A Practical Approach to Cardiac Anesthesia (Practical Approach Series) 5th Ed.

21 Cardiopulmonary Bypass: Equipment, Circuits, and Pathophysiology

Eugene A. Hessel, II, Glenn S. Murphy, Robert C. Groom, and Joseph N. Ghansah


 1. The goal of cardiopulmonary bypass is to provide adequate gas exchange, oxygen delivery, systemic blood flow with adequate perfusion pressure while minimizing the detrimental effects of bypass.

 2. Roller pumps may cause more damage to blood elements and can result in massive air embolism if the venous reservoir becomes empty.

 3. Membrane oxygenators function similarly to natural lungs, imposing a membrane between the ventilating gas and the flowing blood, thereby eliminating direct contact between blood and gas.

 4. During bypass, excessive and rapid warming of blood with the bypass heat exchanger must be avoided to prevent gas coming out of solution risking embolism and to avoid excessive heating of the brain with subsequent potential neurologic damage.

 5. Cardiotomy suction should be minimized or cell salvage techniques used to process the cardiotomy blood as it contains microaggregates of cells, fat, foreign debris, thrombogenic and fibrinolytic elements thought to be major sources of hemolysis, and microemboli during CPB.

 6. More recent data have suggested that the lower limit of autoregulation of the brain is approximately a mean pressure of 70 mm Hg in awake, normotensive subjects. On the basis of this data some clinicians are now using higher (greater than 70 mm Hg) mean pressure on bypass.

 7. Critical oxygen delivery is that point at which maximum oxygen extraction is reached and oxygen consumption starts to fall.


   I. Introduction: It is critical that anesthesiologists who provide care for patients undergoing surgery using cardiopulmonary bypass (CPB) be intimately familiar with the function of the heart–lung (H–L) machine (also referred to as the extracorporeal circuit [ECC]). In this chapter, the components of the H–L machine, the physiologic principles and pathophysiologic consequences of CPB, and the important role the anesthesiologist should play in its optimal and safe conduct are described. In Chapter 8, the medical management of patients during CPB is described.


  The primary goal and function of CPB is to divert blood away from the heart and lungs and return it to the systemic arterial system, thereby permitting surgery on the nonfunctioning heart. In doing so, it must replace the function of both the heart and the lungs. The goal is to provide adequate gas exchange, oxygen delivery, systemic blood flow, and arterial pressure, while minimizing the adverse effect of extracorporeal circulation.This is accomplished by the two principal components of the H–L machine: The artificial lung (blood gas exchanging device or “oxygenator”) and the arterial pump. The “oxygenator” removes carbon dioxide as well as adds oxygen to provide the desired PaO2 and PaCO2, while the arterial pump supplies the energy to maintain systemic blood for arterial pressure and organ perfusion. Because the proximal ascending aorta is often cross-clamped to arrest the heart to facilitate surgery, a cardioplegia delivery system is added to minimize myocardial ischemia. Other components of the ECC include cannulae that connect to the systemic venous and arterial systems, a venous reservoir, a heat exchanger to control body temperature, field or cardiotomy suction, and various safety and monitoring devices. These components will be described in this chapter. The interested reader may find further details on CPB components in the referenced texts [15].

II. Components of the circuit

   A. Overview: The essential components of the H–L machine include the CBP console, oxygenator, venous reservoir, arterial pump, cardioplegia circuit, ventilating circuit, monitoring and safety systems, and various filters. These components can be assembled in a myriad of configurations depending on perfusionist/surgeon preference and patient need. Figure 21.1 shows a detailed schematic of a typical CPB circuit. Desaturated blood exits the patient’s vena cava through a right atrium (RA)/inferior vena cava (IVC) venous cannula and is diverted to the venous reservoir by gravity siphon drainage through large-bore polyvinyl chloride (PVC) tubing. Blood is then drawn from the venous reservoir by the systemic blood pump, which can be either roller or kinetic, and pumped through a heat exchanger (integral to the membrane oxygenator [MO]). Blood then passes through the oxygenator, through an arterial filter, and back into the patient through the arterial cannula inserted into the ascending aorta. Additional parts of the circuit include a recirculation line from the arterial side of the oxygenator, which is used for priming the system and as a blood source for cardioplegia. A purge line is located on the housing of the arterial filter and is kept open during CPB to vent any air from the circuit back to the venous reservoir, or a second reservoir called cardiotomy reservoir.

   Other roller pumps on the H–L machine are used for various functions including delivery of cardioplegic solution, return of shed blood via aspiration, venting of blood from intracardiac sources, or the removal of air from the venous reservoir when collapsible venous reservoir “bag” systems are used. Additional components of the CPB system include microprocessors for console control and electronic data recording, a cooler/heater which serves as an adjustable temperature water source used in conjunction with the circuit heat exchangers, an anesthetic vaporizer for the administration of volatile anesthetic agents, cardioplegia delivery system, various sensors for monitoring arterial and venous blood parameters as well as oxygen concentrations in the ventilating circuit, and various safety devices. Most of the components through which the blood passes are disposable and commercially custom-prepared to meet the specific requirements of individual cardiac teams.

Figure 21.1 Detailed schematic diagram of arrangement of a typical CPB circuit using an MO with integral hard-shell venous reservoir (lower center) and external cardiotomy reservoir. Venous cannulation is by a cavoatrial cannula and arterial cannulation is in the ascending aorta. Some circuits do not incorporate a membrane recirculation line; in these cases the cardioplegia blood source is a separate outlet connector built-in to the oxygenator near the arterial outlet. The systemic blood pump may be either a roller or centrifugal type. The cardioplegia delivery system (right) is a one-pass combination blood/crystalloid type. The cooler–heater water source may be operated to supply water to both the oxygenator heat exchanger and cardioplegia delivery system. The air bubble detector sensor may be placed on the line between the venous reservoir and systemic pump, between the pump and MO inlet or between the oxygenator outlet and arterial filter (neither shown) or on the line after the arterial filter (optional position on drawing). One-way valves prevent retrograde flow (some circuits with a centrifugal pump also incorporate a one-way valve after the pump and within the systemic flow line). Other safety devices include an oxygen analyzer placed between the anesthetic vaporizer (if used) and the oxygenator gas inlet and a reservoir-level sensor attached to the housing of the hard-shell venous reservoir (on the left). Arrows, directions of flow; X, placement of tubing clamps; P and T, pressure and temperature sensors, respectively. Hemoconcentrator (described in text) not shown. (From Hessel EA II. Circuitry and cannulation techniques. In: Gravlee GP, et al., eds. Cardiopulmonary Bypass. Philadelphia, PA: Lippincott Williams & Wilkins; 2008:64, with permission.)

   B. Venous cannulation and drainage

     1. Overview: Blood must be diverted into the H–L machine to keep it from passing through the heart and lungs and thereby permit access to the heart by the surgeon.

     2. Central venous cannulation (Fig. 21.2)

Figure 21.2 Venous cannulation (central, intra-thoracic). Methods of venous cannulation. A: Single cannulation of RA with a “two-stage” cavoatrial cannula. This is typically inserted through the RA appendage. Note that the narrower tip of the cannula is in the IVC, where it drains this vein. The wider portion, with additional drainage holes, resides in the RA, where blood is received from the coronary sinus and SVC. The SVC must drain via the RA when a cavoatrial cannula is used. B: Separate cannulation of the SVC and IVC. Note that there are loops placed around the cavae and venous cannulae and passed through tubing to act as tourniquets or snares. The tourniquet on the SVC has been tightened to divert all SVC flow into the SVC cannula and prevent communication with the RA. (From Hessel EA II. Circuitry and cannulation techniques. In: Gravlee GP, et al., eds. Cardiopulmonary Bypass.Philadelphia, PA: Lippincott Williams & Wilkins; 2008:67, with permission.)

        a. Single simple cannula in RA: Inserted through a purse-string suture in the RA free wall or atrial appendage. This type of cannulation tends to be unstable, does not reliably divert flow into the right ventricle (RV), and is rarely used in adult CPB.

        b. Cavoatrial or “two-stage” single cannula: A single-lumen cannula with a wide proximal portion with drainage slits situated in the RA, and a narrower distal end placed into the IVC. The tip in the IVC makes this cannula more stable. It is usually inserted through a purse-string suture in the atrial appendage. Insertion may be difficult in the presence of RA masses or hardware or a prominent Eustachian valve or Chiari network. Tears produced at the junction of the IVC with the RA are difficult to manage. This is the most common type of cannulation for coronary artery and aortic valve surgery. It may not reliably prevent blood from entering the RV, and may not provide optimal myocardial cooling (especially the RA and RV). Superior vena cava (SVC) drainage and hence venous return to the H–L machine can be compromised if the junction of the SVC to the RA is kinked (which occurs when the heart is lifted up for grafting of vessels in the inferior and posterior-lateral wall).

        c. Bicaval cannulation: Separate cannulae are placed into the SVC and IVC either directly or indirectly through the RA through purse-string sutures. Bicaval cannulation is most effective at totally diverting blood away from the heart. When the right heart must be opened, additional large ligatures or tapes are placed around the SVC and the IVC to prevent any caval blood from entering the atrium (and any air getting from the atrium into the venous drainage). When the caval tapes are tightened, this is termed “complete bypass.” It is critical that when these tapes are tightened that the cannulae or drainage do not become obstructed or venous hypertension (congestion) will occur. This is of particular concern in the SVC because of the potential adverse impact on cerebral perfusion; pressure in the SVC cephalad to the tip of the SVC cannula should be monitored. Bicaval cannulation is necessary whenever the surgeon plans to open the right heart (for surgery involving the tricuspid valve, RA masses, trans-atrial septal approaches to left heart, and for congenital heart surgery) and for mitral valve (MV) surgery. The latter is because retraction of the RA to view the MV causes kinking of the junction of the cavae with the RA which interferes with venous drainage, or if the surgeon approaches the MV through the RA, or for Tricuspid Valve (TV) repair. Bicaval cannulation, by diverting blood away from the right heart, minimizes cardiac warming, especially of the RA and ventricle during cold cardioplegia.

        d. Impact of persistent left superior vena cava (LSVC) on venous cannulation: About 0.3% of the general population (1% to 10% of patients with congenital heart disease) will have a persistent LSVC. This is a remnant of fetal development and drains blood from the junction of the left internal jugular (IJ) and left subclavian veins into the coronary sinus, and hence into the RA. Therefore, placing cannulae in the right SVC and IVC will not be effective in diverting all of the venous drainage away from the RA. One simple solution is for the surgeon to temporarily occlude (snare) the LSVC. However, in about two-thirds of these patients, the left innominate vein is absent or small, and this maneuver may result in venous hypertension and adverse cerebral consequences. In these cases, the surgeon may place a third venous cannula in the LSVC either retrograde via the coronary sinus, or directly into the LSVC through a purse-string suture.

     3. Peripheral venous cannulation: Used for minimally invasive/“port-access” approaches, surgery via left thoracotomy, or for cannulation before entering the chest (electively or emergently when bleeding is anticipated or encountered). Most commonly, venous cannulae are placed via the femoral vein (and rarely the IJ vein). If the femoral vein is used, the cannula is positioned with the tip at the SVC–RA junction. Positioning of the cannula is often guided by transesophageal echocardiography (TEE). Bicaval cannulation is possible with a special IVC cannula designed for this purpose. If a separate IJ venous cannula is placed, it is often inserted by the anesthesiologists shortly after induction and requires special attention on their part in regards to sterile technique, heparinization, and clamping and unclamping of this line for CPB. As peripheral venous cannulae are smaller and longer than directly placed cannulae, resistance to drainage is greater and may require use of augmented venous drainage (see later).

     4. Venous cannulae are plastic (Fig. 21.3). Some are wire-reinforced to minimize kinking. Others designed for direct caval cannulation have curved thin metal or plastic tips for a favorable internal diameter to external diameter (ID:OD) ratio.

Figure 21.3 Venous cannulae. Drawings of commonly used venous cannulae. A: Tapered, “two-stage” RA–IVC cannula. B: Straight, wire-wound “lighthouse” tipped cannula for RA or separate cannulation of the SVC or IVC. C: Right-angled, metal-tipped cannula for cannulation of the SVC or IVC. (From Hessel EA II. Circuitry and cannulation techniques. In: Gravlee GP, et al., eds. Cardiopulmonary Bypass. Philadelphia, PA: Lippincott Williams & Wilkins; 2008:65, with permission.)

     5. Types of drainage

        a. Gravity: Venous drainage is usually accomplished by gravity (siphon effect). This requires that the venous drainage tubing is full of fluid (blood). Drainage is based upon the pressure difference of the column of fluid between the level of the patient and that of the H–L machine (venous reservoir). Flow is influenced by central venous pressure (intravascular volume and venous tone), height differential between the patient and the H–L machine, and resistance in the venous cannula and tubing (length, internal diameter, mechanical obstruction, or malposition of cannulae). “Chattering” of the venous lines suggest excessive drainage or inadequate venous return. Since drainage depends on a siphon, this will be interrupted if the venous line becomes filled with air.

        b. Augmented venous drainage: Used when long or smaller cannulae or venous lines are employed and to permit elevation of the H–L machine to the level of the patients (all designed to decrease the prime volume or for peripheral or port-access cannulation). Two classes of systems are utilized: Vacuum-assisted and kinetic.

           (1) Vacuum-assisted drainage is accomplished by attaching the venous line to a closed “hard-shell” venous reservoir (see below) to which vacuum (usually negative 20 to 50 mm Hg) is applied. Whenever augmented venous drainage is employed there is increased risk of aspirating air from around the venous cannulae, and application of a second purse-string around this site is recommended. There is also additional risk of developing positive pressure in the closed reservoir which can lead to retrograde venous air embolism. This requires inclusion of a positive pressure release valve and heightened attention on the part of the perfusionist.

           (2) Kinetic-assisted drainage is accomplished by inserting a pump (usually centrifugal, but rarely a roller pump). The use of the former is easier to control and minimizes collapse of the cavae or atrium around the tip of the cannulae. This requires close attention by the perfusionist, and as with vacuum-assisted venous drainage increases the risk of air aspiration.

           (3) Studies have not found use of augmented venous return to increase destruction of blood elements nor to aggravate the inflammatory response to CPB.

   C. Arterial cannulation

     1. Overview: The blood from the H–L machine must be returned to the systemic arterial system through an arterial cannula. These cannulae are the narrowest part of the circuit and must carry the entire systemic blood flow (“cardiac output”). The size of cannula is based on the desired blood flow (mainly influenced by patient size) and is chosen to keep blood velocity less than 100 to 200 cm/s and pressure gradients less than 100 mm Hg. Higher flows and pressures (jets) may result in trauma to blood elements and the vessel wall (“sandblasting” and dissection) and potential for reduced flow into side branches. To maximize the ID/OD ratio, the tips of the cannulae are often constructed from metal or hard plastic and the narrowest part of the arterial line is kept as short as possible. Some special tips have been designed to minimize the exit velocities and jet effects (Fig. 21.4). Most commonly the arterial cannula is inserted in the distal ascending aorta, but other sites are also used.

Figure 21.4 Arterial cannulae. Drawings of commonly used arterial cannulae. A: Tapered, bevel-tipped cannula with molded flange near tip. B: Angled, thin-walled, metal-tipped cannula with molded flange for securing cannula to aorta. C: Angled, diffusion-tipped cannula designed to direct systemic flow in four directions (right) to avoid a “jetting effect” that may occur with conventional single-lumen arterial cannulae. D: Integral cannula/tubing connector and luer port (for deairing) incorporated onto some newer arterial cannulae. (From Hessel EA II. Circuitry and cannulation techniques. In: Gravlee GP, et al., eds. Cardiopulmonary Bypass. Philadelphia, PA: Lippincott Williams & Wilkins; 2008:72, with permission.)

     2. Cannulation site options (Table 21.1)

Table 21.1 Arterial cannulation sites

        a. Ascending aorta: This is the most common approach. The cannula is inserted through one or two concentric purse-sting sutures in the distal ascending aorta and directed toward the transverse arch (NOT toward one of the arch vessels). Dislodgement of atheromatous material from the cannulation site is a primary concern. Palpation of the aorta may not be sensitive enough to detect atheroma, and many advocate imaging of the intended site for cannulation with epiaortic ultrasound. When placed in the ascending aorta, some groups use long cannulae directed into the proximal descending aorta to minimize jet effects in the arch, while others use very short cannulae inserted only 1 to 2 cm into the aorta. Dissection associated with ascending aortic cannulation occurs in less than 0.1% of patients. The ascending aorta may not be a suitable cannulation site for various reasons including presence of severe atherosclerotic disease, aortic dissection, and for minimal access surgery, left thoracotomy surgery, and risk of or rescue from hemorrhage during repeat sternotomy.

        b. Femoral or external iliac artery: This is the second most common approach and is used when ascending aortic cannulation is not desirable or feasible. However, this approach has a number of limitations including risk of dissection (0.5% to 1%), risk of atheroembolism (especially into brain and heart), malperfusion of the brain and other organs in the presence of extensive aortic dissection, and ischemia of the cannulated limb. The use of TEE surveillance of the descending aorta is recommended when CPB is initiated and periodically throughout CPB in order to detect the presence of a retrograde dissection. Prolonged femoral cannulation times may result in the release of emboli and acidotic products from the limb with reperfusion and for subsequent development of compartment syndrome in the limb. To minimize leg ischemia, some groups sew a graft onto the side of the femoral artery and insert the arterial cannula into this graft so that blood flows both retrograde and antegrade, while others insert a supplemental arterial cannula into the distal femoral artery.

        c. Axillary/subclavian artery cannulation is often advocated in the presence of aortic dissection or severe atherosclerosis. These vessels are usually free of significant atherosclerosis, and malperfusion and iatrogenic dissection are probably less common than with femoral cannulation. The artery is approached through an infraclavicular incision, and the cannula can be placed either directly into the vessel, or via a graft sewn onto the side of the artery. The right artery is favored since it permits selective cerebral perfusion (SCP) if circulatory arrest is required (see later). If the vessel is cannulated directly (i.e., not through a side arm graft), then the artery in the contralateral upper extremity (radial or brachial) must be used for systemic arterial pressure monitoring during CPB.

        d. Innominate artery: Rarely used approach. There is concern about adequacy of flow around the cannula (which is directed retrograde toward the ascending aorta) into the distal vessel and hence the brain.

   D. Venous reservoir

     1. Overview: The venous reservoir is designed to receive the venous drainage from the patient. The reservoir is placed immediately before the systemic arterial pump to serve as a “holding tank” and act as a buffer for fluctuation and imbalances between venous return and arterial flow. It also serves as a high-capacitance (i.e., low-pressure) receiving chamber for venous return and hence facilitates gravity drainage of venous blood. Additional venous blood may become available from the patient when CPB is initiated and systemic venous pressure is reduced to low levels. Thus, as much as 1 to 3 L of blood may need to be translocated from the patient to the ECC when full CPB begins. This reservoir may also serve as a gross bubble trap for air that enters the venous line, as the site where blood, fluids, or drugs may be added, and as a ready source of blood for transfusion into the patient. One of its most important functions, however, is to provide a source of blood if venous drainage is sharply reduced or stopped; this provides the perfusionist with reaction time in order to avoid “pumping the CPB system dry” and risking massive air embolism. These reservoirs usually include various filtering devices. There are two classes of reservoirs:

        a. Rigid hard-shell plastic, “open” venous canisters: Advantages include ability to handle venous air more effectively, simple to prime, larger capacity, and ability to apply suction for vacuum-assisted venous return. Most hard-shell venous reservoirs incorporate macro- and microfilters often coated with defoaming agents and can also serve as the cardiotomy reservoir (see later) by directly receiving suctioned and vented blood. Their ability to remove gaseous microemboli (GME) varies.

        b. Soft-shell, collapsible plastic bag, “closed” venous reservoirs: These reservoirs eliminate the gas–blood interface and reduce the risk of massive air embolism because they will collapse when emptied and do not permit air to enter the systemic pump. Closed collapsible reservoirs also make the aspiration of air by the venous cannulae more obvious to the perfusionist, but require a way of emptying the air out of the reservoir. When soft-shell reservoirs are used, a separate cardiotomy reservoir is required (see later). Because of reduction of the gas–blood interface, their use may be associated with less inflammatory activation. Data on comparative clinical outcome with use of the two types of venous reservoirs are conflicting and inconclusive [6].

   E. Systemic (arterial) pump: There are currently two types of blood pumps used in the CPB circuit: Roller and kinetic (most commonly called centrifugal) (Table 21.2). In the United States, kinetic pumps are used in approximately 50% of all procedures.

Table 21.2 Comparison of roller versus centrifugal pumps

     1. Roller pump (Fig. 21.5)

Figure 21.5 Roller pump. Drawing of a dual roller pump and tubing. The principle of the roller pump is demonstrated by the hand roller in the lower drawing moving along a section of tubing pushing fluid ahead of it and suctioning fluid behind it. The upper four drawings in sequence (A–D) show how roller pump B first moves fluid ahead of it and suctions fluid behind it (A). As the pump rotates clockwise, the second roller A begins to engage the tubing (B). As the rotation continues there is a very brief period with volume trapped between the two rollers (C) and no forward flow, which imparts some pulsatility. In position D, roller B leaves the tubing, while the second roller A continues to move fluid in the same direction. Not shown are the roller pump backing plate, tubing holders, and tube guides for maintaining the tubing within the raceway. Fluid flows in direction of the arrows. (From Stofer RC. A Technic for Extracorporeal Circulation. Springfield, IL: Charles C. Thomas; 1968:22, with permission.)

        a. Principles of operation: Blood is moved through this pump by sequential compression of tubing by a roller against a horseshoe-shaped backing plate or raceway. A typical pump has two roller heads configured 180° apart to maintain continuous roller head contact with the tubing. The output is determined by the stroke volume of each revolution (the volume within the tubing which is dependent upon the tubing size [internal diameter] and the length of the compressed pathway times the revolutions per minute [rpm]). Flow from a systemic roller pump increases or decreases linearly with rpm. With larger ID tubing (e.g., 1/2-inch ID), lower rpm are required to achieve the same output compared to smaller ID tubing. The total pump output is displayed in milliliters or liters per minute on the pump control panel. Roller pumps are also used to deliver cardioplegia solution, remove blood and air from heart chambers or great vessels, and suction shed blood from the operative field (see later).

        b. Adjustment of occlusion: To minimize hemolysis, the occlusion must be properly set. Occlusion describes the degree to which the tubing is compressed between the rollers and the backing plate. An under-occlusive pump will allow retrograde movement of fluid when the pressure in the downstream location exceeds that generated by the pump, reducing forward flow. Conversely, an over-occlusive pump will create cellular damage (red blood cell [RBC] hemolysis, white blood cell [WBC] and platelet activation) and excessive wear on the tubing with release of microparticles (“spallation”). Occlusion is set by the perfusionist by adjusting the distance between the raceway and each of the roller heads. Typically, occlusion is adjusted to be barely nonocclusive.


        c. Advantages and disadvantages: Roller pumps have the advantages of being simple, effective, low-cost, having a low priming volume, and producing a reliable output which is afterload independent. A primary disadvantage is that since output is afterload independent, if the arterial line becomes occluded, high pressure will develop which may cause rupture of connections in the arterial line. If inflow is obstructed, roller pumps can generate high negative pressures creating microbubbles (“cavitation”) and RBC damage. Roller pumps may cause more damage to blood components and can result in massive air embolism if the venous reservoir becomes empty. They do not adjust to changes in venous return and require more careful attention by the perfusionist.

     2. Centrifugal pumps (Fig. 21.6)

Figure 21.6 Centrifugal pumps. Drawings of centrifugal pump-heads. A cross-sectional view of a smooth, cone-type pump is shown on the top. Blood enters at A and is expelled on the right (B) due to kinetic forces created by the three rapidly spinning cones. Impeller-type pumps with vanes are shown in the bottom drawings. (Modified from Trocchio CR, Sketel JO. Mechanical pumps for extracorporeal circulation. In: Mora CT, ed. Cardiopulmonary Bypass: Principles and Techniques of Extracorporeal Circulation. New York, NY: Springer-Verlag; 1995:222, 223, with permission.)

        a. Principles of operation: Centrifugal pumps consist of a nest of smooth plastic cones or a vaned impeller located inside a plastic housing. When rotated rapidly (2,000 to 3,000 rpm), these pumps generate a pressure differential that causes the movement of fluid. Smaller, vaned, impeller-type rotary (centrifugal) pumps are being used clinically in place of the traditional cone-type centrifugal pump These have smaller prime volumes and may cause less hemolysis.

        b. Advantages and disadvantages: Unlike roller pumps, these devices are totally nonocclusive and afterload dependent (an increase in downstream resistance or pressure decreases forward flow). This has both favorable and unfavorable consequences. Flow is not determined by rotational rate alone, and therefore a flowmeter must be incorporated in the outflow line to quantify pump flow. Furthermore, when the pump is connected to the patient’s arterial system but is not rotating, blood will flow backward through the pump and out of the patient unless the CPB systemic line is clamped or a one-way valve is incorporated into the arterial line. This can cause exsanguination of the patient or aspiration of air into the arterial line (from around the purse-string sutures). On the other hand, if the arterial line becomes occluded, these pumps will not generate excessive pressure and will not rupture the systemic flow line. Likewise, they will not generate as much negative pressure and hence as much cavitation and microembolus production as a roller pump if inflow becomes occluded.

            A reputed advantage of centrifugal pumps over roller pumps is less risk of pumping massive air emboli into the arterial line; centrifugal pumps will become deprimed and stop pumping if more than approximately 50 mL of air is introduced into the circuit. However, they will pass smaller but still potentially lethal quantities of smaller bubbles. A number of studies have demonstrated that centrifugal pumps cause less trauma to blood elements, less activation of coagulation, produce fewer microemboli, and may be associated with better clinical outcomes than roller pumps [6].

     3. Pulsatile flow and pulsatile pumps

        a. Overview: Most roller pumps produce only a low-amplitude, high-frequency pulsatile flow of little physiologic relevance, while centrifugal pumps produce a nonpulsatile flow. The importance of pulsatile flow has long been debated. (See later in this chapter and paper by Murphy et al. [6].)

        b. Many groups use roller pumps and centrifugal pumps that can be programmed to produce pulsatile flow.

        c. A major problem with efforts to produce physiologically effective pulsatile flow in the patient is the dampening effect of various components distal to the arterial pump including the MO, arterial filter, and the arterial cannula. It has been calculated that very little of the pulsatile energy generated is actually delivered into the patient’s arterial system.

   F. The oxygenator (artificial lung or gas exchanging device)

      1. Although numerous types of oxygenators have been used in the past, currently only MOs are used in most parts of the world. These produce less blood trauma and microemboli, permit more precise control of arterial blood gases, and improve patient outcomes as compared with bubble oxygenators. Virtually all current MOs are positioned after the arterial pump because the resistance in the blood path requires blood to be pumped through them, and to minimize the risk of pulling air through the membrane and producing GME. Most oxygenators also include an integral heat exchanger (see later).


          MOs function similarly to natural lungs, imposing a membrane between the ventilating gas and the flowing blood and eliminating direct contact between the blood and the gas. At least three types of membranes are used:

        a. True membranes: These usually consist of thin sheets of silicone rubber wrapped circumferentially over a spool.

        b. Microporous polypropylene (PPL) membranes: Usually configured in longitudinal bundles of narrow hollow fibers, but occasionally as folded sheets of membrane. The pores fill with autologous plasma which serves as the “membrane” through which gas exchange occurs. With excessive pressure in the blood path or over prolonged time, plasma may leak through the membrane (which degrades gas transfer), while excessive negative pressure can lead to entrainment of air emboli. In hollow fiber MOs, the blood typically flows outside the hollow fibers, while ventilating gases flow through the hollow fibers (in a counter-current direction).

        c. Poly-methyl pentene (PMP) diffusion membranes: Hollow fiber MOs made of a new non porous plastic, PMP, are true membranes. This has the advantage of minimizing the risk of plasma leak and microair aspiration and permits prolonged oxygenation(days). Gas exchange occurs by diffusion across this true membrane. An important limitation is that it does not appear to allow transfer of volatile anesthetics and therefore intravenous anesthetics must be employed during CPB. Because of this limitation and because they are more expensive, PMP “diffusion” MOs are not commonly used for conventional CPB, at least in the United States, but because of their reduced risk of plasma leak (“oxygenator pulmonary edema”) are commonly used for long term extracorporeal support (e.g., extracorporeal membrane oxygenation [ECMO]).

      2. MOs were thought to serve as bubble filters and to prevent venous GME from passing into the arterial system, but it is now recognized that the majority of venous gas emboli transit through MOs. The effectiveness of GME removal varies among MOs [6]. This limitation is why teams must make every effort to minimize the entrainment and administration of air into the venous drainage system.

     3. Control of gas exchange and gas supply to the MO: Gas exchange by MO is controlled similarly to normal lungs. Arterial carbon dioxide levels are controlled by flow of fresh gas (commonly called “sweep gas flow”) through the oxygenator (comparable to alveolar ventilation), and arterial PO2 is controlled by varying fractional inspired oxygen (FIO2). Oxygenators require a gas supply system. This typically includes a source of oxygen and air (and occasionally carbon dioxide), which passes through a blender. An oxygen analyzer should be incorporated in the gas supply line after the blender. An anesthetic vaporizer is also placed in the gas supply line near the oxygenator. Volatile anesthetic liquids may be destructive to the plastic components of ECCs; therefore, care must be taken when filling them with volatile agents. A method of scavenging waste gas from the oxygenator outlet should be provided.

   G. Heat exchanger

     1. Overview: The passage of blood through the ECC results in heat loss and patient cooling. To maintain normothermia, heat must be added to the circuit. This is accomplished with a heat exchanger, which may also be used to intentionally cool and rewarm the patient. Heat exchangers consist of heat-exchanging tubes (often metal) through which the blood flows. These tubes are surrounded by water of varying temperatures. As mentioned earlier, heat exchangers are often incorporated in the oxygenator.

     2. Heater–cooler. To control the temperature of the water flowing through the heat exchanger, a heater–cooler device adjusts the water temperature and pumps it through the heat exchanger (counter-current with the blood flow).


      3. Excessive gradients between the blood and water temperature should be avoided. Most groups avoid gradients greater than 10°C. Proper conduct of heating and cooling requires monitoring of the temperature of the water going to the heat exchanger, and of the venous and arterial blood entering and leaving the H–L machine. Excessive warming can lead to gases coming out of solution and causing GME and could cause excessive heating of the brain. Currently, most groups limit the inflow temperature of the arterial blood to 37°C and limit the inflow temperature of the water entering the heat exhanger to 40°C. However, acceptable or optimal temperature gradients have not been conclusively determined.

      4. Separate heat exchangers are also used in the cardioplegia circuits (see below).

   H. Cardioplegia delivery system or circuit

     1. Overview: When the aorta is cross-clamped (distal to the aortic valve but proximal to the arterial inflow cannula) to provide a quiet operative field or access to the aortic valve, the heart is deprived of coronary perfusion and becomes ischemic. This is usually managed by perfusing the heart with cardioplegia solutions. (See also Chapter 23 for further discussion on myocardial protection.)

     2. Route of delivery of cardioplegia solutions

        a. Aortic root: A cannula is inserted in the aortic root (proximal to the cross-clamp). Typically this has a “Y” connector: One limb is connected to the cardioplegia delivery system and the other to suction (to vent the left ventricle [LV] or aspirate air). The cardioplegia solution is delivered into the aortic root and thence into the coronary arteries. This is not effective in the presence of severe aortic regurgitation or when the aortic root is open; it is also less effective in the presence of severe proximal coronary artery stenosis. Ideally, pressure in the aortic root should be measured during administration of the cardioplegia to assure adequate coronary flow.

        b. Directly into the coronary ostia: Special hand-held cannulae are placed directly into the right and left main coronary arteries for delivery of the cardioplegia solutions. This is commonly done in the presence of aortic regurgitation or when the aortic root is open.

        c. Retrograde into the coronary sinus: Balloon-tipped cannulae are inserted blindly (or under direct vision if the RA is opened) into the coronary sinus through a purse- string suture in the low lateral wall of the RA. TEE may be helpful in guiding and assessing placement. Many of these cannulae have a pressure port near the tip so that the pressure in the coronary sinus can be monitored during perfusion (ideally maintained between 30 and 50 mm Hg). Retrograde administration of cardioplegia may be advantageous in the presence of severe coronary artery stenosis or aortic regurgitation and during aortic valve surgery. However, it may provide inferior protection of the RV.

     3. Delivery systems. These vary in their complexity. If blood cardioplegia is being used, blood is taken out of the arterial perfusion line following oxygenation and mixed with the crystalloid cardioplegia solution (usually at a blood:crystalloid ratio of 4 to 6:1). This may be accomplished by using two separate roller pumps, or with a single roller pump which drives two sets of tubing of different sizes (to produce the proper flow ratio). The mixture is then passed through a dedicated heat exchanger. Microfiltration may be added and pressures and temperatures are monitored. More complex delivery systems are in use which permit rapid change of the concentration of various components in the cardioplegia solution.

   I. Cardiotomy, field suction, cell salvage processors, and savers

     1. Overview: During CPB there is often considerable bleeding into the surgical field due to systemic heparinization and persistent pulmonary and coronary venous drainage. It is usually not feasible to discard this large volume of blood via conventional discard suction. Traditionally, this blood is removed from the field by roller pumps on the H–L machine (“cardiotomy suckers”) and returned to the H–L machine via the cardiotomy reservoir (which includes filters and, as noted above, is commonly incorporated into the hard-shell venous reservoir, but must be added separately when a soft-shell reservoir is used). The cardiotomy suction should not be used until the patient is adequately anticoagulated and should be discontinued as soon as reversal with protamine is commenced (to avoid clotting of blood in the H–L machine).


     2. Hazards of cardiotomy blood: Cardiotomy blood contains microaggregates of cells, fat, foreign debris, and thrombogenic and fibrinolytic factors, and is thought to be a major source of microemboli and hemolysis during CPB. For these reasons cardiotomy suction should be used sparingly.

     3. An alternative strategy is to suction the field blood into a cell salvage washer/processor/saver system (or to process the blood that has been suctioned into a stand-alone cardiotomy reservoir with this cell salvage system before returning it to the H–L machine). These devices wash the blood with saline, and separate the red cells from the plasma and saline by centrifugation with the intent of reducing the microemboli, fat, etc. It also removes plasma proteins, platelets, heparin, and some of the WBCs, retaining concentrated RBCs (hematocrit about 70%). Not all processors are equally effective at removing fat, and the salvaged blood may have to be specially filtered before administration. Some of the problems associated with the use of cell processors include delayed availability and turnover time (not adequate in the face of rapid hemorrhage) and loss of platelets and coagulation factors (resulting in a consumptive coagulopathy if >6 bowls or 1,500 mL of blood is processed). Comparative studies of cardiotomy suction versus cell processors have produced conflicting results [6].

   J. Venting

     1. Overview: It is important that both the RV and LV are decompressed during CPB to improve surgical exposure, reduce oxygen demands of the myocardium, and attenuate damage to the heart from overdistention.

      2. The RV is readily visible to the surgeon, and decompression depends on adequacy of venous drainage. The LV is more difficult to observe, has more adverse consequences if distended, and requires various strategies for venting.

     3. Consequences of distention of the left heart:

        a. Stretches myocardium causing ventricular dysfunction

        b. Myocardial ischemia: Impairs subendocardial perfusion and increases myocardial oxygen needs

        c. Increases left atrial pressure, leading to pulmonary edema and hemorrhage

        d. Interferes with surgical exposure

     4. Distention of the left heart is likely to occur when the LV is unable to empty (e.g., on initiation of CPB, in the presence of aortic regurgitation, during cardiac arrest, aortic cross-clamping, and administration of antegrade cardioplegia, during ventricular fibrillation, and following aortic cross-clamp release).

     5. Sources of blood coming into the left heart during CPB:

        a. Bronchial venous drainage (normal ~100 mL/min)

        b. Systemic venous blood that bypasses venous cannulae and passes through right heart and lungs

        c. Aortic regurgitation

        d. Patent ductus arteriosus (PDA) (~1/3,500 adults)

        e. Atrial septal defect (ASD), Ventricular septal defect (VSD)

     6. Assessment of adequacy of decompression of the left heart by inspection or palpation is difficult because of its position and thick walls of the LV. The best method of evaluating the adequacy of LV decompression is TEE.

     7. Methods of venting or decompressing the left heart (see Fig. 21.7 and Table 21.3): Cannulae are inserted in various locations and attached to tubing connected to roller pumps which transfers the blood to the venous or cardiotomy reservoirs. The tubing should first be placed under a level of fluid to assure that it is sucking and not emitting air, and the rate of suction must be constantly adjusted to avoid excessive (risk of damage to heart or aspirating air) or inadequate (overdistention) venting.

Table 21.3 Methods of venting the left heart

Figure 21.7 Sites for venting the left heart. A: Aortic root cannula; one limb of the “Y” is connected to the cardioplegia delivery system and the other limb to suction (siphon or roller pump) for venting the aortic root and hence the LV. B: Cannula inserted at the junction of the right superior pulmonary vein with the left atrium and then threaded through the left atrium and MV and into the LV. C: Cannula inserted directly into the apex of the LV. D: Cannula is inserted into the pulmonary artery. AO, aorta; PA pulmonary artery; LA, left atrium. (From Hessel EA II. Circuitry and cannulation techniques. In: Gravlee GP, et al., eds. Cardiopulmonary Bypass. Philadelphia, PA: Lippincott Williams & Wilkins; 2008:90, with permission.)

        a. Aortic root vent (one limb of the antegrade cardioplegia cannula): This is the most common technique used during coronary artery bypass graft (CABG) surgery. Suction is applied to the antegrade cardioplegia cannula (directly or via a side branch). Aortic root venting of the LV is only effective when the aorta is cross-clamped, when antegrade cardioplegia is not being administered, and when the aortic root is not opened.

        b. LV vent placed via the right superior pulmonary vein: A cannula is inserted at the junction of the right superior pulmonary vein with the left atrium and then threaded through the left atrium and MV into the LV. This method is used during aortic and MV surgery (especially in the presence of aortic regurgitation) and for patients with poor LV function.

        c. LV vent placed directly through the LV apex: This is rarely used today because of difficulty in positioning and bleeding after removal.

        d. Vent placed through the left atrial appendage or the top of the left atrium (into either the left atrium or LV). Rarely used.

        e. Vent placed in the main pulmonary artery: This method minimizes the risk of air entry into the left heart (although it can still occur). It does not provide reliable decompression of the LV in the presence of aortic regurgitation, and closure of the incision in the pulmonary artery can be problematic in patients with pulmonary hypertension.

     8. Complications of venting the left heart include systemic air embolism, bleeding, damage to cardiac structures, dislodgement of thrombi, calcium, tumor, etc., and MV incompetence. A critical complication is inadvertent pumping of air into the heart via these vent lines. This occurs if tubing is misaligned in the roller pump-head or if the pump is in reverse mode. Air can also enter the left heart when these lines are inserted or removed; therefore, the volume and pressure in the left heart should be high at these times and ventilation interrupted.

      9. Any time the heart is opened, even by simply placing a catheter in a chamber, air may collect in the heart. If not removed, this air will embolize with resumption of cardiac contractions. Even right-heart air has the potential to pass into the left heart via septal defects or through the lungs. In addition to vigorous attempts at removal of all air before closing the left heart, the use of venting at the highest point of the aorta is considered the final safety maneuver against systemic air embolism. This is most commonly accomplished using the antegrade cardioplegia cannula (see above) which is placed on suction. TEE is particularly useful in assessing the adequacy of deairing. Some surgical teams often flood the surgical field with carbon dioxide during open cardiac procedures. Because of carbon dioxide’s increased solubility there is a potential for reduction of microemboli.

   K. Ultrafiltration/hemoconcentrators. A hemofiltration device (also referred to as ultrafiltration) consists of a semipermeable membrane which separates blood flowing on one side (under pressure) and air (sometimes under vacuum) on the other side. Water and small molecules (sodium, potassium, water-soluble non–protein-bound anesthetic agents) can pass through it and be removed from the blood, but not protein or cellular components of the blood. Hemoconcentrators are used to eliminate excess crystalloid and potassium, and to raise hematocrit (hemoconcentrate). They may also remove inflammatory mediators and hence reduce the systemic inflammatory response syndrome (SIRS). The device is usually placed distal or after the arterial pump with drainage into the venous limb or reservoir. It can be placed in the venous limb of the circuit but would require a separate pump. Five hundred to 2,000 mL or more of fluid may be removed during an adult case. When using post-bypass (but before reversing heparin) it is referred to as “modified ultrafiltration” or “MUF.” MUF is commonly used in pediatric cases but rarely in adult cases [4,7].

   L. Filters and bubble traps

     1. Overview: CPB generates macro- and microemboli of gas, lipids, and other microparticles (WBCs, platelets, foreign debris) which must be filtered out.

     2. Types and location: Many different types of filters (screen and packed fibers [“in-depth”], made of various materials) with various pore sizes are employed in multiple locations in the ECC. These sites include the venous and cardiotomy reservoirs, as part of the oxygenator, in the arterial and cardioplegia lines, and blood administration sets (including the cell processor), and the gas going to the oxygenator. The “in-depth filters” mainly work by adsorption. The clinical importance of the various types of filters remains controversial [6].

     3. Arterial line filter/bubble trap: Most groups employ a microfilter/bubble trap on the arterial line, especially to reduce air embolization [6]. If used, often a clamped bypass line is placed around the filter in case the filter becomes obstructed and a vent line with a one-way valve runs from the filter/bubble trap to the venous reservoir to vent any trapped air.

      4. The employment of leukocyte-depleting filters in various locations in the circuit is advocated by some, but their benefit remains to be proven.

   M. Safety devices and monitors on the H–L machine

     1. Monitors (see Table 21.4)

Table 21.4 Monitors and safety devices

        a. Microprocessor/console monitor and control. Many of the current commercial H–L machines include a microprocessor-driven monitor that displays and controls various functions of the machine and hemodynamic data from the patient.

        b. Inline venous and arterial oxygen saturation, ± other blood gases, electrolytes, glucose, and hematocrit.

        c. Pressure in arterial line: This should be measured proximal to the arterial line filter/bubble trap and distal to the oxygenator. Prior to initiation of CPB, the pressure should reflect a pulsatile wave that correlates with patient’s arterial pressure (to confirm proper intra-arterial placement of the arterial cannula). Excessively high arterial infusion line pressure during CPB (relative to patient’s arterial pressure) indicates a problem in the arterial flow delivery system or at the cannula tip, bearing in mind that there should always be a considerable (50 to 150 mm Hg) pressure gradient across the aortic cannula. Pressures in the line post-CPB can give clues to true systemic pressures. Some also advocate monitoring the pressure proximal to the MO; high gradients (>100 mm Hg) suggest oxygenator dysfunction.

        d. Level sensor in venous reservoir

        e. Bubble/air detector

        f. Arterial line flowmeter: This is required when using a centrifugal systemic arterial pump, and desirable even when using a roller pump.

        g. Temperature: Water-to-heat exchanger and of blood being delivered to the patient

        h. Oxygen analyzer of gas delivered to oxygenator

     2. Safety devices (see Table 21.4)

        a. Low-level alarm ± servo control/turn-off of arterial pump

        b. High arterial line pressure alarm ± servo control/turn-off of arterial pump

        c. Air/bubble detector and alarm ± servo control/turn-off of arterial pump

        d. One-way check valves in the arterial lines, cardiac vents, and arterial filter/bubble-trap purge lines

        e. Arterial line filter

        f. Bypass line around arterial line filter/bubble trap

        g. Purge line off of arterial line filter/bubble trap

     3. Emergency personnel, supplies, and equipment

        a. Second perfusionist

        b. Battery backup for H–L machine including the pumps and monitors

        c. Portable lighting and flashlights

        d. Backup-oxygen supply (cylinders with regulators)

        e. Hand cranks to drive arterial and other pumps

        f. Spare oxygenator

III. Special topics

   A. Surface coating. Many commercial circuits coat all of the surfaces which come in contact with blood (tubing, reservoirs, oxygenators) with various substances (proprietary) designed to minimize the activation of blood components. Many of these coatings include heparin (which should be avoided in patients with heparin-induced thrombocytopenia [HIT]). The clinical benefits of any or of one type of coating over another remains controversial [6].

   B. Miniaturized or minimized circuits: By reducing the surface area and prime volume, miniaturized circuits reduce the amount of hemodilution (resulting in less transfusions) and are purported to reduce the inflammatory response to CPB (producing improved clinical outcomes) [8]. They often feature a closed (veno-arterial loop) circuit (i.e., no venous reservoir or cardiotomy suction) and kinetic-assisted venous drainage. A sophisticated air detection and elimination system is required, as well as stringent avoidance of air entrance in the venous line. Concerns about safety (especially air embolization), inability to handle fluctuations in venous return (especially if the patient has a large blood volume or experiences exsanguination), and lack of cardiotomy and field suction require careful consideration. These systems require close communication amongst all team members. Miniaturized circuits are rarely used in the United States, but may be used more commonly in Europe. Their use is usually limited to uncomplicated CABG and aortic valve surgery, cases associated with minimal intravascular volume shifts and limited need to scavenge much blood from the surgical field.

   C. Pediatric circuits. The major challenges with pediatric CPB are related to the small blood volume of the patient compared with the prime volume of the ECC, and the small venous and arterial cannulae required by the small-sized vessels. Pediatric cardiac surgery groups and industry have made great strides in miniaturization and reduction of priming volumes (some a little at 100 to 200 mL) by including augmented venous return and elevating and the H–L machine closer to the patient (allowing the use of shorter and narrower tubing). Most pediatric surgery centers in North America include albumen in the prime; packed red blood cell or whole blood and fresh frozen plasma are often used for infants [7]. Some groups exclude arterial microfilters and others use an oxygenator which has an integrated “arterial” microfilter. In distinction to adult CPB, inline arterial blood gas monitoring is employed by the vast majority of North American pediatric centers. (See additional discussion in Chapter 14.)

   D. Cerebral perfusion during circulatory arrest

     1. Overview: Circulatory arrest is often required for conduct of surgery involving the aortic arch, and in congenital heart surgery. Deep hypothermia (<18°C) is used to minimize cerebral injury. For periods of circulatory arrest exceeding 25 to 30 min, two strategies for cerebral perfusion are utilized. The benefits and preference of one over the other remain controversial. (See additional discussion in Chapters 24 and 25.)

     2. Retrograde cerebral perfusion (RCP): The arterial line from the H–L machine is connected to the SVC cannula (in the case of bicaval cannulation), or to a cannula inserted through a purse-string in the SVC. The SVC is occluded between the entrance of the catheter and the junction with the RA. Cold blood (15 to 18°C) is then pumped at flow rates of 250 to 500 mL/min and pressures maintained between 20 and 40 mm Hg. It may be desirable to measure the pressure via a catheter placed directly into the right internal jugular (IJ) vein, since valves may reduce the amount of flow and pressure being delivered into this vein. If pressure is measured in this location, it is probably prudent to keep the pressure <25 mm Hg to minimize cerebral edema. An additional benefit of RCP is that atheroemboli and air in the carotid vessels may be “washed out.”

     3. Antegrade cerebral perfusion or selective cerebral perfusion (SCP): Catheter(s) (sometimes with balloon cuffs) connected to the arterial line are inserted into the right innominate or carotid artery or the left carotid and subclavian artery. Cold blood is then infused at a rate of about 10 mL/kg/min and a pressure of about 30 to 70 mm Hg. This technique provides more cerebral blood flow than RCP but adds a risk of arterial trauma and embolization. If the systemic arterial cannula had been inserted into the right subclavian artery (see above), then selective perfusion of the right carotid artery can be accomplished by occluding the proximal innominate artery. If the arterial cannula in the subclavian has been placed via a graft sewn onto the side of the artery, the pressure in the right radial or brachial line provides monitoring of the cerebral perfusion pressure. Obviously this provides only unilateral perfusion and relies on an adequate Circle of Willis to perfuse the left side of the brain.

   E. Less common cannulation

     1. Minimally invasive or port access CPB involves use of smaller incision and often smaller or specially designed venous and arterial cannulae for transthoracic placement, or peripheral cannulation. This may require use of augmented venous drainage and increase the risk of aortic dissection associated with femoral artery inflow. Peripherally placed retrograde coronary sinus catheters (via the right IJ vein by the anesthesiologist), and aortic balloon occlusion catheters and antegrade cardioplegia cannulae (passed via the femoral artery) are used. These require TEE and fluoroscopic guidance for proper placement.

     2. Right thoracotomy. This approach gives excellent views of the MV and RA, but aortic cannulation, occlusion of the ascending aorta, administration of antegrade cardioplegia, and deairing of the LV are problematic. Often some of the minimally invasive cannulation techniques mentioned above are employed.

     3. Left thoracotomy. This approach is used for surgery on the descending thoracic aorta and occasionally for redo MV surgery and CABG surgery for revascularization of the lateral or posterior heart. Venous cannulation is problematic. Peripheral cannulation of the RA via the femoral vein is a commonly employed option (see above). For descending thoracic aortic surgery, isolated partial left heart bypass (LHB) can be accomplished cannulating the left atrium or ventricle directly or via a purse-string in the left superior pulmonary vein or left atrial appendage for venous outflow and the distal aorta or femoral artery for arterial return. The ECC for partial LHB does not require an oxygenator or reservoir and may not include a heat exchanger, typically employs a centrifugal pump and is heparin coated and permits minimal systemic heparinization, but it only supplies oxygenated blood to the lower half of the body. Flows are typically about 1 to 1.5 L/min/M2 and adjusted to adequately decompress the left heart and maintain adequate pressures in the lower and upper part of the body. Perfusion of the upper half of the body (especially to the heart and brain) is provided by the LV and oxygenation to both parts must be provided by the patient’s own lung. Management of partial LHB is quite challenging and requires excellent communication between the anesthesiologist and perfusionist. TEE assessment of LV filling is extremely valuable [9]. See also Chapter 25.

IV. Priming

   A. Overview: The ECC (including venous and arterial lines) must be filled with fluid (“primed”) before use and all air in the circuit eliminated. Circuits are usually primed with asanguineous fluids. To minimize hemodilution much effort has recently been directed at reducing the priming volume of ECC (as low as 1,000 to 1,250 mL for adults).

   B. Consequence of asanguineous primes: Priming results in hemodilution with reduction of hematocrit, plasma proteins (decreased oncotic pressure), and coagulation factors. Controversy surrounds the acceptable lower level of hematocrit [6]. However, prior to initiation of CPB, the predicted hematocrit should be estimated to determine if the team wishes to add RBCs to the prime.

      yBaseline hematocrit used for this calculation should be obtained immediately prior to CPB to account for any crystalloids administered prior to CPB.

   C. Retrograde autologous priming (RAP) is a method of reducing hemodilution. Before commencing CPB, arterial blood is drained retrograde to displace asanguineous prime in the arterial line (which is sequestered in a collection bag). Immediately before going on CPB, venous blood is also allowed to drain out of the patient through the venous line into the collection bag (“antegrade autologous priming”). With this method 500 to 1,000 mL of asanguineous prime may be eliminated. However, it is associated with a reduction of the patient’s blood volume and may result in hypotension. Placing the patient in Trendelenburg position and administration of a vasopressor is usually required. The majority of randomized trials have demonstrated that RAP reduces perioperative packed RBC transfusions and is usually safe. However, this technique can be very dangerous in certain high-risk patients, leading to hypotension and requiring initiation of immediate bypass to rescue the patient. Further studies are needed to determine the effect of retrograde priming on major morbidity and mortality.

   D. Composition of the prime: Many formulations are in use. Most use a balanced electrolyte solution without glucose. Much controversy surrounds the need to add colloid, and use of both albumen and hydroxyethyl starches (HES) have been advocated (although use of high-molecular-weight HES [HetastarchTM] is not recommended by the FDA in the United States). Many add mannitol to the prime and most include heparin (about 2,500 units/L).

   E. Priming of the circuit. The perfusionist fills the circuit with the priming fluid and circulates it employing various maneuvers to remove all air. Often before introducing the prime the circuit is flushed with carbon dioxide which is more easily removed from the prime than air bubbles. Usually, a prebypass microfilter is temporarily included in the circuit during this recirculation process to remove any foreign particles.

   F. Final disposition of prime at the end of CPB: Usually, as much of this volume as possible is returned to the patient before removing the arterial line. That which is left may either be pumped directly (sometimes through a hemoconcentrator) into an IV line in the patient, or placed in an IV bag for administration by the anesthesiologist (preserving platelets and protein but also containing heparin which will require neutralization), or first processed by a cell-washing device and hemoconcentrated (does not contain heparin).

V. Complications, safety, and risk containment

   A. Incidence of adverse events: Three surveys covering practice between 1994 and 2005 have reported rates of adverse events associated with CPB of 1/35, 1/135, and 1/198 (average 1.3%), with a severe injury or death rates of 1/1,286, 1/1,453, and 1/3,220 (average 0.04%) [1012].

   B. Specific complications

     1. Aortic dissection occurs in about 0.06% of cases (about 0.05% following cannulation of ascending aorta, 1% with femoral artery cannulation, and 0.8% with axillary/subclavian cannulation); adverse outcome occurs in about 20% of these cases. It is usually associated with arterial cannulation but can also be related to cross-clamping, aortic vents, and aorto-coronary anastomoses. Dissection often presents as low arterial pressure, high arterial line pressure, loss of venous return (hidden blood loss), bluish discoloration of ascending aorta, and signs of ischemia in various organs. It is best diagnosed with ultrasound imaging (TEE or epiaortic ultrasound). Although dissection usually presents early it can occur at any time during CPB. Repair of the dissection is usually required (after moving the arterial cannula to an alternate site, and induction of deep hypothermic circulatory arrest). If recognized early in the course of CPB, especially when associated with femoral cannulation, the situation may resolve by simply coming off CPB.

     2. Massive air embolism occurs in about 0.01% of cases with an adverse outcome in 35% of events. It is almost always preventable with vigilance and employment of various safety devices and techniques. The most common cause is pumping air out of an empty venous reservoir; other causes are related to LV vents, pressurization of cardiotomy or venous reservoirs, and premature ejection by the LV before adequate deairing. Prevention is key. Successful management, when it occurs, requires the collaborative efforts of all members of the team. Establishment of protocols and team practice during simulation are essential. Further discussion can be found in Chapter 8 and management is detailed in Table 8.3.

     3. Oxygenator failure occurs in about 0.05% of cases with adverse outcomes in 10% of cases. This is usually manifested by hypoxemia, but can also present with hypercapnia. Excessive oxygen consumption (reflected by a low venous saturation) or excessive carbon dioxide production (often related to uptake from the field suction if the field is being flooded with carbon dioxide) should be ruled out. Failure of an oxygenator is often preceded by increased pressure gradients through the oxygenator. Systematic examination of entire oxygen supply system to the oxygenator should be conducted. Disconnections, especially at the anesthesia vaporizer, are common causes, as is malfunction of the blender. An alternate source of oxygen should be explored. If the cause is thought to be the MO and the heart is still functioning, separation from bypass should be considered. If the heart is arrested, then the oxygenator must be replaced. To accomplish this, circulation must be interrupted. If time permits, cooling the patient as rapidly as possible is prudent. With proper training, supplies, and practice, perfusionists should be able to do this in less than 3 min [13].

     4. Arterial pump malfunction: Malfunction of the arterial pump occurs in about 0.06% of cases. This can be due to electrical or mechanical failure and can result in either excessive flow (“run-away”) or cessation of function. Hand cranks must be available.

     5. Inadequate anticoagulation and clotting of the circuit has been encountered in about 0.03% of cases. It is usually related to inadequate heparinization. Periodic monitoring of anticoagulation is required (see also Chapter 19). While on CPB the ACT can overestimate heparin levels, therefore most groups advocate adding heparin (~100 units/kg) hourly even if the ACT is satisfactory. Monitoring heparin levels (e.g., by protamine titration) may alleviate this discrepancy. A catastrophic accident is inadvertent administration of protamine during CPB (often due to miscommunication). Another potential problem is clotting of the ECC after CPB due to continued use of the cardiotomy suction after administration of protamine. If this occurs, it is impossible to resume bypass if required urgently to rescue a patient [11]. It is therefore prudent to discontinue the use of cardiotomy suction as soon as protamine administration is initiated.

     6. Dislodgment of cannulae is encountered in about 0.06% of cases and tubing rupture in about 0.03%. Vigilant care of these cannulae must be maintained. If a venous line dislodges, an “air-lock” occurs (see below), and arterial flow will have to be stopped to avoid pumping air out of a depleted venous reservoir. Dislodged arterial cannulae also require that the arterial pump be stopped until the cannula is placed back into the arterial system (after excluding all air). Tubing rupture (usually in the high-pressure part of the circuit) again requires cessation of CPB while the defect is repaired.

     7. Obstructed venous return and air-lock. This complication poses two problems: elevation of venous pressure (with possible decreased cerebral perfusion, distention of the right heart, and decreased venous return to the H–L machine) and risk of pumping air if the reservoir becomes empty. Vigilant monitoring of the pressure in the SVC and of the level in the venous reservoir is essential. The entire venous drainage circuit must be examined. Sites of obstruction include clamps, kinks, malposition of cannulae, and an air-lock. An air-lock occurs when a bolus of air fills the venous drainage line and interrupts the siphon. The source of the air leak must be corrected and then the air emptied into the venous reservoir by sequential elevation and lowering of the venous line.

     8. Erroneous systemic blood flow: Due to miscalibration or malfunction of the arterial pump or the flowmeter; may lead to excessive or inadequate blood flow.

     9. Vasoplegia. Vasodilation during and following CPB is a frequently encountered problem which if persists is associated with adverse outcome. It is defined as low arterial pressure despite a normal or elevated systemic blood flow (cardiac output). Evaluation begins with the differential diagnosis of low arterial pressure (e.g., artifactual measurement, low blood flow [malfunctioning or miss calibrated arterial pump, open arteriovenous shunts], reduced viscosity [severe anemia], allergic reaction, infusion of vasodilating drugs, aortic dissection and sepsis). Purported risk factors for the vasoplegic syndrome include preoperative Angiotensin conversion enzyme-1 (ACE-I), angiotensin receptor blockers (ARBs) and calcium-channel blockers, and Ventricular assist device (VAD) placement. The next decision is whether or not to treat the low pressure (see discussion in subsequent section on pathophysiology). This is based upon the severity of the hypotension, patient factors [e.g., age, coexisting disease (vascular, renal, diabetes, central nervous system [CNS])] and whether there is evidence of inadequate tissue perfusion (venous saturation, acid–base status and lactates, cerebral oximetry, urine output). Therapeutic options include raising blood flow, raising hematocrit, and adding a vasoconstrictor. Phenylephrine and norepinephrine are most commonly used initially, with vasopressin (2 to 4 units/hr) and methylene blue (1.5 mg/kg bolus ± 1 mg/kg/hr) often used as second-line agents.

     10. Heater–cooler dysfunction occurs in about 0.16% cases and can lead to excessive heating or cooling. Close monitoring of water inflow temperature and temperature of arterial blood leaving the oxygenator (which contains the heat exchanger) is essential to detect these problems. If malfunction is detected, the heater–cooler must be inactivated and other means of temperature control initiated.

     11. Electrical failure (H–L machine only or entire room) has been reported in 0.08% of cases. All modern H–L machines should have battery backup which should be tested before each case. If battery backup fails, hand cranking of the arterial pump must be initiated immediately. Often overlooked is the loss of field suction and left heart vents, which could impair surgical exposure and lead to loss of venous return; thus, the pumps driving these systems must also be hand cranked. Electrical failure may also cause loss of critical monitoring and lighting.

     12. Gas (oxygen) supply failure is encountered in about 0.03% of cases. Since most teams do not monitor flow into the oxygenator, this complication may present as a cause of oxygenator failure (see above). A standby oxygen supply (cylinder) must be available.

     13. High arterial line pressure demands immediate attention. It may be a sign of systemic arterial hypertension, malposition of the arterial cannula, arterial dissection, kinks or clamps in the arterial cannula or line, inadequate cannula size, excessive flow, or an obstructed arterial line filter. Arterial flow must be reduced until the problem is identified and rectified.

     14. Distension of right and/or left heart (see section on venting above)

   C. Risk containment. This requires the active and continuing participation of all members of the team (surgeons, perfusionists, anesthesiologists, and nurses).

      1. Vigilance on the part of all members of the team

     2. Special monitoring of the adequacy of perfusion is discussed in Part II of this chapter and in Chapter 8 and by Murphy et al. [6]. Two issues deserve special attention:

        a. The CNS. Many advocate the use of cerebral near-infrared spectroscopy (NIRS) (i.e., cerebral oximetry) or other monitors of cerebral perfusion/function (e.g., transcranial Doppler, processed electroencephalography) to detect problems with venous drainage and arterial cannulation and malperfusion (e.g. with dissection) [14]. This is discussed further in Chapter 24.

        b. TEE. TEE is useful not only to diagnose cardiac abnormalities pre-CPB and evaluate surgical repairs but to assist with the conduct of CPB. Some of these applications include the following:

           (1) Evaluating atherosclerosis in the aorta (often adding epiaortic scanning) as it relates to cannulation and placement of clamps

           (2) Evaluating placement of cannulae, especially retrograde coronary sinus cannula, LV vents, and transfemorally placed IVC cannulae, and intra-aortic balloon pump (IABP)

           (3) Detect devices, masses (thrombi and tumors), and anatomic abnormalities that could affect cannulation

           (4) Assess adequacy of decompression of the LV

           (5) Detect and evaluate deairing of left heart

           (6) Detect aortic dissection and malperfusion of arch vessels

      3. Education, practice, experience, retraining, certification, and recertification

      4. Communication between surgeon, perfusionist, and anesthesiologist. Each must warn team members of actions that could impact all the others, and of any variance with normal course of CPB or deviations in expected parameters. Commands must be positively and verbally acknowledged.

     5. The “two-minute drill”: Although serious complications occur in only about 0.25% to 0.1% of the time, it is prudent to wait about 2 min after going on “full bypass” and arresting the heart, to assure that all is going well and to rule out serious complications which can be most easily managed by discontinuing CPB before the heart is arrested, and resuming normal circulation. One should confirm the following endpoints:

        a. Able to achieve targeted flow

        b. Adequate venous return and not losing volume

        c. Adequate oxygenation from the oxygenator

        d. Acceptable arterial pressure and have excluded arterial dissection as the cause of hypotension

        e. RV and LV are decompressed

        f. Acceptable systemic venous pressure

        g. Acceptable arterial line pressure

        h. Acceptable venous oxygen saturation

      6. Planning, development of, and adherence to protocols for routine as well as unusual types of CPB and of complications

      7. Use of prebypass check-lists and for other key times during CPB and for adverse events

      8. Use of safety equipment and alarms

      9. Appropriate preventive maintenance program, replacement of old equipment, and familiarity and testing of new equipment

     10. Team practice at diagnosing and management of major complications

     11. Periodic audit, team meetings, quality assurance (QA) and quality improvement (QI). Use of a registry to measure variation and benchmarking

     12. Automated control and regulation of the H–L machine

   D. Key role of the anesthesiologist in the conduct of CPB. Anesthesiologists are in a unique position to assist with the conduct and management of CPB. They are able to observe both the surgical field and the H–L machine, and can facilitate communication between the perfusionist and surgeon. Their detailed knowledge about the patient’s medical history and the patient’s course prior to arrival in the operation room and prior to CPB gives them a unique perspective.

   Anesthesiologists should oversee the anesthetic and vasoactive drug management and patient monitoring during CPB and make recommendations concerning the safe and appropriate conduct of the CPB. They should be vigilant for any adverse events and assist with the management of the complications mentioned above. Finally, anesthesiologists should be involved in collaboration with the surgeons and perfusionists in the development of protocols for the conduct of safe CPB, participate in practice sessions for handling emergency situations and complications, and be involved in education, and QA/QI activities related to perfusion.


I. Introduction

Improvements in the design of the CPB circuit and greater understanding of the physiologic insult of CPB have contributed to the relative safety of modern cardiac surgery. Despite advancements in technology and knowledge during the past several decades, a variety of minor and major complications are observed following CPB.

Major physiologic trespasses introduced by CPB include: (i) Alterations of pulsatility, blood flow patterns, and pressure; (ii) exposure of blood to nonphysiologic surfaces and shear stresses; (iii) hemodilution; (iv) systemic stress response; and (v) varying degrees of hypothermia (or hyperthermia during rewarming). Improving the safety of CPB will depend on greater understanding of these aberrations [15].

II. CPB as a perfusion system

   A. Circulatory control during CPB. “Cardiac output” on CPB is the pump flow rate, which can be set at any level desired, but is limited by the amount of venous return. Systemic and venous blood pressures are partially dependent on the patient’s autonomic tone, but can be manipulated by increasing or decreasing venous drainage and by administering vasopressors or vasodilators. Thus, the circulation during CPB is controlled in large part by the perfusionist and the anesthesiologist.

     1. Systemic blood flow. Systemic blood flow is determined by pump flow of the H–L machine, which is set by the perfusionist. This should be guided by the patient’s age, temperature, depth of anesthesia, and hematocrit.

        a. Pump flow rates are usually expressed as L/min/M2. In awake patients, it is generally accepted that a cardiac index less than 2.0 to 2.2 L/min/M2 is not sufficient to provide tissues with an adequate oxygen supply. This also appears to be the lower limit of sufficient cardiac output during normothermic CPB.

        b. At moderate hypothermia (about 30°C) and with a hematocrit of about 24% in an adult, flow is often set at about 2.4 L/min/M2, which meets the oxygen needs of an anesthetized patient. Whether this remains true at higher temperatures (e.g., 35°C) remains to be determined.

        c. With increasing degrees of hypothermia, the patient’s oxygen demand decreases, and consequently pump flow rates may be reduced significantly. Kirklin and Barratt-Boyes [16] calculated curves relating oxygen consumption to pump flow rates at different temperatures and describe “best-fit” lines for measured oxygen consumption (VO2) at varying nonpulsatile flow rates (Q) from several animal studies (Fig. 21.8). The small xs on each curve represent flow rates at each temperature used clinically during CPB at the University of Alabama at that time.

Figure 21.8 Nomogram relating oxygen consumption (VO2) to perfusion flow rate (Q) and temperature. (From Kirklin JW, Barratt-Boyes BG. Hypothermia, circulatory arrest, and cardiopulmonary bypass. In: Kirklin JW, Barratt-Boyes BG, eds. Cardiac Surgery. 2nd ed. New York, NY: Churchill-Livingstone; 1993:91, with permission.)

        d. Maximal flow rate during CPB is limited by venous return from the patient, which is influenced by the height of the operating table above the H–L machine, placement, resistance, and size of venous cannulae and lines, blood volume, and venous tone. Maximal flow is also limited by the capacity of the H–L machine and size of the arterial cannula. High flows through the arterial cannula produce high pressure gradients and turbulence, which damage the blood and produce adverse jet effects (e.g., dislodgment of atheroemboli).

        e. At the present time, there are no data defining a minimal (or maximal) safe flow rate during hypothermic or normothermic CPB, nor has the optimal flow rate which results in the most favorable organ perfusion (and best clinical outcomes) been determined.

     2. Arterial pressure. As in the normal state, arterial pressure is the product of cardiac output (i.e., pump flow during CPB) and systemic vascular resistance (SVR). The latter is determined by blood viscosity and by smooth muscle tone in the arterioles. Viscosity is principally influenced by hematocrit and temperature, both of which often change considerably during CPB.

        a. At normothermia, if hematocrit falls from 40% to 20%, viscosity (and hence SVR) will fall about 50% and at a constant pump flow (cardiac output) the mean arterial pressure (MAP) will fall about 50%.

        b. SVR is also determined by vascular tone, which is influenced by pulsatility (SVR lower with pulsatile flow), sympathetic nervous system activity, depth of anesthesia, catecholamines, angiotensin and arginine vasopressin (AVP) and other local vasoactive substances (e.g., nitric oxide and endothelin), acid–base and electrolyte status, various mediators of the systemic inflammation reaction (SIR), and administration of vasoactive drugs.

        c. The optimal range of arterial pressures during CPB remains controversial. If pressure is too low, perfusion of critical vascular beds may be compromised, especially if vascular disease is present. Conversely, excessive arterial pressure increases noncoronary collateral flow to the heart during aortic cross-clamping (hence “washing out” myocardial protection with cardioplegia) and bronchial flow to the lungs (which increases blood return to left heart) and places strain on arterial clamps and suture lines.

        d. Many clinicians maintain MAP of 50 to 60 mm Hg during adult CPB. This value is based on data suggesting a lower limit of autoregulation of the brain of 50 mm Hg in awake subjects. However, the lower limit of cerebral autoregulation may be as low as 20 to 30 mm Hg during hypothermic bypass with moderate hemodilution [17]. This explains why short periods of hypotension (MAP of 30 mm Hg or less) are well tolerated for brief periods of time.


        e. More recent data have suggested that the lower limit of autoregulation of the brain is approximately 70 mm Hg in awake, normotensive subjects, while during CPB in adults it has been found to average 66 mm Hg (but with a 95% prediction interval of between 43 and 90 mm Hg) and to NOT be related to preoperative blood pressure or hypertension, age, gender, diabetes or a history of prior cerebrovascular accident[6,18]. On the basis of these data, some clinicians use higher MAP (≥70 mm Hg) on CPB. Higher perfusion pressures may improve tissue perfusion in high-risk patients (older, hypertensive, peripheral vascular disease) and enhance collateral blood flow when emboli obstruct blood vessels.

        f. Observational studies examining the association between hypotension on CPB (usually defined as an MAP <50 mm Hg) and adverse outcomes have yielded conflicting results (outcomes either unchanged or worsened) [6]. In a randomized trial of low (targeted pressures of 50 to 60 mm Hg) or high (targeted pressures of 80 to 100 mm Hg) pressure on CPB, the combined incidence of adverse cardiac and neurologic outcomes was lower in the high-pressure group [19].

        g. There is insufficient evidence to recommend an optimal arterial pressure during CPB. In addition, the best method of increasing arterial pressure which may beneficially impact clinical outcomes is undetermined (increasing flow, increasing hematocrit, use of vasoactive medications). The decisions relating to arterial pressure on CPB should be based on an assessment of the benefits and risks of higher or lower perfusion pressures for each patient. In the near future it may be possible to individualize MAP during CPB utilizing real-time monitoring of cerebral autoregulation [18].

     3. Venous pressure. Venous pressure is determined by blood volume, venous tone (sympathetic nervous system, depth of anesthesia, vasoactive drugs), resistance to flow out of the venous cannula (placement, size, kinks, distortion of heart), height of operating table, and total systemic flow.

          Venous return is normally by gravity (siphon), but sometimes is augmented by applying vacuum or suction to the venous lines. Elevated venous pressure can seriously compromise organ perfusion and can lead to peripheral edema.

     4. Distribution of blood flow. In addition to total blood flow, one must be concerned about flow in each organ [20]. Recent studies have noted a hierarchy of distribution of blood flow during normothermia and hypothermia as total flow is reduced [21]. Even at “normal flow” (i.e., 2.4 L/min/M2), muscle flow is significantly reduced during CPB. As flow is progressively reduced, first splanchnic, then renal, and eventually (only at extremely low flows) cerebral flows are reduced.

   B. Circulatory changes during CPB

     1. Changes at onset of CPB. At commencement of CPB, there is usually a fall in systemic blood pressure due to a decrease in SVR. This phenomenon results from the following:

        a. Decreased blood viscosity secondary to hemodilution by the pump-priming fluid

        b. Decreased vascular tone secondary to the following:

           (1) Dilution of circulating catecholamines

           (2) Temporary hypoxemia. Hypoxemia from initial circulation of pump asanguineous priming fluid may lead to decreased vascular tone.

           (3) Low pH and low calcium and magnesium levels in the priming fluid

     2. Circulatory changes during hypothermic CPB

        a. Increased SVR. There may be considerable patient-to-patient variations in SVR during CPB. However, as CPB progresses, there will generally be a steady increase in systemic pressure due to increasing SVR if flow rates are kept constant. The observed increase in SVR during the course of CPB is due to several factors:

           (1) Decreased vascular cross-sectional area from closure of portions of the microvasculature

           (2) Vasoconstriction brought on by the following factors:

             (a) Hypothermia

             (b) Increasing levels of circulating catecholamines, AVP, endothelin, and angiotensin II

           (3) Increase in blood viscosity secondary to hypothermia and rising hematocrit (due to urine output or translocation of fluid into the interstitial compartment)

        b. Decreased SVR. Transient decreases in SVR and systemic pressure may be observed shortly after infusion of cardioplegic solutions, especially if the solutions contain nitroglycerin.

     3. Circulatory changes during the rewarming phase of CPB

        a. As the perfusate temperature is increased to rewarm the patient, variable circulatory responses are observed depending on the anesthetics used, patient hematocrit, underlying disease, and other factors. SVR and MAP increase frequently during initial rewarming from 25 to 32°C, but then usually decrease as temperature increases above 32°C.

        b. A more consistent decrease in SVR and MAP usually occurs with release of the aortic cross-clamp and reperfusion of the heart. Despite cardioplegia and hypothermia, there is some degree of ongoing metabolic activity and utilization of myocardial energy stores during the ischemic period. This results in coronary vasodilation and a marked increase in coronary blood flow and a decrease in arterial pressure. In addition, when the heart is reperfused, accumulated metabolites are washed out of the heart into the general circulation. Some of these metabolites, most notably adenosine, are potent vasodilators which aggravate the decrease in SVR.

     4. Changes in the microcirculation and adequacy of tissue perfusion during CPB

        a. During CPB, cardiac output and arterial pressure can be easily maintained at “normal” values. However, several observations suggest that tissue perfusion and oxygen delivery can be impaired to varying degrees during CPB, including the following:

           (1) Postoperative organ dysfunction, both temporary and permanent

           (2) Variable decreases in oxygen consumption during normothermic CPB at flows and pressures that are comparable to pre-CPB values

           (3) Increases in serum lactate levels

        b. The microcirculation lies between the precapillary arterioles and the postcapillary venules and includes the capillary bed, interstitial fluid space, and microcirculatory lymphatics. Normal microcirculatory physiology is poorly understood and requires further clarification. However, it is clear that microcirculatory function during CPB may be impaired by the following:

           (1) Constriction of precapillary arteriolar sphincters caused by catecholamines, angiotensin, vasopressin, thromboxane, endothelin, and decreased release of nitric oxide (NO)

           (2) Increased interstitial fluid volume (edema)

           (3) Decreased lymphatic drainage

           (4) Loss of pulsatile flow

           (5) “Sludging” in the capillaries due to hypothermia

           (6) Altered deformability of RBCs

           (7) Microaggregation and adhesion of white cells, platelets, and fibrin onto the endothelium related to the SIR

           (8) Microemboli (gas, lipids, cellular aggregates), primarily from the cardiotomy suction

        c. Attempts to optimize microcirculatory function during CPB may include use of vasodilators to inhibit arteriolar constriction, addition of mannitol and colloid (e.g., albumin) to the pump-priming fluid to inhibit interstitial fluid accumulation, use of pulsatile perfusion techniques, hemodilution to a hematocrit between 20% and 30% to optimize capillary flow, use of microfiltration, minimizing return of unprocessed cardiotomy suction blood directly into the H–L machine, and anti-inflammatory strategies.

     5. Pulsatile versus nonpulsatile flow during CPB. One of the major physiologic derangements introduced by CPB is loss of pulsatility of flow. Intuitively, it seems desirable to reproduce normal flow patterns as closely as possible during CPB. However, there is considerable controversy about the merits of and need for pulsatile perfusion compared with conventional nonpulsatile perfusion [6].

        a. How to produce pulsatile flow. Several methods are commonly used to achieve arterial pulsations during CPB:

           (1) If partial CPB is being used, venous drainage can be reduced to permit some cardiac ejection.

           (2) If an intra-aortic balloon is in place, it can be used to impart pulsatility to the flow.

           (3) Pulsations can be produced by roller pumps, and to a lesser degree by centrifugal pumps, designed to rotate at varying speeds.

        b. Damping effects of the aortic cannula. The first two methods of producing pulsations are more effective because they generate the pulse in the aorta itself. Although many pumps can generate a pulsatile outflow, the amount of pulsatile energy transmitted into the aorta is limited by the damping effects caused by the narrow aortic cannula, MOs, and arterial microfilters. This makes it unlikely that much pulsatile power can be transferred into the patient by roller pumps [22].

        c. Putative benefits of pulsatile flow

           (1) Transmission of more energy to the microcirculation, which improves tissue perfusion, lymphatic flow, and cellular metabolism

           (2) Reduction of adverse neuroendocrine responses (mainly vasoconstrictive) to nonpulsatile flow that emanate from baroreceptors, the kidneys, and the endothelium

        d. Liabilities of attempting to generate pulsatile flow

           (1) Increased cost and complexities

           (2) Requires use of larger arterial cannulae

           (3) Is associated with higher nozzle velocities out of the arterial cannula (risking vascular injury and thromboembolism)

        e. Clinical outcome. Clinical outcome data have been conflicting. A recent evidence-based review concluded that existing data were insufficient to support recommendations for or against pulsatile perfusion to reduce the incidence of complications following CPB [6].

III. Adequacy of perfusion

   A. How to define. There is no generally accepted definition of optimal perfusion during CPB. Perfusion can be considered acceptable if the patient survives without evidence of organ dysfunction. However, the primary objective of optimal perfusion is to produce a healthy, productive long-term survivor of cardiac surgery. Therefore, perfusion during CPB should accomplish the following goals:

     1. Maintain adequate oxygen delivery to all organs (arterial oxygen content and delivery)

     2. Avoid activation of undesirable reactions, e.g., neuroendocrine stress response and inflammation

     3. Minimize microembolization and disturbance of the coagulation system

     4. Maintain adequate systemic blood flow and arterial pressure

   B. Monitoring. Detailed monitoring of organ perfusion and function during CPB is not usually employed during clinical CPB; however, research studies of this nature have provided much useful information to guide improvements in conducting CPB.

     1. Global perfusion

        a. Oxygen consumption (VO2) measurement, although not commonly used clinically, has provided much insight into the proper conduct of CPB. It can be easily calculated from simultaneously measured arterial and venous oxygen contents and pump flow rate:

            Kirklin and Barratt-Boyes [16] suggested that maintaining VO2 at 85% of the predicted maximum for a given temperature will provide adequate oxygen delivery (Fig. 21.8).

            Normal oxygen consumption in an awake normothermic resting person is about 120 to 140 mL/min/M2 and during deep general anesthesia may be about 90 to 110 mL/min/M2 [23], and falls about 5% per degree centigrade of hypothermia (i.e., below 37°C).

            A fall in oxygen consumption during CPB while the other determinants of oxygen consumption (e.g., depth of anesthesia, muscle relaxation, temperature) remain stable suggests impaired oxygen delivery (even if mixed venous oxygen saturation remains stable).

        b. Oxygen delivery (DO2 =Cao2 ×pump flow) may be one of the most important determinants of adequacy of perfusion. The DO2 calculation incorporates two critical variables that determine tissue oxygenation (arterial oxygen content and pump flow rate) into a single measure. During CPB, DO2 can be improved by raising hematocrit values (transfusion or hemoconcentration), increasing inspired oxygen concentrations, or increasing pump flow rates (e.g., significant hemodilution on CPB can be compensated for by increasing flow rates). Normal oxygen delivery in awake subjects is about 700 mL/min/M2 and the lower limit of normal is about 350 mL/min/M2. This value is lower in the hemodiluted patient under CPB (typically 200 to 300 mL/min/M2). Therefore, the safe margin between oxygen supply and demand may be reduced on bypass.


            Critical oxygen delivery is that point at which maximum oxygen extraction is reached and oxygen consumption starts to fall (i.e., become flow dependent). Tissue hypoxemia and systemic acidosis (lactic) begin to occur at this point. The existence of a critical level of oxygen delivery during CPB has been the subject of active debate. Some studies have found this threshold to be 330 mL/min/M2 at 35°C, 272 mL/min/M2 at 32 to 34°C, and 243 mL/min/M2 at 32°C in humans prior to or during CPB [23,24]. However, other investigators have observed a direct linear relationship between DO2 and VO2, and have been unable to define a critical DO2 level.

        c. Mixed venous oxygen saturation (SvO2), content (Cvo2), partial pressure (PvO2), or oxygen extraction ratio (OER) provide clues to the adequacy of the balance of oxygen delivery (DO2) to oxygen demand (VO2). OER is the ratio of VO2/DO2. Normally, SvO2 is about 75% and OER about 25%. When these two values approach 50%, critically compromised oxygen delivery is suggested. Inline monitoring of mixed venous saturation or partial pressure is advocated during CPB, and clinicians should strive to maintain the SvO2 at 80%. Unfortunately, global venous oxygen saturation values can fail to detect regional ischemia if the vascular bed is small or if there is too little desaturated blood returning from a poorly perfused bed [25]. Furthermore, SvO2 has been found not to be sensitive to critical oxygen delivery nor to the development of hyperlactatemia during CPB [26]. Thus, although a low venous oxygen saturation should always be remedied, a normal or high venous saturation does not assure adequate perfusion to all organs.

        d. Metabolic and lactic acidosis

        e. Cerebral oximetry (e.g., NIRS). In addition to its role as a monitor of the adequacy of cerebral oxygen delivery, recent studies have suggested that it may be a useful surrogate to assess adequacy to total body oxygen delivery [27]. As flow is maintained to the brain at the expense of other organs, decreases in cerebral oxygenation suggest that flow to all other tissues is impaired. A critical issue with this technology is categorizing thresholds which identify pathologic brain perfusion. At the present time, clinical studies suggest that reductions in cerebral oxygenation of 20% to 25% relative to baseline are associated with organ dysfunction and adverse outcomes [27].

     2. Organ-specific perfusion

        a. Brain. Electroencephalography (raw and processed, e.g., bispectral index) and transcranial Doppler are used to monitor cerebral perfusion, but their value is debated. Cerebral oximetry can be used as an indirect monitor of cerebral blood flow; however, only limited data support a beneficial effect of this monitoring on clinical outcomes [28]. Although some have suggested that such monitoring should be a standard of care, this is an unresolved controversy [14,29,30]. Jugular venous saturation, pressure, and temperature may give additional insight into how well the brain is being supported. In the research setting, release of cerebral enzymes (e.g., S-100, enolase) into cerebrospinal fluid or systemic blood is used as sensitive indicators of CNS injury.

        b. Heart. Monitoring the electrocardiography (ECG), TEE, myocardial temperature, myocardial tissue pH, PCO2, PO2, coronary sinus lactates, and cardiac enzymes have been advocated to assure adequate support of the heart.

        c. Kidney. Urine output is the simplest measure of renal function. However, different blood flow patterns, varying perfusion pressures, hypothermia, and the presence or absence of diuretics in the pump-priming fluid may affect urine output and render it an inaccurate indicator of overall tissue perfusion. Measurement of kidney-specific tubular proteins (e.g., cystatin C, KIM, NGAL, GST, NAG, α1-MG, NEP, and RBP) in the urine may represent more sensitive indicators of renal injury than changes in postoperative creatinine and creatinine clearance/glomerular filtration rate [31].

        d. Splanchnic bed. No monitoring is usually employed clinically at this time, but use of gastric tonometry (saline or air, pH, or PCO2), Doppler assessments of mucosal blood flow, hepatic blood flow measurement, and hepatic venous oxygen saturation monitoring have been used in clinical investigations.

IV. Hypothermia and CPB

   A. Effects of hypothermia on biochemical reactions. The Q10 for chemical reactions is a measure of changes in rate of reaction for each 10°C change in temperature. For human tissues, Q10 is approximately 2. That is, for each 10°C decrease in body temperature, the rate of reaction (i.e., metabolic rate or oxygen consumption) is roughly halved.

   B. Effects of hypothermia on blood viscosity. Hypothermia increases blood viscosity. In the early history of CPB, hemodilution was not performed. In contemporary cardiac surgical practices, patients are typically hemodiluted to hematocrits of 20% to 30% during CPB (due to the inevitable hemodilution of the patient’s red cell mass with the asanguineous pump-priming solution). Although oxygen-carrying capacity is decreased from hemodilution, oxygen delivery may be improved due to decreased viscosity and enhanced microcirculatory flow. Data suggest that viscosity remains stable if hematocrit (%) matches temperature (in °C) during hypothermia. The optimal degree of hemodilution during hypothermic CPB has not been determined. Recent studies have demonstrated an association between severity of hemodilution on CPB (hematocrits below 22% to 23%) and morbidity and mortality [32]. Clinicians should avoid both excessively high hematocrits (increased blood viscosity and decreased microcirculatory flow) and low hematocrits (inadequate oxygen content) during hypothermic CPB.

   C. Changes in blood gases associated with hypothermia

     1. Changes in oxygen–hemoglobin dissociation curve. As the temperature decreases, the affinity of oxygen for hemoglobin increases (i.e., the oxygen–hemoglobin dissociation curve is shifted to the left). A lower partial pressure of oxygen in the tissues is required to remove the same amount of oxygen from the hemoglobin molecule.

     2. Changes in solubility of O2 and CO2. As temperature decreases, gases become more soluble in liquid. For a given partial pressure more gas will be dissolved in the plasma. This is more significant for CO2 due to a higher solubility in plasma at any given temperature.

     3. Neutrality of water. Neutral water is water in which the [H+] is equal to the [OH–]. At 37°C, the pH of neutral water is 6.8. At 25°C the pH of neutral water is 7. As temperature decreases, the pH at which water is “neutral” changes in a linear fashion. The neutral pH increases 0.017 units for each degree Celsius decrease in temperature (Fig. 21.9).

Figure 21.9 Blood pH of exotherms and homeotherms, and pH of neutral water as a function of body temperature. (From Ream AF, Reitz BA, Silverberg G. Temperature correction of Pco2 and pH in estimating acid-base status: An example of the Emperor’s new clothes? Anesthesiology. 1992;56:42, with permission.)

     4. Differing strategies for measuring and managing blood gases during CPB. Blood gases are measured at 37°C in blood gas analyzing machines. If the patient’s body temperature is lower than 37°C, pH and PaCO2 can be corrected to determine their actual values at the patient’s temperature. If the patient’s temperature is 27°C and the pH and PCO2 as measured at 37°C are 7.4 and 40, respectively, then the pH and PaCO2 corrected to a body temperature of 27°C would be about 7.55 and 25. Conversely, if the pH and PaCO2 as measured at 37°C are 7.25 and 55, then the corrected values at 27°C would be about 7.4 and 40.

          At issue are the appropriate temperature-corrected pH and PaCO2 values during hypothermia. One method (pH-stat) attempts to keep the temperature-corrected pH and PaCO2 at 7.4 and 40, respectively. The other method (α-stat) attempts to keep the ratio of OH–/H+ ions constant so that enzyme systems function appropriately. This will be accomplished if the uncorrected pH and PaCO2 as measured at 37°C are 7.4 and 40 mm Hg, respectively. The rationale for these two regimens is discussed in further detail in Chapters 814, and 24.

V. Normothermia and CPB

   A. Potential advantages of normothermic CPB. By the late 1960s, hypothermia became a near-ubiquitous practice for patients undergoing CPB. Recently, there has been a trend to conduct CPB at near-normothermic levels [33]. The putative advantages include avoiding adverse effects of hypothermia (increased viscosity, reduced microcirculatory flow, leftward shift of the oxygen–hemoglobin dissociation curve), decreased duration of bypass (time spent cooling and rewarming not required), and avoiding the hazards of overheating (especially of the brain).

   B. Potential disadvantages of normothermic CPB. Normothermic bypass narrows the ratio of oxygen demand (VO2) to oxygen delivery (DO2). This suggests the need to maintain hematocrit and pump flow rates higher than accepted during hypothermic CPB. However, minimal safe pump flow rates and hematocrit values during normothermic CPB have not been determined. SVR and MAP also tend to be lower; more fluids and vasoconstrictors and higher pump flow rates are typically used at higher CPB temperatures.

   C. Clinical outcomes following normothermic CPB. Neurologic injury is a primary concern of conducting bypass at normothermia. The two largest randomized studies examining the effect of temperature management on neurologic outcomes reached conflicting conclusions; in patients undergoing warm CPB, the incidence of stroke was reported to be increased or not different compared to patients undergoing hypothermic bypass [34,35]. Normothermic CPB was introduced primarily to improve myocardial protection, and some evidence suggests that myocardial injury is reduced and myocardial function is improved following bypass at warmer temperatures [6]. Renal or hematologic function does not appear to be improved by normothermic CPB [6]. At the present time, many cardiac centers are using mild degrees of hypothermia (about 35°C), the so-called “tepid” bypass, which may offer substantial cerebral protection without the disadvantages of deeper levels of hypothermia. Current evidence does not clearly support one temperature management strategy over another, and it is likely that the ideal temperature for CPB varies with the physiologic requirements of the patient and surgery.

VI. Systemic effects of the CPB

CPB is a highly unphysiologic experience that triggers an “explosion” of adverse events (Fig. 21.10).

Figure 21.10 The “explosion” of adverse events triggered by cardiopulmonary bypass. (From Eleferiades JA. Mini-CABG. A step forward or backward? The “pro” point of view. J Cardiothorac Vasc Anesth. 1997;11:661, with permission.)

   A. Causes and contributors of adverse systemic effects of CPB:

      1. Microemboli (gas and particulate matter)

      2. Activation of the inflammatory and coagulation systems

      3. Altered temperature, cooling and warming

      4. Exposure of blood to foreign surfaces

      5. Reinfusion of shed blood and transfusion of blood products

      6. Hemodynamic alterations (abnormal flow rate and pattern, abnormal arterial and venous pressures)

      7. Ischemia and reperfusion (especially of heart, lungs, and gut)

      8. Hyperoxia

      9. Hemodilution (with anemia and reduced oncotic pressure)

   B. Blood

     1. Coagulation and fibrinolytic systems and tissue factor (TF). Changes in the coagulation cascade, platelets, and the fibrinolytic cascade are discussed in Chapter 19.

      2. Changes in formed elements

        a. RBCs

           (1) RBCs become stiffer and less deformable during CPB, which may interfere with microcirculatory blood flow and increase susceptibility to hemolysis.

           (2) During CPB, RBCs are exposed to nonphysiologic surfaces and shear stresses which may cause their destruction. The degree of hemolysis is increased by both higher flow rates and the accompanying increase in rate of shear, and by gas–fluid interfaces in the ECC. As red cells are lysed, the free hemoglobin produced is bound to haptoglobin. When the amount of free hemoglobin generated exceeds the binding capacity of haptoglobin, serum hemoglobin concentrations increase and hemoglobin is filtered by the kidney, resulting in hemoglobinuria. Cardiotomy suction is a major contributor to hemolysis during CPB.

        b. Leukocytes. CPB affects primarily neutrophils (polymorphonuclear leukocytes [PMNs]) and, to a lesser degree, monocytes. Shortly after the onset of CPB there is a marked decrease in circulating PMNs. This is due to sequestration in the pulmonary circulation and intravascular and extravascular accumulation in the microcirculation of heart and skeletal muscle. Blockage of vessels by PMNs or microcirculatory derangements induced by substances released from PMNs may contribute to organ dysfunction after CPB. Circulating PMN levels increase dramatically with rewarming. Neutrophils released from the pulmonary circulation and younger cells released from the bone marrow contribute to the observed neutrophilia.

            Effects of CPB on host defense functions of PMNs are controversial. Studies demonstrating decreased responsiveness of PMNs to chemotactic and aggregating stimuli indicate impaired defense mechanisms. However, other studies show that the bacteriocidal activity of PMNs is increased for up to 3 days after CPB.

     3. Changes in plasma proteins

        a. Denaturation. Proteins are molecules with highly specific structures. When proteins approach a gas–liquid interface, strong electrostatic forces at that interface produce varying degrees of molecular unfolding by disrupting internal bonds (i.e., denaturation). Some of the consequences of this protein denaturation include the following:

           (1) Altered enzymatic function. Denatured proteins lose some or all of their function. This may be one mechanism by which coagulation becomes impaired during and after CPB.

           (2) Aggregation of proteins. Denatured proteins have a tendency to aggregate and produce precipitates.

           (3) Altered solubility characteristics. Denatured proteins are less soluble in plasma and cause increased blood viscosity.

           (4) Release of lipids. Denaturation of lipoproteins and the protein fractions of chylomicrons may result in aggregates which occlude small vessels.

           (5) Absorption of denatured proteins onto RBC membranes may cause them to become “sticky.” The resulting RBC aggregates promote capillary sludging and microcirculatory dysfunction.

        b. Reduced colloid osmotic (oncotic) pressure (COP). Because of the hemodilution associated with use of asanguineous priming solutions, plasma protein concentration and hence COP fall with onset of CPB if no colloids are added to the CPB circuit. There is controversy about the need and benefits of avoiding the fall in COP by use of albumin or artificial colloids (e.g., dextrans, starches) in the prime.

     4. Activation of humoral cascade systems—see later.

   C. Fluid balance and interstitial fluid accumulation during CPB. The following equation, based upon Starling’s hypothesis, is thought to describe the fluid fluxes at the microcirculatory level:

Tissue fluid accumulation = K[(Pc - Pis) - δ(pc - pis)] - Qlymph

     where K is the filtration coefficient (“permeability”) of capillary membrane, Pc the mean intracapillary hydrostatic pressure, Pis the mean interstitial hydrostatic pressure, δ the reflection coefficient to macromolecules, πc the intracapillary oncotic pressure, π is the interstitial oncotic pressure, and Qlymph the lymph flow out of interstitium.

   CPB shifts this balance toward accumulation of interstitial fluid by affecting several of these variables. Membrane permeability is increased by activation of the SIR and intermittent ischemia/reperfusion. Plasma oncotic pressure falls due to the use of asanguineous priming fluids. Inadequate venous drainage may increase mean capillary hydrostatic pressure, whereas immobility, lack of pulsatile flow, and loss of negative intrathoracic pressure impede lymphatic flow.

   D. Heart. Some degree of myocardial injury and cell necrosis occurs during CPB which may result in myocardial stunning and dysfunction. However, frank myocardial infarction is relatively uncommon (although the ECG and cardiac enzyme changes which identify myocardial infarction have not been precisely defined).

   Factors that contribute to myocardial injury include those that affect microcapillary perfusion in general, but also ventricular distention, prolonged ventricular fibrillation, coronary air embolization, hypotension, catecholamines, endotoxemia, and ischemia/reperfusion associated with aortic cross-clamping. It is thought that higher perfusion pressures are desirable to maintain adequate myocardial perfusion in patients with cardiac hypertrophy or severe coronary artery disease during those periods of CPB when the ascending aorta is unclamped. Potent volatile anesthetics appear to be protective of the heart. (See also Chapter 23.)

   E. CNS. Cerebral dysfunction (ranging from subtle neurocognitive dysfunction to frank stroke or coma) is not infrequent following CPB. Its etiology is multifactorial and includes hypoperfusion, macroemboli, microemboli, and the inflammatory response to CPB. The causes of cerebral dysfunction and strategies to minimize adverse cerebral outcomes are discussed further in Chapter 24.

   F. Kidneys. Renal dysfunction after CPB, which ranges from a rise in creatinine and release of renal tubular proteins to frank renal failure requiring renal replacement therapy (e.g., hemodialysis), is a persistent cause of morbidity and mortality in cardiac surgical patients.

     1. Significance of urine output during CPB. Urine output is a crude indicator of renal function, but there is no correlation between the amount of urine output during CPB and the incidence of postoperative renal failure. Urine output is greater when MAP is higher, when pulsatile perfusion is used, and when mannitol is added to pump-priming fluids.

     2. Decreased glomerular and tubular function. Evidence of temporary glomerular and tubular dysfunction is present in many patients early post-CPB [36]. Tubular function is depressed by hypothermia and by reductions in renal blood flow.

     3. Renal blood flow. Global renal blood flow usually decreases during CPB as a result of diminished flow rates and pressures or loss of pulsatility. As in other low-flow states, there is a redistribution of renal blood flow from the cortex to the outer medulla. This redistribution of blood flow appears to be less severe during pulsatile perfusion.

     4. Hemoglobinuria. Intravascular hemolysis and hemoglobinuria can cause acute tubular necrosis. The mechanism is precipitation of pigment in the renal tubules with subsequent blockage of tubular flow or glomerular–tubular injuries caused by red cell stroma and other substances liberated from lysed RBCs.

     5. Renal failure. The reported incidence of acute renal failure requiring renal replacement therapy ranges from 1% to 3%. Development of renal failure appears to depend more on the preoperative renal function and postoperative hemodynamic status than on various manipulations used to maintain urine output during CPB [36]. However, Ranucci et al. [24] observed that a critically low oxygen delivery during CPB was an important predictor of acute renal failure after coronary artery surgery.

   G. Splanchnic, visceral, and hepatic effects. The incidence of major gastrointestinal (GI) complications (e.g., bleeding, ulcers, mesenteric ischemia or infarction, cholecystitis, pancreatitis), postoperatively, is low (1% to 2%), but is associated with high mortality rates (36% to 65%). Risk factors include advanced age, open ventricle operations, emergency procedures, prolonged bypass times, use of vasopressors, and postbypass low cardiac output syndrome [37].

   Although global splanchnic blood flow (and hepatic venous oxygen saturation) appears to be preserved during CPB, the hierarchy of regional blood flow suggests that splanchnic flow will be compromised early whenever systemic flow is reduced during CPB [20,21] (and is likely reduced further with administration of vasoconstrictors, such as phenylephrine, norepinephrine, and AVP). Furthermore, many subjects exhibit increased intestinal permeability, decreased gastric or intestinal mucosal pH and increased mucosal PCO2 (by tonometry), decreased mucosal blood flow (by Doppler flowmetry), and endotoxemia, all of which suggest that mucosal ischemia occurs frequently during CPB. GI ischemia may play a primary role in the development of SIR syndrome and other GI complications.

   As with renal failure, hepatic dysfunction is more dependent on hemodynamic status before and after CPB than on any direct effect of CPB. Jaundice may occur in up to 23% of patients after cardiac surgery, but severe jaundice (bilirubin levels at least 6 mg/dL) occurs in only 6% of patients. The risk of postoperative jaundice is increased in the setting of high RA pressures, persistent hypotension after CPB, or significant transfusion.

   H. Lungs. Pulmonary dysfunction after CPB may range from mild decreases in PaO2, related to the nearly ubiquitous postoperative atelectasis, to full-blown respiratory failure resembling the adult respiratory distress syndrome (ARDS).

     1. Pulmonary dead space and ventilation–perfusion mismatching after CPB. Extravascular lung water increases during CPB. Development of intrapulmonary shunts and increased dead space ventilation result in less efficient matching of ventilation to perfusion. Increased dead space is reflected in greater end-tidal–arterial CO2 gradients after CPB in the majority of patients. Ventilation–perfusion (V/Q) mismatching results in increased alveolar-arterial O2 gradients and decreased PaO2. These are the most consistent abnormalities noted after CPB.

     2. Pulmonary sequestration of neutrophils and release of vasoactive compounds. PMNs sequestered in the lung during CPB may undergo release reactions causing localized, intense vasoconstriction or membrane damage with subsequent edema formation, resulting in increased dead space ventilation and V/Q mismatching.

     3. Changes in pulmonary vascular resistance (PVR) and hypoxic pulmonary vaso- constriction

        a. PVR. Numerous factors influence pulmonary vascular tone after CPB. Many congenital cardiac lesions are associated with increased PVR. Catecholamine and endothelin levels rise during and after CPB, which may also contribute to increases in PVR. Efforts to limit increases in PVR include hyperventilation and administration of sodium bicarbonate to maintain an alkaline pH, maintenance of high inspired oxygen concentrations, and avoidance of catecholamine infusions that can induce pulmonary vasoconstriction.

        b. Hypoxic pulmonary vasoconstriction. Direct effects of CPB and hypothermia on hypoxic pulmonary vasoconstriction are poorly understood. However, use of volatile anesthetics and vasodilators after CPB may interfere with hypoxic pulmonary vasoconstriction, leading to V/Q mismatching and decreasing PaO2.

     4. Postpump pulmonary dysfunction. Most patients exhibit some immediate decrease in PaO2 postoperatively, and it is difficult to predict which patients will go on to develop more serious pulmonary insufficiency. Although full-blown respiratory failure after CPB is now relatively rare, its incidence is directly related to preoperative pulmonary dysfunction, duration of CPB, and postoperative hemodynamic status. Events during CPB that may contribute to development of pulmonary dysfunction include the following:

        a. Decreased pulmonary blood flow resulting from the following:

           (1) Emboli of various compositions leading to localized areas of ventilation/perfusion (V/Q) mismatching

           (2) Localized vasoconstriction due to elevated endogenous and exogenous catecholamines, endothelin, or substances released from PMNs trapped in the pulmonary capillaries

        b. Membrane damage resulting in increased capillary permeability and edema formation from the following:

           (1) Complement activation

           (2) Vasoactive compounds released from PMNs

           (3) Oxygen-free radicals

        c. Edema formation from increased pulmonary hydrostatic pressure caused by inadequate LV venting and increased bronchial blood flow

        d. Ischemia/reperfusion of the lungs and remote organs

        e. Atelectasis from lung deflation during CPB, which occurs in up to 60% of patients

        f. SIR

     5. Methods to reduce lung injury associated with CPB

        a. Optimal lung management during CPB. Most data suggest that ventilation with 100% oxygen during CPB is detrimental to the lungs. The majority of studies suggest that application of 5 cm of water PEEP and/or continuous ventilation utilizing low tidal volumes during CPB does not improve pulmonary function or gas exchange following CPB. In contrast, the use of recruitment maneuvers at the end of CPB before resuming ventilation does appear to improve post-CPB oxygenation, particularly if PEEP is used after the recruitment maneuver.

        b. Prophylactic administration of anti-inflammatory drugs such as aprotinin and corticosteroids to help preserve lung function during CPB has been advocated by some investigators. Although both drugs reduce the SIR to CPB, the effect of these agents on postoperative gas exchange and pulmonary complications remains controversial.

   I. Inflammation. All patients undergoing cardiac surgery will experience a systemic proinflammatory response to the procedure. Although the trauma of surgery itself leads to a degree of inflammation, the CPB circuit seems to accentuate this inflammatory response.

   This syndrome represents an unphysiologic activation of the innate immune system resulting in a whole-body SIR resembling that associated with sepsis and trauma [38]. It presents as a spectrum of responses ranging from near-universal evidence of mild inflammation (fever, leukocytosis), to more significant clinical signs (tachycardia, increased cardiac output, decreased SVR, increased oxygen consumption, increased capillary permeability) to frank organ dysfunction (cardiac, renal, pulmonary, GI, hepatic, CNS) and finally to the multiple organ dysfunction syndrome (MODS) and death. There are several possible variables that may explain why the SIR differs significantly between patients. These factors include pre-existing medical conditions, extremes of age, the extent and duration of the surgery, and genomic makeup of the patient.

   Normal inflammation is a localized protective response that is composed of cellular as well as humeral components. When the localized inflammatory response is excessive, it may spill over to the rest of the body; the same appears to be true when the injurious agent is systemic (e.g., CPB), and thus the inflammatory response becomes generalized leading to diffuse and remote end-organ damage [3942].

     1. Activation. Nonspecific activators of the inflammatory response include surgical trauma, blood loss or transfusion, and hypothermia. CPB may also activate the inflammatory response by three distinct mechanisms, which include the following:

        a. Contact activation. When blood comes into contact with the CPB circuit, plasma proteins are adsorbed to the surface of the circuit. Platelets then adhere to the surface of this protein layer. Contact activation results in the activation of the complement, coagulation, and fibrinolytic systems and the kallikrein–bradykinin cascades.

            Another cause of contact activation is the use of cardiotomy suction. Blood aspirated with the cardiotomy suction is contaminated with TF, tissue and fat fragments, free plasma hemoglobin, thrombin, TF activator, and fibrin degradation products. These contaminants as well as a blood–air interface and turbulent flow lead to activation of complement and coagulation systems.

        b. Ischemia–reperfusion. Reperfusion injury refers to damage to tissue caused when blood supply returns to the tissue after a period of ischemia. The absence of oxygen and nutrients from blood creates a condition in which the restoration of circulation results in inflammation and oxidative damage from the oxygen rather than re-establishment of normal function. The reintroduction of oxygen with restored blood flow results in formation of oxygen-free radicals that damage cellular proteins, DNA, and plasma membranes. WBCs carried to the area release a host of inflammatory substances such as interleukins. In addition, leukocytes may also build up in small capillaries leading to obstruction and more ischemia.

        c. Endotoxemia. Endotoxin is a lipopolysaccharide from the cell wall of gram-negative bacteria. Endotoxin binds with lipopolysaccharide-binding protein and this complex stimulates the release of tumor necrosis factor (TNF) by macrophages, potentially triggering the development of SIRS. During CPB there is transient endotoxemia. This phenomenon is thought to result from splanchnic hypoperfusion, which causes damage to the GI mucosa and translocation of endotoxin.

     2. Propagation. Once the inflammatory response is triggered, components of the immune and coagulation systems are activated (as well as the endothelium), which propagate the SIR. These components include the following:

        a. Complement. The complement system is a cytotoxic host immune defense system composed of a cascade of 30 or so proteins that interact to recognize and kill pathogens. Activation is triggered by one of three pathways—classical, alternative, and lectin. Mechanisms of activation and propagation include the following:

           (1) Exposure of blood to the foreign surface and the blood–gas interface of the CPB circuit activates the alternative pathway.

           (2) Endotoxin from the GI tract during CPB activates both the alternative and classical pathways.

           (3) Heparin–protamine complexes in the post-CPB period activate the classical pathway.

            The result of complement activation is the release of several active substances including the following:

             (a) The anaphylatoxins C3a and C5a, which stimulate leukocyte activation and chemotaxis, increase production of proinflammatory cytokines, oxygen-derived free radicals, proteolytic enzymes, and leukotrienes, and increase capillary permeability.

             (b) iC3b, which binds to the CPB circuit and the vascular endothelium and acts as a ligand for leukocyte adhesion

             (c) C5b-9, which causes cell lysis, increases leukocyte adhesion molecule expression, and stimulates proinflammatory cytokine production

The degree of complement activation in patients undergoing CPB appears to correlate with the extent of postoperative complications.

        b. The release of other mediators

           (1) Cytokines

Cytokines are small proteins and polypeptides that mediate and regulate immunity, inflammation, and hematopoiesis. They are released by activated monocytes, tissue macrophages, lymphocytes, and endothelial cells. In addition, blood products and reinfused shed blood contain significant concentrations of cytokines.

   Cytokines may exert proinflammatory or anti-inflammatory effects. Proinflammatory cytokines are produced predominantly by activated macrophages and are involved in the upregulation of inflammatory reactions. These include TNF-α, interleukin (IL)-1, IL-2, IL-6, IL-8, and interferon γ. Elevations of proinflammatory cytokine levels post-CPB have been associated with adverse outcome following cardiac surgery.

   Anti-inflammatory cytokines belong to the T-cell–derived cytokines and are involved in the downregulation of inflammatory reactions. Those released during CPB include IL-10, IL-1 receptor antagonist (IL-1ra), and transforming growth factor-β (TGF-β) [41].

           (2) Nitric oxide. NO is produced by the vascular endothelium by the conversion of L-arginine to L-citrulline by constitutive NO synthase (cNOS). The effects of NO include regulation of vasomotor tone, inhibition of platelet and neutrophil aggregation, and immunomodulation. CPB leads to the upregulation and release of inducible NO synthase (iNOS), potentially resulting in post-CPB vasodilation, vascular permeability, and end-organ dysfunction [38].

           (3) Leukotrienes. Leukotrienes are released during CPB and act as chemokines and potent vasoconstrictors of smooth muscle. They may also play a role in neutrophil-related injury, ARDS, and MOF.

           (4) Platelet-activating factor (PAF). PAF is an endogenous phospholipid intracellular signaling messenger as well as an inflammatory and neurotoxic agent. Its neurotoxic actions include increasing intracellular calcium concentrations, disrupting the blood–brain barrier, and reducing cerebral blood flow.

           (5) Tissue factor. During CPB, TF is expressed by many cells including the endothelium and blood monocytes. TF-initiated blood coagulation through the extrinsic pathway leads to the release of a myriad of proinflammatory cytokines.

           (6) Kallikrein–bradykinin. Kallikrein plays a seminal role in activation and amplification of the inflammatory response. Bradykinin increases vascular permeability and release of tissue plasminogen activator.

           (7) Collagenases, gelatinases, and metalloproteases

           (8) Endothelins. These are a family of peptides, the most important of which is endothelin-1, which are produced by endothelial cells (as well as smooth muscle and cardiac myocytes). Endothelin-1 is a potent vasoconstrictor that has been associated with pulmonary and systemic hypertension and myocardial dysfunction.

        c. The endothelium. The endothelium is an active participant in a variety of physiologic and pathologic processes. It plays a major role in regulating vascular tone, membrane permeability, coagulation and thrombosis, fibrinolysis, and inflammation. It attracts and directs the passage of leukocytes into areas of inflammation through the expression of adhesion molecules [42].

            CPB causes extensive activation and dysfunction of the endothelium due to ischemia/reperfusion, exposure to inflammatory mediators, surgical manipulation, and hemodynamic shear stresses. Expression of adhesion molecules mediates the binding of neutrophils to the endothelium and translocation into the interstitium. Resultant neutrophil degranulation worsens the state of the endothelium further, leading to capillary leak and edema. Proinflammatory cytokines inhibit the production of NO, shifting the balance from vasodilation to vasoconstriction. Endothelial cell activation also shifts the endothelial anticoagulant phenotype to a procoagulant phenotype.

        d. Coagulation and fibrinolysis. The coagulation–fibrinolysis cascade is closely intertwined with the inflammatory response. The activation of one system plays a key role in the other.

        e. The cellular immune response. Leukocytes are made up of granulocytes, monocytes, and lymphocytes. Granulocytes include neutrophils, eosinophils, and basophils. Monocytes become tissue macrophages after they migrate into tissue. Lymphocytes are made up of B- and T-cells and natural killer cells.

            CPB results in the activation of leukocytes, leukocyte-endothelial cell adhesion, and transmigration. The activation may be mediated by contact activation through the CPB circuit and via a variety of inflammatory mediators.

           (1) Neutrophils. Interaction of neutrophils with the endothelium takes place in distinct steps:

             (a) Neutrophil “rolling” on the endothelium. This process is thought to be mediated by P-selectin, an adhesion molecule, which is expressed on both endothelial cells and neutrophils during CPB.

             (b) Interaction of more adhesion molecules on the neutrophils and endothelium causes a firm neutrophil adherence: β 2 integrins (e.g., CD11a/CD18, CD11b/CD18, ICAM-1).

             (c) Migration of the neutrophils into the interstitial space

   Activated neutrophils can cause tissue damage and end-organ dysfunction via microvascular occlusion or release toxic metabolites and enzymes.

           (2) Monocytes and macrophages. Monocytes phagocytize microorganisms and cell fragments, produce and secrete chemical mediators, and produce cytotoxins. The cytokine monocytes release play an important role in directing neutrophils and monocytes to sites of inflammation.

           (3) Lymphocytes. B and T lymphocytes decrease in number and function following CPB, which may render the patient immunosuppressed and at risk for infections.

           (4) Platelets. Through the production and release of leukotrienes, serotonin, chemokines, PF4, and other substances, platelets contribute to the inflammatory response.

           (5) Endothelium. See above.

     3. Consequences of the inflammatory response to CPB. In most patients, the early SIR resolves without significant injury as a result of discontinuation of the stimulus, dissipation of mediators, or the action of naturally occurring antagonists (e.g., IL-10). The variability in the expression, consequences, and outcome of the inflammatory response to CPB is the subject of much speculation. Some factors which influence this include the preoperative condition of the patient, type and complexity of the surgery, and, perhaps most importantly, underlying genetic polymorphism.

     4. Strategies to minimize the inflammatory response to CPB. Various strategies have been promoted to minimize the SIR syndrome, including surface modification of the ECC (e.g., heparin-coated circuits), use of centrifugal pumps, pulsatile flow, and MOs, priming with colloids, ultrafiltration, leukodepletion or leukofiltration, use of minimized circuits (reduced surface area and elimination of the venous reservoir), improved management of shed blood (e.g., elimination of or minimizing use of cardiotomy suction, use of RBC-washer/salvage devices), off-pump techniques, administration of corticosteroids or aprotinin, complement inhibitors, protease inhibitors (e.g., aprotinin), anticytokine therapy, antiendotoxin therapy, digestive decontamination, and administration of antibodies to block adhesion molecules or their activation [38,41,43]. Although many of these strategies have been effective in attenuating the inflammatory response to CPB, few have been proven to reduce major complications in the postoperative period.

   J. “Stress response” to CPB—Endocrine, metabolic, and electrolyte effects. CPB is associated with a marked exaggeration of the stress response associated with all types of surgery. This is manifested by large increases in epinephrine, norepinephrine, AVP (or antidiuretic hormone), adrenocorticotropic hormone, cortisol (mainly after bypass), growth hormone, and glucagon. Elevated catecholamines may have adverse effects on regional and organ blood flow patterns. Catecholamines also increase myocardial oxygen consumption, which may adversely affect the balance of myocardial oxygen supply and demand at the time of reperfusion. Other stress hormones also increase catabolic reactions, leading to increased energy consumption, tissue breakdown, and possible impairment of wound healing.

   Hyperglycemia is invariably encountered during CPB, especially in patients with diabetes mellitus. Contributors to hyperglycemia include decreased insulin production, insulin resistance (possibly related to stress hormones), decreased consumption (related to insulin resistance and hypothermia), increased glycogenolysis and gluconeogenesis (related to stress hormones), and increased reabsorption of glucose by the kidney. Observational studies have demonstrated an association between post-CPB hyperglycemia and increased morbidity and mortality. Although control of blood glucose values during and following CPB is currently strongly advocated, the optimal “target range” for serum glucose concentrations has not been determined.

   Renin, angiotensin II, and aldosterone levels all tend to rise during CPB. Many patients display the so-called sick euthyroid syndrome with reduced tri-iodothyronine (T3), thyroxine (T4), and free thyroxin levels but normal thyroid-stimulating hormone levels. The etiology of this phenomenon is unclear, but it provides the rationale for administration of thyroid hormone in some patients with low cardiac output syndrome.

   Calcium and magnesium. Both ionized calcium and total and unfiltratable fractions of magnesium commonly fall, whereas potassium levels may fluctuate widely during CPB. The latter may be related to diuretics, catecholamines, preoperative spironolactone (aldactone) and β -blockers, potassium-containing cardioplegia, and renal dysfunction. The importance of maintaining normal levels of these ions to preserve normal muscle and cardiac function and prevent dysrhythmias is apparent.

VII. Summary

CPB is performed safely and effectively throughout the world. This is a result of a combination of sophisticated equipment used for extracorporeal circulation and well-trained and educated perfusionists. The responsibility for safe CPB is shared by surgeons, anesthesiologists, and perfusionists in order to manage cardiovascular surgery with the lowest possible patient risk. Despite technologic advances in circuit design that have occurred over the past several decades, aberrations of normal physiology are imposed by CPB, which may result in postoperative organ dysfunction. Post-CPB organ dysfunction constitutes a spectrum ranging from mild dysfunction in one or more organ systems to death resulting from multiorgan failure. It should be emphasized that placing a patient on CPB is a physiologic trespass against that patient. Absence of significant damage caused by CPB depends primarily on a particular patient’s ability to compensate for the derangements introduced by that trespass.


 1. Ghosh S, Falter F, Cook DJ. Cardiopulmonary Bypass. Cambridge, MA: Cambridge University Press; 2009.

 2. Hessel EA. Cardiopulmonary bypass equipment. In: Estafanous FG, et al., eds. Cardiac Anesthesia. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2001.

 3. Schell RM, Hessel, EA II, Reves JG. Cardiopulmonary bypass. Chapter 10. In: Reves JG, Reeves S, Abernathy JH III, eds. Atlas of Cardiothoracic Anesthesia. 2nd ed. New York, NY: Springer; 2009.

 4. Groom RC, Stammers AH. Extracorporeal devices and related technologies. In: Kaplan JA, et al., eds. Kaplan’s Cardiac Anesthesia. 6th ed. St. Louis, MO: Elsevier/Saunders; 2011.

 5. Hessel EA II. Circuitry and cannulation techniques. In: Gravlee GP, Davis RF, Stammers AH, et al., eds. Cardiopulmonary Bypass: Principles and Practice. 3rd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2008.

 6. Murphy GS, Hessel EA, Groom RC. Optimal perfusion during cardiopulmonary bypass: An evidence-based approach. Anesth Analg. 2009;108:1394–1417.

 7. Groom RC, Froebe S, Martin J, et al. Update on pediatric perfusion practice in North America: 2005 survey. J Extra Corpor Technol. 2005;37:343–350.

 8. Zangrillo A, Garozzo FA, Biondi-Zoccai G, et al. Miniaturized cardiopulmonary bypass improves short-term outcome in cardiac surgery: A meta-analysis of randomized controlled studies. J Thorac Cardiovasc Surg.2010;139:1162–1169.

 9. Hessel EA. Bypass techniques for descending thoracic aortic surgery. Semin Cardiothoracic Vasc Anesth. 2001;5:293–320.

10. Charrière JM, Pélissié J, Verd C, et al. Survey: Retrospective survey of monitoring/safety devices and incidents of cardiopulmonary bypass for cardiac surgery in France. J Extra Corpor Technol.2007;39:142–157.

11. Jenkins OF, Morris R, Simpson JM. Australasian perfusion incident survey. Perfusion. 1997;12:279–288.

12. Mejak BL, Stammers A, Rauch E, et al. A retrospective study on perfusion incidents and safety devices. Perfusion. 2000;15: 51–61.

13. Groom RC, Forest RJ, Cormack JE, et al. Parallel replacement of the oxygenator that is not transferring oxygen: The PRONTO procedure. Perfusion. 2002;17:447–450.

14. Edmonds HL Jr. 2010 Standard of care for central nervous system monitoring during cardiac surgery. J Cardiothorac Vasc Anesth. 2010;24:541–543.

15. Gravlee GP, Davis RF, Kurusz M, eds. Cardiopulmonary Bypass: Principles and Practice. 3rd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2007.

16. Kirklin JW, Barratt-Boyes BG. Hypothermia, circulatory arrest, and cardiopulmonary bypass. In: Kirklin JW, Barratt-Boyes BG, eds. Cardiac Surgery. 2nd ed. New York, NY: Churchill Livingstone; 1993:61–127.

17. Grovier AV, Reves JG, McKay RD, et al. Factors and their influence on regional cerebral blood flow during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg. 1984;38:592–600.

18. Joshi B, Ono M, Brown C, et al. Predicting the limits of cerebral autoregulation during cardiopulmonary bypass. Anesth Analg. 2012;114(3):503–510.

19. Gold JP, Charlson ME, Williams-Russo P, et al. Improvement of outcomes after coronary artery bypass. A randomized trial comparing intraoperative high versus low mean arterial pressure. J Thorac Cardiovasc Surg.1995;110:1302–1311.

20. Rudy LW, Heymann MA, Edmunds LH. Distribution of systemic blood flow during cardiopulmonary bypass. J Appl Physiol. 1973;34:194–200.

21. Slater JM, Orszulak TA, Cook DJ. Distribution and hierarchy of regional blood flow during hypothermic cardiopulmonary bypass. Ann Thorac Surg. 2001;72:542–547.

22. Wright G. Mechanical simulation of cardiac function by means of pulsatile blood pumps. J Cardiothorac Vasc Anesth. 1997;11:299–309.

23. Cavaliere F, Gennari A, Martinelli L, et al. The relationship between systemic oxygen uptake and delivery during moderate hypothermic cardiopulmonary bypass: Critical values and effects of vasodilation by hydralazine. Perfusion. 1995;10:315–321.

24. Ranucci M, Romitti F, Isgro G, et al. Oxygen delivery during cardiopulmonary bypass and acute renal failure after coronary operations. Ann Thorac Surg. 2005;80:2213–2220.

25. Schmidt FX, Philipp A, Foltan M, et al. Adequacy of perfusion during hypothermia: Regional distribution of cardiopulmonary bypass flow, mixed venous and regional venous oxygen saturation. Thorac Cardiovasc Surg.2003;51:306–311.

26. Ranucci M, Isgro G, Romitti F, et al. Anaerobic metabolism during cardiopulmonary bypass: Predictive value of carbon dioxide derived parameters. Ann Thorac Surg. 2006;81:2189–2195.

27. Murkin JM, Adams SJ, Novick RJ, et al. Monitoring brain oxygen saturation during coronary bypass surgery: A randomized, prospective study. Anesth Analg. 2007;104(1):51–58.

28. Taillefer M-C, Denault AY. Cerebral near-infrared spectroscopy in adult heart surgery: Systematic review of its clinical efficacy. Can J Anaesth. 2005;52:79–87.

29. Hessel EA. CNS monitoring: The current weak state of the evidence. J Cardiothorac Vasc Anesth. 2011;25(4):e15 [Online only article].

30. Edmunds HL, Jr. Reply to Dr. Hessel. J Cardiothorac Vasc Anesth. 2011;25(4):e16.

31. Wagener G, Jan M, Kim M, et al. Association between increases in urinary neutrophil gelatinase-associated lipocalin and acute renal dysfunction after adult cardiac surgery. Anesthesiology.2006;103:485–491.

32. Habib RH, Zacharias A, Schwann TA, et al. Adverse effects of low hematocrit during cardiopulmonary bypass in the adult: Should current practice be changed? J Thorac Cardiovasc Surg.2003;125:1438–1450.

33. Cook DJ. Changing temperature management for cardiopulmonary bypass. Anesth Analg. 1999;88:1254–1271.

34. The Warm Heart Investigators. Randomized trial of normothermic versus hypothermic coronary bypass surgery. Lancet. 1994;343:559–563.

35. Martin TD, Craver JM, Gott JP, et al. Prospective, randomized trial of retrograde warm blood cardioplegia: Myocardial benefit and neurological threat. Ann Thorac Surg. 1994;57:298–304.

36. Aronson S, Blumenthal R. Perioperative renal dysfunction and cardiovascular anesthesia: Concerns and controversies. J Cardiothorac Vasc Anesth. 1998;12:567–586.

37. Hessel EA II. Abdominal organ injury after cardiac surgery. Semin Cardiothorac Vasc Anesth. 2004;8:243–263.

38. Laffey JG, Boylan JF, Cheng DCH. The systemic inflammatory response to cardiac surgery. Implications for the anesthesiologist. Anesthesiology. 2002;97:215–252.

39. Pintar T, Collard CD. The systemic inflammatory response to cardiopulmonary bypass. Anesthesiol Clin North America. 2003;21:453–464.

40. Menasche P, Edmunds LH Jr. The inflammatory response. In: Cohn LH, Edmunds LH, eds. Cardiac Surgery in the Adult Patient. New York, NY: McGraw-Hill Professional; 2003:349–360.

41. Bennett-Guerrero E. Systemic inflammation. In: Kaplan JA, Reich DL, Savino JS, eds. Kaplan’s Cardiac Anesthesia. 6th ed. Philadelphia, PA: Elsevier/Saunders; 2011:178–192.

42. O’Brien ERM, Nathan HJ. Coronary physiology and atherosclerosis. In: Kaplan JA, Reich DL, Lake CL, et al., eds. Kaplan’s Cardiac Anesthesia. 5th ed. Philadelphia, PA: Elsevier/Saunders; 2006:94–96.

43. Asimakopoulos G, Gourlay T. A review of anti-inflammatory strategies in cardiac surgery. Perfusion. 2003;18:7–12.