The roller head pump, shown in Figure 28.2A, consists of a raceway where two roller spindles oriented 180° from each other rotate around a central axis and compress the raceway tubing. This raceway tubing is also referred to as the “boot line.” The double-headed roller pump is a positive volume displacement pump. Sequential compression of the boot line propels the blood forward. The amount of compression of the tubing is set in a manner to ensure forward flow without

being over- or underocclusive, otherwise known as the minimally nonocclusive setting. This compression setting is simply referred to as the “occlusion” and is set during priming of the circuit. Proper occlusion is commonly set at a level that allows for the fluid level in the arterial limb to fall at a rate of 1 cm/min when held 75 cm above the venous reservoir fluid level. An overocclusive setting risks damaging the formed elements of the blood as well as increasing the risk of rupturing the boot line. An underocclusive setting risks damaging blood elements by allowing for sloshing in the raceway and compromises effective forward flow. This becomes an important safety issue if flow is not being monitored with an arterial line flow probe. Roller head pumps in modern heart-lung machines have the advantage of very low primes, since it is common for the arterial head to be positioned close to the venous reservoir and oxygenator. These pump heads also operate independently of vascular and system resistance, thereby allowing precise low-flow management, which is especially important in pediatric perfusion. To prevent circuit disruptions with load-independent roller head pumps, it is standard practice to incorporate servoregulating pressure monitoring to prevent excess system pressures. Circuit system pressure normally ranges between 100 and 250 mm Hg, but with kinked tubing or a misplaced line clamp can instantly spike to >500 mm Hg. Servoregulation adjusts arterial pump flow automatically to prevent pressures which could cause circuit connections to fail with subsequent loss of cardiopulmonary support.

The centrifugal head pump, shown in Figure 28.2B, operates on the principle of a constrained vortex. Impellers or vanes on a spinning disc within plastic housing create negative pressure at the pump’s central axis inlet and positive displacement at the pump’s normally singular outlet, positioned at its outer circumference. As with roller head systems, a centrifugal head system is positioned in the bypass circuit to aspirate blood from the venous reservoir and to then pump it into the oxygenator. Centrifugal heads have the theoretical advantage of creating less trauma to the formed elements of the blood (2,3). However, a best-evidence review in patients undergoing elective coronary artery bypass grafting failed to find a significant difference in this regard as compared to roller head pumps (4). Centrifugal heads are afterload-sensitive. An afterload-sensitive system greatly reduces the chances of overpressurizing a pump circuit that can lead to component and system failure, otherwise referred to as “blowing up a circuit” (5). This benefit comes at the expense of less-constant pump flow when system resistance varies (changing systemic vascular resistance, pressure on the arterial limb of the circuit by the surgical team, cannula position change, etc.). Centrifugal heads also reduce the risk of massive arterial air embolism in the case of an emptied venous reservoir, because forward flow is stopped if large amounts of air enter the constrained vortex. Disadvantages of centrifugal systems relative to roller systems for pediatric perfusion include cost and increased prime volume. As with roller head pumps, it is standard practice to incorporate servoregulating arterial circuit pressure monitoring. Additionally, centrifugal systems commonly incorporate a one-way valve in the arterial limb of the bypass circuit to prevent retrograde flow back through the device which could occur during low-flow and no-flow conditions. Automated pump arterial line clamps which activate to prevent retrograde flow may also be used.

There are two distinct types of oxygenators: bubble oxygenators and membrane oxygenators. Bubble oxygenators will not be discussed here, because essentially only membrane oxygenators are currently used in the care of pediatric cardiac surgical patients (6,7).

Membrane oxygenators can be classified as either microporous or true membrane. The gas exchange membrane for pediatric CPB is nearly universally of the microporous type. Microporous membranes are made of 200 to 300 µm diameter polypropylene tubes with hydrophobic pores 0.03 to 0.07 µm in diameter. These pores cover at least 50% of the membrane surface. Blood flow is normally on the outside of the tubes, while countercurrent gas flow is through the inside of the tubes. This is referred to as extraluminal blood flow and is preferred over intraluminal flow, which has a higher risk of membrane clots and shunts which could quickly decrease device performance. Upon exposure to blood, the micropores are designed to become covered with a thin proteinaceous layer through which gas exchange occurs without blood and gas coming in direct contact.

The other type of membrane oxygenator is referred to as a true membrane or a solid membrane type because an intact membrane separates the gas and blood. Solid membranes are constructed of methyl silicon rubber that is thin enough (<25 µm) to permit diffusion of gas, with CO₂ diffusing across the membrane five times as easily as O₂. These membranes are more limiting to diffusion than microporous systems, and therefore require a greater gas exchange surface area. Hence, the priming volume to provide comparable gas exchange is larger. They also do not share the important advantage of microair removal from the blood flow path that microporous-type systems have. A major advantage of true membranes is that they have a useful life measured in days. For this reason, true membranes are still used in the setting of extracorporeal membrane oxygenation (ECMO), although their use in this setting is decreasing as advances in microporous systems have increased their useful life.

While modern membranes themselves offer little impedance to gas diffusion, there are other characteristics of membrane oxygenators that limit gas exchange. The primary determinants of O₂ and CO₂ exchange across a membrane are the solubility and diffusability of each in blood and the partial pressure gradient across the membrane. Because O₂ is less soluble and diffusible in blood than CO₂ (ratio of 1:25), the thickness of the blood film has a greater influence on oxygen exchange. Approximately 90% of the resistance to diffusion in a membrane oxygenator system is diffusion of O₂ into blood (8). The pressure gradient for O₂ can be increased by increasing the O₂ concentration in the fresh gas flow. However, because of the lower diffusability of O₂, a thick blood film or boundary layer will result in poor O₂ delivery to the erythrocytes most distant from the gas exchange pores, even when the pressure gradient for O₂ is high. Modern membranes have design features to induce eddy currents, which serve to limit the boundary layer and maximize red blood cell exposure to the micropores. As CO₂ is more soluble and diffusible than O₂, CO₂ exchange is mainly dependent on the pressure gradient for CO₂ across the membrane. The rate of CO₂ removal can be increased by increasing the oxygenator’s sweep flow rate of ventilating gas, which is analogous to increasing alveolar ventilation. The rate of CO₂ removal can also be enhanced by decreasing the CO₂ content of the sweep gas. Most adult oxygenator systems do not contain CO₂ in the sweep gas but pediatric programs utilizing pH-stat blood gas management commonly do, and therefore CO₂ content of the ventilating gas becomes an important consideration.

The properties of O₂ transfer into and CO₂ removal from the blood are defining characteristics of the different oxygenators. The manufacturing techniques for oxygenator bundle wrapping, the overall oxygenator membrane surface area, and the blood flow path through the device will primarily determine the gas transfer capabilities and maximum recommended blood flow rate which can be handled by the device. To note, over the course of several hours of use, the gas exchange capacity of microporous membranes deteriorates as a protein layer coats the micropores. The designed useful life for gas exchange in microporous oxygenators, as stated by their manufacturers, is normally 6 hours. Pediatric perfusion systems must appropriately size the oxygenator to meet the anticipated individual patient needs for gas exchange and blood flow.

Microporous systems have been proven to be immensely effective. Because their efficient gas exchange properties and microair removal are difficult to match, they have lead the marketplace for cardiac surgery for more than two decades. A disadvantage of microporous membranes is that air can be drawn across the micropores into the blood flow path under unique negative pressure situations. Negative pressure in the oxygenator can occur at any time but practically speaking this can occur during or after bypass. On bypass, if there is an abrupt stoppage of blood flow, the fluid momentum in the oxygenator has the potential to pull ventilating gas into the blood flow path. Excessive suction while drawing blood samples, especially with low flow rates or regional perfusion, can draw ventilating gas into the blood flow path. After bypass, air can be drawn into the blood flow path during traditional
arteriovenous modified ultrafiltration (MUF). Postoxygenator ALFs are therefore extremely important safety devices.

The venous reservoir has blood flow into it directly from the bypass circuit venous line as well as from the cardiotomy filter. Venous return from the patient has its own dedicated filtration system to process the relatively clean venous blood. Venous reservoirs may be open or sealed. Open reservoirs are open to the atmosphere so that air is automatically purged. Sealed reservoirs have a single venting port which may be easily sealed for use with vacuum-assisted venous drainage (VAVD). VAVD requires the use of a sealed reservoir.

Cardiotomy suction serves as an important source of blood conservation. Most systems use a roller pump to provide cardiotomy suction, with the collected blood directed to a cardiotomy filter that then drains into the venous reservoir. As the cardiotomy suction is frequently used during cannula placement before initiation of CPB, it is essential that adequate heparinization be achieved before use of cardiotomy suction so that clot does not form in the cardiotomy filter or venous reservoir.

The cardiotomy filter is normally located higher than and behind the venous reservoir. This superior arrangement allows the cardiotomy filter, which may have more dynamic volume hold up during filtering, the time necessary to work before blood drains into the circulating volume of the venous reservoir. Placing the cardiotomy filter posteriorly in the venous reservoir allows the perfusionist to visualize and focus on the forward facing venous reservoir level to constantly ensure a safe operating level.

The cardiotomy filter has a higher filtration capacity than the venous return filter to handle blood from the field suction devices and ventricular vent line. Cardiotomy blood always has air mixed in and may also contain particulate contaminants which necessitate increased filtering capacity. While cardiotomy suction has its disadvantages, most clinicians consider it obligate, especially in pediatric cardiac surgery, because of the volume of blood lost to the field with collateral circulations and general bleeding during the course of surgery. If cardiotomy suction was not used in these situations, there would be a regular need for homologous blood transfusion simply to maintain a sufficient volume within the CPB system. And, the loss of plasma via cell saver salvage may also be unacceptable.

Blood from the pericardial well collected by cardiotomy suction during bypass and returned to the venous reservoir is known to activate the extrinsic coagulation pathway (9). This aspirated pericardial well blood contaminated by tissue contact is the most significant activator of the coagulation system
during bypass (10). Pericardial aspirate is rich in tissue factor (TF) and procoagulant cellular (primarily platelet) derived microparticles (10,11,12). It has also been demonstrated that elimination of cardiotomy suction return to the venous reservoir or washing of pericardial cardiotomy blood prior to return reduces inflammatory mediator generation and thrombin, neutrophil, and platelet activation (13). Cardiotomy suction is the major source of blood trauma and hemolysis during CPB due to the unavoidable aspiration of air and blood (10,14).

In addition to cardiotomy suction, bypass circuits utilize a roller head pump to provide ventricular venting. The vent line collects blood from the ventricular vent and normally returns it to the cardiotomy filter.

Tubing sizes for congenital heart surgery range in internal diameter sizes from 1/8 to ½ inches, and will vary within each custom set-up. Prime volumes for different size tubing are shown in Table 28.2.

Institutions generally have accepted maximum values for system pressure and achievable flow rates for the different tubing sizes. Tubing packs are then defined based on anticipated pump flow and line lengths required within a CPB setup. For example, a program may start with 3/16-inch tubing for the major tubing components (arterial and venous limbs, boot line) for patients with an expected maximum bypass flow rate of 600 mL/min. As patient weight and expected flow rates increase, the different tubing components are then upsized. The venous limb size is normally increased first, followed by the boot line, and finally the arterial limb until adult bypass circuit sizes are achieved. An adult circuit commonly has a 3/8-inch arterial limb with ½-inch venous and boot lines. The venous limb tends to be the largest tubing prime component, but can be reduced in size with augmented venous return (kinetic and vacuum assisted drainage) techniques. Table 28.3 provides an example of tubing selection based on its position and maximum recommended pump flow rate.

The left ventricular and other vent lines and the field suction lines are other primary components of a custom tubing pack. They are not normally primed before bypass and so their volumes have less impact on the system. However, their sizing is particularly important for small patients because they can be intermittently primed with blood during bypass. Generally, these lines are primed at various times during bypass, but are free of blood during the weaning phase of bypass. For example, an aortic root vent line may be primed with aortic root blood after aortic cross-clamping and during rewarming to aid in de-airing. This line is then either disconnected or
allowed to deprime to the bypass circuit before separation from bypass.

Treatment of the bypass circuit tubing, cannulas, and oxygenators with heparin and other surface-modifying agents (SMAs) has been undertaken by manufacturers in an effort to increase the circuit’s biocompatibility (18,19,20). In theory, use of SMAs should result in decreased activation of platelets and attenuation of the inflammatory response induced by contact activation (21). However, solid evidence linking use of SMAs to meaningful improvements in important clinical outcomes such as reduced blood loss, reduced incidence of renal dysfunction, or shorter duration of ventilatory support in either children or adults undergoing cardiac surgery with CPB is lacking (21,22). Despite this, SMAs are clearly in widespread use in pediatric cardiac surgery programs; a 2011 review of international pediatric perfusion practice reported that 85% of programs were using CPB circuits that incorporated SMAs (1,7).

Arterial inflow is provided most often with a single arterial cannula placed in the ascending aorta proximal to the innominate artery. Alternative arterial inflow strategies include femoral arterial cannulation for reoperations, double arterial cannulation for interrupted aortic arches, and ductus arteriosus/pulmonary artery cannulation for ascending aorta and aortic arch hypoplasia. The surgeon must be cognizant of the arterial cannulation strategy required and its potential effects on systemic perfusion, operative strategy and blood pressure monitoring sites.

Sizing of arterial cannulas is by their outer diameter. The lumen must be large enough to allow for anticipated pump flow rates while not creating excessively high system pressures. If a cannula is too large, it can obstruct native heart output, particularly in the ascending aortic position as this output is critical during cannulation and the initiation and weaning phases of bypass. An arterial cannula that is too small, in addition to limiting flow, leads to high pressures, increased flow velocities, jetting against the arterial wall, and high shear forces which may damage the formed elements of the blood (23,24). The pressure drop across an arterial cannula is normally limited to 100 mm Hg. Arterial cannulas come in a variety of styles and sizes to meet the varied needs of congenital cardiac patients (Fig. 28.6). The outer diameter for arterial cannulas ranges from 6 to 24F to accommodate neonates to older adults. Arterial cannulas commonly have wire-wound bodies to resist kinking. Additionally, arterial cannulas should not become too rigid during hypothermia since this can put undue stress on the aortic wall.

The venous limb of the bypass circuit connects to the venous cannulas. The majority of adult cardiac procedures are extra-cardiac in nature (coronary artery bypass grafts and aortic valve replacements) and may be performed with only a single right atrial cannula for venous drainage. In contrast, most congenital heart operations are intra-cardiac. Venous drainage is typically provided with bicaval cannulation, which allows for total CPB. Venous cannulas are normally rated for achievable flow based on the pressure drop across their length in the setting of gravity siphon drainage (GSD). A pressure drop less than 35-40 mm Hg is commonly used as a limit. Venous cannulas ideally have an internal lumen that is large enough to handle the expected portion of the bypass flow diverted to it while not having an external diameter that completely occludes the vessel. Venous cannulas commonly have side port holes, in addition to the tip hole, which importantly affect flow performance. Blocking these side ports by inserting too large a cannula can diminish flow performance and desired flow rates (Fig. 28.7). It follows that a maximized internal-to-outer (IO) ratio of cannula diameter is advantageous. For this reason, metal- and hard plastic-tipped venous cannulas are commonly used in pediatric settings. These materials lend
themselves to more favorable IO ratios than standard cannula materials such as polyvinylchloride.

Electrochemical neutrality occurs when there are equal concentrations of hydrogen (H⁺) and hydroxyl (OH^–) ions. Electrochemical neutrality is important for maintenance of a constant transcellular H⁺ gradient necessary for many cellular processes and optimal protein and enzyme function. In humans, electrochemical neutrality (ratio of H⁺:OH^– = 1) occurs at an intracellular pH of 7.1, which is close to the pH of neutrality of water (7.0). Buffer systems maintain the extracellular pH at 7.4 and the intracellular pH at 7.1. The imidazole group of the amino acid histidine is present in many cellular and plasma proteins, including hemoglobin. Together with phosphate (mainly intracellular) and bicarbonate, these buffers maintain an equal concentration of H⁺ and OH^– ions, and thereby electrochemical neutrality.

When water (and blood) is cooled, the dissociation constant (pK) decreases, resulting in an equal reduction in the concentration of H⁺ and OH^– ions. As pH is the inverse logarithm of the H⁺ concentration, and because the H⁺ concentration decreases with cooling, the pH of the water (and blood) will increase. For each 1°C decrease in temperature, the pH of water will increase by 0.017 and the pH of blood will increase by 0.0147 (45).

Changes in cellular pH during hypothermia are mediated through Paco₂ homeostasis. As temperature decreases, the solubility of CO₂ in blood increases. If the total CO₂ content of blood is held constant, this increase in CO₂ solubility will result in a reduction in Pco₂. For example, if the total CO₂ content is held constant and the measured Pco₂ at 37°C is 40 mm Hg, then the measured Pco₂ at 20°C will be 16 mm Hg. This causes pH to increase as temperature decreases and electrochemical neutrality is maintained.

Blood gas analyzers evaluate samples at a temperature of 37°C. If a blood sample is lower than 37°C, the analyzer will warm the sample to 37°C and then measure the various parameters. Therefore, when a blood gas sample is drawn from a hypothermic patient and sent to the blood gas laboratory, the sample is warmed to 37°C before measurement. The values obtained at 37°C are called the temperature-uncorrected values. These values are then converted to temperature-corrected values at the patient’s temperature using a nomogram. The nomogram accounts for temperature-induced changes in pH, O₂ solubility, and CO₂ solubility in a closed-blood system. When pH and Paco₂ are measured at 37°C and then corrected to a lower temperature, the electrochemically neutral pH will be higher and the corrected Paco₂ will be lower than the values at 37°C. Temperature correction of blood gas values is independent of whatever blood gas strategy is being used to manage the patient and is only accounting for the physiochemical changes of temperature on pH and solubility of CO₂ and O₂.

α-Stat and pH-stat are methods of acid-base management and directly influence blood flow to the brain and other organs.

In α-stat regulation, electrochemical neutrality is maintained at whatever temperature the patient is at. This is achieved by allowing the pH to increase as the patient is cooled. The pH will be alkalotic and the Paco₂ decreased in the temperature-corrected blood gas and normal in the temperature-uncorrected blood gas. For example, when a blood gas is taken from a patient at 27°C, the temperature uncorrected (at 37°C) pH and Paco₂ will be 7.4 and 40 mm Hg, respectively, while the temperature corrected (27°C) pH and Paco₂ will be 7.55 and 26 mm Hg, respectively (Table 28.6). With a-stat management, the total CO₂ content of the patient and bypass circuit is held constant.

With pH-stat regulation, electrochemical neutrality is lost in that the pH is maintained at 7.4 and the Paco₂ at 40 mm Hg at whatever colder temperature the patient is at, that is, normal values on the temperature-corrected blood gas. As shown in Table 28.6, at temperature-uncorrected values the pH will be acidotic and the Paco₂ higher than normal. A pH-stat strategy is achieved by adding CO₂ to the ventilating gas during hypothermic CPB to increase Paco₂ and decrease the pH. In contrast to a-stat regulation, in which total CO₂ content is kept constant, pH stat regulation results in an increase in total CO₂ content.

It is important to point out that at a patient temperature above 30°C there is essentially no difference between a-stat and pH-stat management. The difference between these strategies becomes more marked as temperature progressively decreases and is not clinically relevant until patient temperature is below 30°C. For practical purposes, it is easier to use uncorrected blood gases aiming for a pH of 7.4 and a Paco₂ of 40 mm Hg at 37°C during a-stat management and blood gases corrected for patient temperature aiming for a pH of 7.4 and a Paco₂ of 40 mm Hg during pH-stat management.

Landmark studies by Greeley and colleagues found that when a-stat regulation is used, pressure-flow autoregulation is preserved during moderate hypothermic (25°C-32°C) bypass, but is lost or attenuated during deep hypothermic (18°C-22°C) bypass (46,47). Cerebral blood flow (CBF) and cerebral O₂ consumption (CMRo₂) are appropriately coupled with a-stat regulation. In contrast, the addition of CO₂ during pH-stat management produces loss of autoregulation and uncoupling of CBF and CMRo₂ (48). As a result, CBF varies linearly with arterial blood pressure and luxuriant cerebral blood flow exists with CBF far in excess of that dictated by cerebral metabolic rate. This hyper-perfusion state is the result of reduced CMRo₂ induced by hypothermia and cerebral vasodilation resulting from a disproportionately high Paco₂ for the degree of hypothermia present. Cerebral metabolism is exponentially related to temperature, and the temperature coefficient in neonates, infants, and children is 3.65 versus 2.4 in adults (49,50).

The etiology of brain injury during CPB is a major consideration in the decision to use an α-stat or pH-stat strategy. Because of atherosclerotic disease, adults are more susceptible than children to brain injury resulting from macro- and microembolism of atherosclerotic debris. By linking CBF with CMRo₂ and thereby avoiding the cerebral hyperperfusion associated with pH-stat, α-stat regulation is most commonly used for adult cardiac surgery. Global and regional hypoperfusion is a bigger concern in the pediatric population. Additionally, lower temperatures, with shift of the oxyhemoglobin dissociation curve to the left, are more frequently required in children. Consequently, pediatric CPB more frequently uses the pH-stat strategy. Additional benefits of pH-stat include suppression of CMRo₂, which may prolong oxygen availability during periods of circulatory arrest, and more efficient brain cooling (51). The acidosis associated with pH-stat increases the pulmonary vascular resistance and thereby decreases pulmonary blood flow, which has the benefit of reducing the steal of blood from the systemic and hence cerebral circulation in patients with significant aortopulmonary collaterals (52). For an extensive review supporting the use of pH-stat in pediatric cardiac surgery, the reader is referred to reference (44).