁶⁴ The incidence of acute renal dysfunction after surgery with bypass to correct congenital heart defects is 17%, ranging from 0.7% for arterial septal defect closure to 59% for arterial switch operations.²⁶⁵ Deep hypothermic cardiac arrest subjects the kidney to additional ischemia reperfusion injury.²⁶⁶Acute renal failure after CPB is uncommon in children, with fewer than 3% requiring dialysis perioperatively.^265,267 Infants who undergo cardiac surgery routinely receive diuretics or a peritoneal dialysis catheter, the latter prophylactically in some instances.^268,269 Although some have attributed the improved survival with early peritoneal dialysis to the prevention of fluid overload, others have attributed it to a more rapid clearance of CPB-induced proinflammatory cytokines.²⁷⁰ Further study is required to clarify the mechanism of action of early peritoneal dialysis. In our center, neonates and children with a complex heart defect usually receive peritoneal dialysis immediately postoperatively to prevent fluid overload.

Recovery of hepatic and gastrointestinal function follows hemodynamic recovery, but may require several days. Therapy is mainly supportive. Splanchnic and renal perfusion can be monitored noninvasively using somatic oximetry. Somatic oxygenation may predict renal dysfunction and predict organ failure. Interventions based on the somatic NIRS may improve outcome.²⁷¹

Immune System Effects

Leukocytes are activated by the CPB circuit, although their numbers may be depleted by leukocyte filters, which are sometimes used to attenuate the inflammatory response. Despite the theoretical possibility that this may increase the risk of infection or neutrophil dysfunction, this has not been observed in published studies or clinical practice.²⁷²

Endocrine System Effects

The magnitude of the inflammatory and endocrine responses after cardiac surgery depends in part on the duration of the surgical procedure and CPB.²⁷³ In children undergoing brief operating times, postoperative blood concentrations of cortisol, adrenocorticotropic hormone, and β-endorphins are significantly greater than those in children undergoing prolonged operation times. In contrast, the serum concentrations of the proinflammatory cytokines IL-6, IL-1β, and TNF-α are similar in the two groups. Adrenocorticotropic hormone and cortisol concentrations correlated positively with the blood concentrations of IL-1β, IL-6, and TNF-α in the group of children with prolonged operation times.

The plasma concentrations of both epinephrine and cortisol increase after cardiac surgery.²⁷⁴ In children, pre- and post-bypass cortisol and norepinephrine increase significantly during isoflurane anesthesia when 2 µg/kg of fentanyl is used rather than 25, 50, 100, or 150 µg/kg.²⁷⁵ No significant increase in the blood concentrations of these hormones occurred with any of the fentanyl doses of 25 µg/kg or greater. In addition to cardiovascular stability, continued use of larger doses of opioids during bypass minimizes the stress responses and stabilizes hemodynamics during and after bypass, but may delay recovery.²⁷⁶ Also, growth hormone, glucose and insulin, lactate, glutamate, aspartate, and free fatty acid concentrations increase after cardiac surgery, whereas total triiodothyronine concentrations decrease.²⁷⁷ Limiting the amount of opioids balances the negative effects of inflammation and stress with the opportunity to fast-track children’s recovery after surgery for congenital heart disease.^278,279