Cardiopulmonary Bypass: Principles and Practice

Previous Chapter | Next Chapter >




Quick Links to Sections in this Chapter


















E. A. Hessel: Department of Anesthesiology, University of Kentucky School of Medicine, Chandler Medical Center, Lexington, KY 40536.

A. G. Hill: Inova Fairfax Hospital, Falls Church, VA 22042.


Back to Quick Links

The primary function of cardiopulmonary bypass (CPB) is to divert blood away from the heart (both the right and left side and usually the lungs as well) and return it to the systemic arterial system, thus allowing cardiac surgery. Therefore, it must replace the function of both the lung (gas exchange) and heart (provide energy to ensure circulation of blood). Typically, blood is drained by gravity via cannulas in the superior and inferior vena cavae (SVC, IVC) or IVC and right atrium (RA) (cavoatrial position) to the heart–lung machine where it is pumped (with a roller or centrifugal pump) through the artificial lung (most often a membrane-type "oxygenator") back into the systemic vasculature via an arterial cannula placed in the ascending aorta. In the past, when bubble oxygenators were used, the pump was placed after the oxygenator, drawing arterialized blood from a reservoir.

Because of the need to offset the cooling during the extracorporeal passage of blood and the frequent desire to intentionally cool and then rewarm the patient, a heat exchanger is included as part of the oxygenator, either before or contiguous with the gas exchange unit.

Peripheral cannulation, using the femoral or other veins and arteries, is occasionally used electively for cardiac surgery when central cannulation is not technically possible, for initiating bypass before opening the chest, for emergent situations, for aortic surgery, and for extracorporeal membrane oxygenation. Left heart bypass or proximal aorta bypass (with "venous cannulation" of the left atrium, left ventricle, or proximal aorta) and distal infusion into the distal aorta or femoral artery, incorporating only an extracorporeal pump (commonly a centrifugal pump with minimal anticoagulation), is sometimes used for aortic surgery.

Besides the major venous and arterial connections and the oxygenator, heat exchanger, and pump, there are many other components to the heart–lung machine (Fig. 5.1). An adjustable clamp or remote venous line occluder regulates the main venous drainage line and a separate tubing clamp is used on the systemic flow line whenever the patient is not on CPB to prevent backflow out of the arterial cannula, particularly when a centrifugal pump is used. The venous reservoir serves as a buffer for fluctuations in venous drainage and is a source of fluid for rapid transfusion. It usually is placed before a membrane oxygenator (before the pump but often physically attached to the membrane oxygenator housing). When a bubble oxygenator is used, it is usually incorporated as part of the oxygenator and before the pump. Various fluids, such as blood and crystalloid solutions, and drugs may be added to this reservoir. Several suction devices and systems, usually using one or more of the roller pumps, may aspirate blood and gas from the open heart chambers (hence the term "cardiotomy suction"), surgical field, aortic root (during aortic cross-clamping as a left ventricular vent and after unclamping as an air vent), and left ventricular vent. This blood is then passed into the cardiotomy reservoir, which may be incorporated in the housing of an open (hard-shell) venous reservoir or may first flow into an external cardiotomy reservoir before emptying into a separate venous reservoir or bubble oxygenator.

FIG 5.1. Detailed schematic diagram of arrangement of a typical cardiopulmonary bypass circuit using a membrane oxygenator 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 membrane oxygenator 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.

A cardioplegia delivery and/or coronary perfusion system is another component that typically uses one of the roller pumps for administering blood or cardioplegic solution into the coronary arteries, aortic root, or coronary sinus. This circuit usually includes a separate heat exchanger and may include a reservoir and sometimes a recirculation line from the surgical field, which is used when cardioplegic solution is not being administered into the heart, although a single-pass delivery system is more commonly used. Often, arterial blood is simultaneously mixed with crystalloid-based cardioplegic solution (usually in a 4:1 blood-to-fluid ratio) to produce blood cardioplegia (see Chap. 13).

A source of oxygen, air, and sometimes carbon dioxide, with appropriate flow meters and blenders, supplies ventilating gas to the oxygenator, usually through an in-line anesthetic vaporizer. Although at times hot and cold water is supplied from wall outlets to a mixing valve for adjusting water temperature in the heat exchangers, most commonly a dedicated stand-alone water cooler and heater is used for this purpose. A number of filters (macro or micro) are often included at various sites in the CPB circuit (e.g., cardiotomy reservoir, venous reservoir, oxygenator, and arterial line). Also included are sampling ports (pre- and postoxygenator), pressure monitoring sites such as the cardioplegia-coronary perfusion delivery line and the arterial line (after the systemic pump but before the arterial filter), and arterial and venous in-line blood-gas monitors. Temperature monitoring sites, such as water inflow and outflow for major heat exchanger, venous and arterial blood, cardioplegic solution, and water bath, are also present.

Whenever a centrifugal pump is used, a flowmeter must be included in the systemic outflow line. Various safety devices and monitors, besides those already mentioned, are frequently incorporated into the CPB circuit, including: a bubble trap on the arterial line, often incorporating a microfilter and purge line that includes a one-way valve that drains back to the venous or cardiotomy reservoir; a bypass line that goes around the arterial filter in case the latter becomes obstructed; an air bubble detector on the systemic flow line; and a low-level alarm on the venous or arterial reservoir. A hemoconcentrator is sometimes attached between the systemic flow line, or some other source of blood under pressure, and the venous or cardiotomy reservoir.


Back to Quick Links

Principles of venous drainage

Venous drainage is usually accomplished by gravity siphonage. However, recently there has been a renewed interest in applying suction to the venous lines, a technique that had been discarded early in the history of CPB. Siphonage places two constraints on successful venous drainage. First, the venous reservoir must be below the level of the patient and, second, the lines must be full of blood (or fluid) or else an air lock will occur and disrupt the siphon effect. The amount of venous drainage is determined by the pressure in the central veins (patient's blood volume), the difference in height of the patient and the top of the blood level in the venous reservoir or entrance of venous line into a bubble oxygenator (negative pressure exerted by gravity equals this height differential in centimeters of water), and the resistance in the venous cannulas, venous line and connectors, and venous clamp, if one is in use.

The central venous pressure is influenced by intravascular volume and venous compliance, which is influenced by medications, sympathetic tone, and anesthesia. Excessive drainage (i.e., drainage faster than blood is returning to the central veins, which may be caused by an excessive negative pressure caused by gravity) may cause the compliant vein walls to collapse around the ends of the venous cannulas (manifested by line "chattering" or "fluttering") and intermittent reduction of venous drainage. This may be ameliorated by partially occluding the clamp on the venous line, which may paradoxically increase venous drainage, or by increasing the systemic blood flow. Obviously, the ultimate limit to venous flow is the amount of blood returning to the great veins from the body.

Types and sizes of cannulas

Venous cannulas are either single or two stage (cavoatrial) (Fig. 5.2). The latter have a wider portion with holes in the section designed to sit in the RA and a narrower tip designed to rest in the IVC. Cannulas are usually made of a flexible plastic; most are wire reinforced to prevent kinking. They may be straight or right angled. Some of the latter are constructed of hard plastic or metal for optimal inner diameter (ID) to outer diameter (OD) ratio. The venous cannulas are typically the narrowest component of the CPB venous system and thus are a limiting factor for venous drainage. Knowing the flow characteristics of the particular catheter, which should be provided by the manufacturer or established by benchtop testing, and the required flow (about one third of total flow from SVC and two thirds of total flow from IVC), one can select the appropriate venous cannula for a patient. For example, a 1.8-m2 patient (total estimated flow, 5.4 L/min; SVC, 1.8 L/min; IVC, 3.6 L/min) at a siphon (gravity) gradient of 40 cm would require at least a 30 French (F) SVC, a 34F IVC, or a single 38F single-stage catheter (1,2). These requirements are easily met by various 36F to 51F cavoatrial cannulas (3). Delius et al. (4) offered a new method for describing the performance of cannulas used in extracorporeal circulation called the M number. They reported the M numbers of several currently available cannulas and provided a nomogram for determining the M number and for predicting the pressure gradient across any cannula at any flow based on this number.

FIG 5.2. Drawings of conventional venous cannulae. A: Standard, tapered, two-stage cavoatrial cannula for insertion into the right atrium (RA) and inferior vena cava (IVC). B: Wire-reinforced cannula for atrial or caval cannulation. C: Cannula with right-angled tip (usually made of metal or hard plastic because the thin wall optimizes the ratio of internal to external diameters). This type of cannula is often used for congenital or pediatric cases and may be inserted directly into the vena cava near its junction with the RA.

Although venous (and arterial) cannulas are considered disposable and are not intended for reuse, current cost-containment pressures have led to reconsideration of this practice (5).

Connection to the patient

Usually, the venous connection for CPB is accomplished by inserting cannulas into the RA. Three basic approaches are used (Table 5.1 and Fig. 5.3): bicaval, in which separate cannulas are inserted into SVC and IVC; single atrial; and cavoatrial (i.e., the two-stage approach). The latter has a wider proximal section with holes that lie within the RA and a narrower extension with end and side holes that extends into the IVC. When bicaval cannulas are used, tapes are frequently placed around the cavae and passed through small tubes so they may be cinched down as tourniquets or snares around the cannula. This forces all the patient's venous return to pass to the extracorporeal circuit, preventing any systemic venous blood from getting into the right heart and any air (if the right heart is opened) from getting into the venous lines. This is referred to as caval occlusion, or total CPB.

FIG 5.3. Methods of venous cannulation. A: Single cannulation of right atrium (RA) with a "two-stage" cavoatrial cannula. This is typically inserted through the right atrial appendage. Note that the narrower tip of the cannula is in the inferior vena cava (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 superior vena cava (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 cannulas 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.

Other ways of accomplishing this include the use of elastic tapes placed around the cavae and held together with vascular clips (6) and the use of specially designed external clamps that go around the cavae and their contained cannulas (7). Cuffed venous cannulas may be used, either specially designed for this purpose (e.g., model 191037, Medtronic DLP, Inc., Grand Rapids, MI) (8), or cuffed endotracheal tubes (9,10). The latter may be helpful in emergency cases and when dissection around the vena cava to place tapes could be particularly difficult or dangerous. When there is a hole in the atrium and it is not possible (or there is not enough time) to insert a purse-string suture or the suture breaks, a cuffed endotracheal tube may also be used for venous drainage (11). After insertion, the cuff is inflated and gentle traction tamponades the hole in the atrium so adequate venous drainage may be provided.

Arom et al. (12) and Bennett et al. (13) compared the efficiency of the various approaches for venous drainage (Table 5.1). Bicaval cannulation with caval occlusion is required any time the right heart is entered. This approach may provide the best caval decompression if properly positioned. However, caval cannulas cause greater interference with venous flow (and hence cardiac output) when not on CPB (i.e., after cannulation but before going on bypass and after bypass but before decannulation). When the caval tapes are tightened, no provision for decompression of the right heart (atrium and ventricle) is provided. If the right ventricle is not able to eject, then coronary sinus blood returning to the RA must be removed by opening or venting the right heart or releasing the caval tourniquets. This would be aggravated by presence of a left superior vena cava (LSVC). When the aorta is cross-clamped, coronary sinus flow is greatly reduced. However, the problem of right heart decompression recurs whenever cardioplegia or direct coronary perfusion is administered.

Bicaval cannulation without caval tourniquets is often preferred for mitral valve surgery because the retraction necessary often distorts the cavoatrial junctions, interfering with venous drainage if only a single atrial cannula is used. Right heart decompression is much better than when caval tourniquets are used but may not be as good as with atrial cannulation.

Single atrial cannulation has the advantage of being simpler, faster, and less traumatic, with one less incision, and provides fairly good drainage of both the cavae and the right heart. It interferes least with caval return when off bypass. However, the quality of its drainage of the cavae and right heart is sensitive to positioning, especially with distortion of the heart (e.g., "circumflex position" when lifting the heart to make an anastomosis to posterior branches of the circumflex coronary arteries). The cavoatrial cannula has many advantages of a single right atrial cannula but may provide superior drainage of the right heart, especially in the circumflex position, perhaps by providing some stability to the position of the atrial holes (13).

Although drainage of the IVC remains good with cavoatrial cannulation in the circumflex position, drainage of the SVC is often compromised. Proper location of the atrial holes is critical to optimal drainage by this cannula (12), and adequacy of decompression of the right heart and myocardial temperature must be monitored and appropriate adjustments made when needed. Some controversy has occurred regarding the effect of the type of venous cannulation on the adequacy of myocardial protection during aortic cross-clamping with cardioplegic arrest. The concern is that with atrial cannulation alone, relatively warm (about 25 to 30°C) blood returning from the body may bathe the right heart and interfere with myocardial protection (17).

Bennett et al. (14) studied the effects of venous drainage on myocardial preservation in a dog model and compared cavoatrial cannulation with biatrial cannulation with or without caval tourniquets. They observed the greatest myocardial cooling, the slowest rewarming (between doses of cardioplegic solution), and the least evidence of myocardial ischemia with cavoatrial cannulation, which they attributed to superior decompression of the right heart. The fact that most surgeons use a cavoatrial cannula for coronary artery bypass grafting surgery with apparent good results corroborates these observations. Specially designed swirl-tip atriocaval catheters (model VC2, Medtronic DLP, Inc.) and 45-degree two-stage cannulas (Research Medical, Inc., Midvale, UT) (18) may facilitate venous drainage, especially during limited access surgery.

Taylor and Effler (19) and Kirklin and Barratt-Boyes (2) reviewed the surgical technique of venous cannulation. Single atrial cannulas are usually inserted through the right atrial appendage after placing a purse-string suture. Bicaval cannulas are usually placed through separate incisions, although some surgeons may place both through a single incision in the atrial appendage. The SVC cannula is usually passed through the right atrial appendage. The IVC cannula is usually passed through a purse-string suture placed in the posteroinferior portion of the lateral wall of the RA near the IVC and avoiding the right coronary artery. The cavoatrial junctions may be dangerously thin. Some surgeons place purse-string sutures directly in the SVC and IVC, but this could cause narrowing when closed.

At times, venous cannulation is accomplished peripherally, usually via the femoral or iliac veins. This is used for emergency closed cardiopulmonary assist, for support of particularly ill patients before induction of anesthesia, for prevention or management of bleeding complications during sternotomy for reoperations (20), and for certain types of aortic and thoracic surgery. The key to adequate flow rates with peripheral cannulation is use of as large a cannula as possible and advancing the catheter into the RA guided by transesophageal echocardiography (TEE), if available. Specially designed, commercially available (e.g., Medtronic BioMedicus, Inc., Eden Prairie, MN), long, ultrathin, nonkinkable, wire-reinforced catheters are available for this purpose. Insertion may be facilitated by use of an internal stylet and guidewire. Jones et al. (21) documented flows of up to 3.6 L/min (25F) to 4.0 L/min (27F and 29F) with simple gravity drainage. Using another brand of femoral venous catheter (model Femflex II, Research Medical, Inc.) and gravity drainage, Merin et al. (20) obtained flows of up to 2.5 L/min with 20F catheters and flows of 3.5 to 4.5 L/min with 28F catheters. This flow can be augmented by use of kinetic or vacuum assistance, which is discussed below.

Alternatively, venous catheters intended for percutaneous CPB may be used. Westaby (22) suggested that in cases where IVC drainage alone does not provide adequate venous return, adding a 32F cannula inserted into the SVC via a cut-down in the right internal jugular vein is effective. In contrast, Flege and Wolf (23) described using the right internal jugular vein as the sole source of venous drainage for conduct of CPB using percutaneously placed 21F 20-cm-long femoral arterial catheters (Medtronic DLP, Inc.) advanced into the RA and augmented venous drainage. Bicaval femoral venous cannulas (29F and 33F) are available (Medtronic DLP, Inc.) that allow drainage of the SVC and IVC while isolating the RA by snaring the SVC and IVC around the cannula (24).

Persistent left superior vena cava

An LSVC is present in about 0.3% to 0.5% of the general population but in 2% to 10% of patients with congenital heart disease and in up to 40% when such patients have abnormal sinus. It usually drains into the coronary sinus and then into the RA (25–28). In about 10% of cases, usually associated with other congenital heart disease, the LSVC drains into the left atrium. In some cases, there are defects in the wall between the coronary sinus and the left atrium permitting intercommunication between the left atrium and RA (e.g., coronary sinus type atrial septal defect).

The presence of an LSVC should be suspected when a large coronary sinus is noted on echocardiography (differential diagnosis includes right-sided venous hypertension, tricuspid regurgitation, and stenosis of the ostium of the coronary sinus) (29). Sometimes the LSVC itself can be seen on echocardiography posteriorly and laterally to the left atrium above the atrio-ventricular groove beside the aorta. Its presence can be confirmed by injection of agitated saline echocontrast into a left arm vein and noting passage into the coronary sinus before its arrival into the RA. The surgeon should suspect an LSVC when the (left) innominate vein is small or absent.

The presence of an LSVC poses a number of problems during cardiac surgery. It may confuse and complicate passage of a pulmonary artery catheter or interfere with administration of retrograde cardioplegic solution (30). The latter is difficult because the coronary sinus is usually quite large in this circumstance and thus the balloon on the retrograde catheter does not seal and cardioplegic solution leaks into the RA. Furthermore, the cardioplegic solution may run off into the persistent LSVC, and the cardioplegic solution will be diluted with systemic venous blood draining down the LSVC. Patients with this anomaly may be more prone to atrial tachycardia and may have other congenital cardiac abnormalities. Finally, the presence of an LSVC poses obvious problems if the right heart is to be entered or with right heart decompression and adequacy of venous return if bicaval cannulation is used because of the flow of the additional systemic venous blood into the RA.

If the right heart is not going to be opened and a single or two-staged venous cannula is used for venous drainage and retrograde cardioplegia is not used, the presence of an LSVC poses no problems. If the right heart needs to be entered, several options are available. If an adequately sized innominate vein is present (true in about 30% of cases), the LSVC can simply be occluded during CPB. However, one must be wary of the rare possibility of the associated anomaly of atresia of the coronary sinus, in which case the LSVC provides the main outlet for cardiac venous drainage and occlusion of the LSVC could injure the myocardium (31). This condition should be suspected if the coronary sinus is not enlarged and by failure of echocontrast injected in the left arm or LSVC from entering the RA via the coronary sinus during TEE. Another obvious circumstance, in which the LSVC cannot be occluded despite the presence of an adequately sized innominate vein, is when there is associated absence of the right SVC (true in about 20% of cases).

If the innominate vein is absent (true in about 40% of cases) or small (true in about 33% of cases), occlusion of the LSVC may cause serous venous hypertension and potentially cerebral injury or ischemia and should not be done without documenting acceptable venous pressure in the LSVC cephalad to the occlusion. Otherwise, some other arrangement must be made to provide drainage of the LSVC. Use of cardiotomy suction in the coronary sinus ostium may be adequate, but cannulation, usually via a cannula passed retrograde into the LSVC through the ostium of the coronary sinus, is preferred, with a caval tape (tourniquet) placed around the LSVC. Alternatively, a cuffed caval cannula or endotracheal tube may be used (26). A caval cannula could be placed directly into the LSVC through a purse-string suture placed externally. Finally, in small infants, induction of deep hypothermia with CPB cooling using a single venous cannula followed by circulatory arrest obviates the need for extra cannulation of the LSVC.

Augmented venous return

Early in the history of CPB, suction pumps (roller or finger) were used for venous drainage, but because they were difficult to control, these were discarded in favor of the more simple and effective gravity siphon method described above. Recently, there has been a renewed interest in use of regulated suction to overcome the resistance of longer and/or narrower venous cannulas used during limited (transthoracic) access or peripheral venous (e.g., jugular or femoral veins) access. With these narrower and longer venous cannulas, gravity siphon alone may not provide adequate flow for full CPB even with a maximal height differential between the patient and the venous reservoir.

Three methods to augment venous return have been described. One is to place a roller pump in the venous line between the venous cannula and the venous reservoir (32). This carries a high risk of generating excess negative pressure and collapsing the RA or great veins around the cannula tip and requires constant attention and adjustment of the roller pump flow rate. A second method substitutes a kinetic (centrifugal) pump for the roller pump in the venous line and is referred to as kinetic-assisted venous drainage (KAVD) (33). The third method involves applying a regulated vacuum to a closed hard-shell venous reservoir attached to the venous line. This method is referred to as vacuum-assisted venous drainage (34,35). This system is relatively simple and does not require regulation of a second pump. When a pump (roller or centrifugal) is used to generate the suction, some advocate placing a shunt around the pump using two Y-connectors. When a centrifugal pump is used, the shunt is clamped but is opened in the event entrained venous line air deprimes the centrifugal pump. When a roller pump is used, the shunt is only partially occluded to prevent generation of excessive negative pressures.

With all three systems, a conventional systemic pump (centrifugal or roller) then pumps blood out of the venous reservoir through the oxygenator and to the patient. Fried et al. (36) described use of a single pump for KAVD. Using their method, one centrifugal pump both aspirates venous blood and pumps it to the patient. This requires that the venous reservoir is T-ed into the venous drainage line to sequester or add blood to the system. This method reduces the problem of balancing the flow of two pumps but runs the risk of systemic air embolization (37,38).

Use of any of these systems requires careful regulation of the degree of negative pressure applied to the venous line. This is best accomplished by monitoring the pressure in the venous line about 10 cm before the inlet to the venous pump (roller or kinetic) or the hard-shell reservoir (if using a vacuum-assisted system). The negative pressure (or vacuum) measured at this site should not exceed –60 to –100 mm Hg (33). It is also desirable to observe the RA directly or via TEE. When the vacuum-assisted system is used, the degree of vacuum applied should be controlled with a vacuum regulator that can be adjusted and display low levels of suction in 10-mm Hg increments. It is important that vacuum should never be applied when there is no forward blood flow through the oxygenator to prevent air from being pulled across the microporous membrane into the blood path. The reservoir also should be open to the atmosphere when vacuum is not being applied to prevent overpressurization of the venous reservoir with reduction of venous return and risk of retrograde or antegrade air embolization. The venous reservoir should have a low positive (about +15 mm Hg) and a high negative (about –150 mm Hg) relief valve. Usually, adequate venous drainage is achieved with speeds of 1,000 to 1,200 revolutions per minute (rpm) of the kinetic pump or application of 20 to 60 mm Hg vacuum to the venous reservoir.

When a vacuum-assisted system is used with a closed reservoir, the degree of vacuum within the reservoir is influenced not only by the amount of vacuum applied to the system but also the relative flow of blood and air into the reservoir (from the venous line and cardiotomy suction and vents) and blood out of the reservoir (by the systemic pump).

There are a number of potential problems and risks associated with the use of augmented venous return methods. Excessive negative pressure may cause hemolysis because red blood cells are more easily damaged by negative than positive pressure (39,40). There also may be collapse of right atrial, tricuspid valve, or venous structures around the cannula tip resulting in impaired venous return and "chattering" in the venous line and possible damage to cardiovascular structures. Application of additional negative pressure (beyond gravity) increases the risk and amount of air aspiration from holes in walls of the RA or great veins. Air may also enter through a patent foramen ovale if the left heart is open or through any intravenous lines or introducers that may be in place (these should be closed or placed in occlusive infusion pumps during augmented venous return). Any aspirated air may cause an air lock or can deprime a centrifugal pump and stop blood flow or fill the hard-shell venous reservoir and contribute to systemic air embolization. If vacuum is applied to the closed reservoir system during a no-flow state, there is the theoretic risk of pulling air across the microporous membrane into the blood path with subsequent systemic air embolism. If the venous reservoir is not open to atmosphere when vacuum is not being applied or when KAVD is in use, the venous reservoir can become overpressurized with reduction of venous return and increase the risk of retrograde (through cardiac vents or cardioplegia line) or antegrade (systemic arterial line) air embolization. When a pump is being used to augment venous return, there also is potential for imbalance of flow between the venous drainage and the systemic flow pump, resulting in a change in intravascular volume in the patient or a risk of systemic air embolism. Thus, these methods of augmenting venous drainage require that the perfusionist must be even more attentive than when using conventional gravity siphon drainage.

Complications associated with achieving venous drainage

These include atrial dysrhythmias, laceration and bleeding of the atrium, air embolization (especially if the atrial pressure is low, which could cause systemic embolization with potential right-to-left shunts), laceration of the vena cavae (the IVC is particularly prone to this), and malposition of the tips (or atrial portion of the cavoatrial catheter), including inserting the tips into the azygos, innominate, or hepatic veins or across an atrial septal defect into the left heart. Placement of the low atrial purse-string suture for cannulation of the IVC requires retraction of the heart, which may have adverse hemodynamic consequences and sometimes is deferred until the patient is placed on bypass with a single cannula (SVC). Placing tapes around the cavae may result in lacerations of the cavae themselves or branches off the cavae or, when encircling the SVC, the right pulmonary artery. Once the venous cannulas are in place, they may interfere with venous return and cardiac output until CPB is initiated. Placing venous cannulas may displace central venous or pulmonary artery catheters inserted for hemodynamic monitoring (41). Caval tapes may occlude these lines and, conversely, their presence may prevent tight caval occlusion by the tapes. Further, these monitor lines may become caught in atrial purse-string sutures, causing malfunction and preventing their removal (42). Finally, the cavae may become obstructed when purse-string sutures placed in the cavae are closed after cannula removal (43).

Causes of low venous return

Reduced venous drainage may be due to reduced venous pressure, inadequate height of the patient above the CPB reservoir, malposition of the venous cannulas (sometimes due to surgical manipulation of heart), or obstruction or excess resistance of the lines and cannulas. Inadequate venous pressure may be caused be venodilation with drugs (e.g., nitroglycerin, inhalation anesthetics) or hypovolemia. Kinks, air lock, or insertion of a pulmonary artery balloon catheter into a cannula (44) may cause obstruction, or the cannulas may be too small. During rewarming, the tendency for kinking of cannulas is potentially aggravated because of softening of the tubing and/or surgical manipulation of the heart.


Back to Quick Links


Many different types of cannulas are available, made of various materials (Figs. 5.4 and 5.5). Some that are designed for insertion into the ascending aorta have right-angled tips; some are tapered; and some have flanges to aid in fixation and prevent introduction of too great a length into the aorta. The arterial cannula is usually the narrowest part of the extracorporeal circuit. High flow through narrow cannulas may lead to high pressure gradients, high velocity of flow (jets), turbulence, and cavitation with undesirable consequences, which are discussed below.

FIG 5.4. Conventional arterial cannulas. A: Bevel-tipped tapered "blue line" Texas Heart Institute (THI)-type cannula with molded flange near tip. B: Similar cannula without a flange ("Bardic" type) that also can be used for femoral arterial cannulation.

FIG 5.5. Newer arterial cannulas. A: Metal-tipped right-angled cannula with plastic molded flange for securing cannula to aorta. B: Similar design but with a plastic right angle tip and molded flange. C: (Left) Diffusion-tipped angled cannula designed to direct systemic flow in four directions to avoid a "jetting effect" that may occur with conventional single-lumen arterial cannulas. An inverted cone occludes the tip. (Right) Drawing with arrows depicts flow patterns. D: Integral cannula connector and luer port (for de-airing) incorporated into some arterial cannulas; newer arterial cannulas may contain a self-venting cap (not shown) for removal of air during insertion.

Hemodynamic evaluations of arterial cannulas have traditionally been based on measurement of the pressure drop. A useful descriptive characteristic of an arterial cannula is its performance index (pressure gradient versus OD at any given flow) (45). The narrowest portion of the catheter that enters the aorta should be as short as compatible with safety, and thereafter the cannula size should enlarge to minimize the gradient. Long catheters with a uniform narrow diameter are undesirable. The use of metal or hard plastic for the tip provides the best ID-to-OD ratio. Pressure gradients exceeding 100 mm Hg are associated with excessive hemolysis and protein denaturation (46). Thus, it is preferable to select a cannula that will provide adequate flow with no more than 100-mm Hg pressure gradient. Drews et al. (47) suggested that in small-sized cannulas, the right angle configuration (as compared with straight configuration) may aggravate hemolysis. New approaches to hemodynamic evaluation of arterial cannula include velocimetry (48) and detailed analysis of flow patterns using laser Doppler anemometry (49), color Doppler ultrasound, and high-field magnetic resonance imaging (50), but the clinical relevance of these studies is yet to be demonstrated.

The jetting effect produced by small cannulas may damage the interior aortic wall, dislodge atheroemboli and cause arterial dissections, and disturb flow into nearby vessels. Muehrcke et al. (49) described a new aortic cannula that has a closed tip and internal cone designed to reduce exit forces and velocities to reduce these adverse jet effects (Fig. 5.5C). Hemolysis rates were similar while pressure gradients were intermediate compared with a number of other cannulas in common use.

Brodman et al. (45) evaluated 29 different types of arterial cannulas. They found that an 8-mm OD high-flow aortic arch cannula (model 15235, 3M Healthcare, Inc., Ann Arbor, MI) and an 8-mm OD aortic cannula with or without flange (models 1858 and 1860, CR Bard, Billerica, MA) were best (gradient less than 50 mm Hg at flows of 5 L/min), whereas several others were unacceptable (gradient more than 100 mm Hg at flows of 4 L/min). For cannulas not studied by them, one should refer to gradient-flow data provided by the manufacturer or conduct benchtop tests. Unfortunately, the data of Brodman et al. may underestimate clinical gradients because they used water rather than blood or a blood analogue as the fluid in their studies. Size and shape of aortic cannulas did not influence rate of transcranial Doppler-detected microemboli in one study (51).

Connection to the patient

Ascending Aorta

In the early days of CPB, arterial inflow was via the subclavian or femoral artery (52), but currently it is usually via a cannula inserted into the ascending aorta (53). The advantages of this approach over the femoral or iliac arteries (Table 5.2) include ease, safety, and the fact that it does not require an additional incision. The surgical technique for aortic cannulation has been reviewed in detail by many others (2,19,54,55). The site for cannulation is selected based on the type of cannula to be used, the operation planned (how much of the ascending aorta must be available), and the quality of the aortic wall (56).

Atherosclerosis with or without calcification frequently involves the ascending aorta and poses problems regarding arterial cannulation and application of clamps and vascular grafts. Dislodgement of atheromatous debris either by direct mechanical disruption or from the "sand-blasting" effect of the jet coming out of the arterial cannula is thought to be a major cause of perioperative stroke (57–59). Atherosclerosis is also considered a risk factor for perioperative aortic dissection (60) and postoperative renal dysfunction (61).

Traditionally, surgeons have relied on palpation to detect these changes and select sites for cannulation, cross-clamping, and so on, and this should continue to be a component of the evaluation of the aorta (62). However, this method is much less sensitive and accurate than epivascular ultrasonic scanning (63–66). Mills and Everson (56) recommend using a 10- to 20-second period of venous inflow occlusion to reduce systemic arterial pressure to 40 to 50 mm Hg to improve the reliability of palpation of the ascending aorta. Specially designed epivascular probes, or transthoracic probes, encased in sterile sheaths with ultrasound jelly or lubricant inside and saline covering the aorta may be used for epiaortic scanning but do require some time and effort on the part of the surgeon. The use of a saline-filled glove placed between the probe and the aorta as a step-off may improve the image. Unfortunately, TEE, which is more convenient, is not sensitive enough because of limited views of the ascending aorta (65–67). However, some believe it can be used as a screening method to determine which patients need epiaortic scanning. If no significant atherosclerosis is detected in the ascending, transverse, or proximal descending aorta, it has been suggested that epiaortic imaging is not necessary (68), but others disagree (66,69).

Epiaortic and TEE should be considered complimentary (66). Beique et al. (59) suggested using epiaortic scanning in all patients who have a history of transient ischemic attacks, strokes, severe peripheral vascular disease, palpable calcification in the ascending aorta, calcified aortic knob on chest x-ray, those older than 60 years, and those with TEE findings of moderate aortic atherosclerosis. Others advocate epiaortic scanning of the ascending aorta in all patients over 50 years of age (70). If atherosclerosis is detected, then sites for insertion of cannulas, grafts, and application of vascular clamps are modified. If extensive atherosclerosis precludes arterial cannulation in the ascending aorta, then the femoral route should be considered (see below). However, in this case, the transverse and descending aorta should be evaluated by TEE to rule out extensive atheroma that might be embolized into the brain or elsewhere with retrograde flow from a femoral cannula. If such is the case, then axillary-subclavian cannulation should be considered.

If atheroma is extensive in the ascending or transverse aorta, some clinicians have suggested using a long arterial cannula that is inserted in the ascending aorta and threaded around into the proximal descending aorta to reduce the "sand-blasting" effect (71). Others have advocated doing an endarterectomy under deep hypothermic circulatory arrest if severely protruding or mobile atheroma are detected (72), but in a recent study this has been associated with a higher stroke rate (35% versus 12%) and mortality (19% versus 12%) than when endarterectomy was not performed (73).

If the ascending aorta is totally calcified and rigid (so-called porcelain aorta), then entirely different strategies for cannulation and surgery must be used (74). These include no clamping of the ascending aorta, femoral or axillary-subclavian arterial cannulation, and, in selected cases, graft replacement or endarterectomy of the ascending aorta during deep hypothermic circulatory arrest (70,75). Unfortunately, graft replacement of the atherosclerotic ascending aorta may be a high-risk procedure (76). If there is no intraluminal debris, Liddicoat et al. (77) used an intraluminal balloon designed for port-access surgery that is inserted through a purse-string suture in an atherosclerotic free portion of aortic arch to occlude the aorta. Others used a Foley catheter in a similar manner (78). Studies using historical control subjects suggest improved neurologic outcome with echocardiographic-based modification of surgical techniques in handling the ascending aorta (59,64,79).

If possible, the intrapericardial aorta is chosen for aortic cannulation because this segment best resists tearing or dissection. Many surgeons insert two concentric purse-string sutures into the aortic wall. Surgeons differ as to whether these should be shallow, deep, or full-thickness bites (2,80). Most surgeons then incise and dissect away the adventitia within the purse-string suture. Most avoid using a partial occluding clamp, except in pediatric patients, to minimize clamp trauma to the aorta. Optimal arterial blood pressure during cannulation (mean arterial pressure of about 70 to 80 mm Hg, systolic pressure of about 100 to 120 mm Hg) is important: if too high, there is a greater chance of tears and dissection and blood loss and spray; if too low, the aorta tends to collapse, it is harder to make an incision and insert the cannula, and there is a greater risk of damaging the back wall of the aorta. An appropriately long full-thickness incision is then made, and the leak is controlled with a finger or by approximating the adventitia or by simultaneously inserting the cannula. Alternatively, Garcia-Rinaldi et al. (81) suggested only incising down to the intima during aortic cannulation.

Dilators are routinely or selectively used. If a right-angled tip is used, it is often initially directed toward the heart and then rotated 180 degrees to confirm intraluminal placement. Slight or vigorous back bleeding is allowed to eliminate air or atheromatous debris and to further confirm intraluminal placement, which can be additionally confirmed by noting a pulsatile pressure in the systemic flow line pressure monitor on the CPB circuit, which should approximate the radial artery pressure. Proper position of the cannula tip is critical. Most surgeons insert only 1 to 2 cm of the tip into the aorta and direct it toward the middle of the transverse arch to avoid entering the arch vessels. Others have advocated threading a long cannula into the proximal descending aorta to reduce the velocity and turbulence in the aortic arch to reduce the "sand-blast" effect and emboli (71).

Many potential complications of aortic root cannulation exist (52,82), including: inability to introduce the cannula (interference by adventitia or plaques, too small an incision, fibrosis of wall, low arterial pressure), intramural placement, dislodgment of atheroemboli (56,64), and air embolism from the cannula or if the aortic pressure is very low, injury to the back wall of the aorta; persistent bleeding around the cannula or at the site after its removal; malposition of the tip to a retrograde position or even across the aortic valve, against the vessel wall, or into the arch vessels (52); abnormal cerebral perfusion; obstruction of the aorta in infants (83); aortic dissection; and high CPB line pressure. High systemic flow line pressure may be a clue to malposition of the tip against the vessel wall or into an arch vessel, cannula occlusion by the aortic cross-clamp, aortic dissection, a kink anywhere in the inflow system, including a line clamp still on, or use of too small a cannula for the intended CPB flow.

Inadvertent cannulation of the arch vessels or the direction of a jet into an arch vessel may cause irreversible cerebral injury and reduced systemic perfusion (52,83–88). Suggestive evidence includes: high systemic line pressure in the CPB circuit; high pressure in the radial artery if supplied by the inadvertently cannulated vessel (52) (or low pressure if not supplied by the cannulated vessel); unilateral facial blanching when initiating bypass with a clear priming solution (86); asymmetric cooling of the neck during perfusion cooling (88); and unilateral hyperemia, edema, petechia, otorrhea, or dilated pupils. Before CPB, palpation of the carotids may reveal asymmetric pulsation (reduced on cannulated side) and the opposite may be observed during pulsatile bypass (increased pulsation on cannulated side) (88). Before CPB, the radial artery catheter may reveal sudden damping if the cannula is inserted in the arch vessel supplying the monitored radial artery (89).

It has been suggested that the Coanda effect (in which a jet stream adheres to the boundary wall and hence produces a lower pressure along the opposite wall) may be associated with carotid hypoperfusion (90). This has been shown to occur experimentally and may account for some cerebral dysfunction after CPB using aortic cannulation. Salerno et al. (91) detected major electroencephalographic abnormalities due to malposition of a cannula in 3 of 84 patients undergoing arch perfusion, possibly on the basis of the Coanda effect.

Antegrade aortic dissection associated with ascending aortic cannulation has been reported to occur in between 0.01% and 0.09% of cases (Table 5.3) (54,60,80,92). Aortic dissection should be suspected when any of the following are observed: a sudden decrease in both venous return and arterial pressure, excessive loss of perfusate, increased CPB systemic line pressure, evidence of decreased organ perfusion (oliguria, dilated pupil, electroencephalographic changes, electrocardiographic evidence of myocardial ischemia), blue discoloration of the aortic root (because of subadventitial hematoma), and bleeding from needle or cannulation sites in the aortic root. Subadventitial hematomas tend to be less extensive and to be softer and usually resolve when incised. TEE may be useful in diagnosing aortic dissection (93). Gott et al. (92) and Still et al. (80) discussed this complication in detail.

Management usually involves prompt cessation of CPB; recannulation distal to the dissection (usually femoral but occasionally into the distal aortic arch); induction of deep hypothermia and a period of circulatory arrest while the aorta is opened and the extent of the injury analyzed and repaired either by direct closure, use of a patch, or replacement of ascending aorta with a tubular graft. Occasionally, small injuries can be repaired off bypass by closed plication (92). Survival of those recognized and treated in the operating room has ranged from 66% to 85%. When not recognized until postoperatively, survival has been 50% or less (Table 5.3). Bleeding and infected or noninfected false aneurysms are late complications of aortic cannulation (82).

Femoral Artery

Cannulation of the femoral or iliac arteries (exposed via a retroperitoneal suprainguinal approach) is indicated when there is an aneurysm of the ascending aorta or when it is otherwise unsatisfactory for cannulation (56,64). It may be indicated when there is no space available due to multiple procedures involving the ascending aorta, for peripheral cannulation under local anesthesia in unstable patients, during reoperations prophylactically (20), when bleeding comlications occur during reentry, or when an antegrade dissection complicates aortic cannulation. Femoral cannulation requires a second incision and limits the size of the cannula that can be used. Hence, the adverse consequences of fluid jetting effects and high pressure gradients are more likely. Lees et al. (94) found no difference in the distribution of blood flow and vascular resistance between retrograde (femoral artery infusion) and antegrade (aortic root infusion) flow in monkeys.

Femoral cannulation is associated with many complications (20,53,91), including: trauma to the cannulated vessel, such as tears, dissection, late stenosis or thrombosis, and bleeding; lymph fistula; infection; embolization; and limb ischemia. Because the retrograde perfusion cannula usually totally occludes the blood supply to the cannulated limb, ischemic complications (acidosis, compartment syndrome, muscle necrosis, and neuropathy) may develop if cannulation exceeds 3 to 6 hours (95–97). The risk of distal ischemia can be minimized by placing a Y-connector or leur-lock port in the arterial line and attaching a smaller cannula (e.g., 8F to 14F pediatric arterial cannula) inserted distally through the same arteriotomy (96) or a 8.5F introducer catheter inserted into the distal superficial femoral artery using the Seldinger technique (98) to maintain perfusion of the leg.

Alternately, VanderSalm (99) advocated suturing a 10-mm polytetrafluroethylene graft end-to-side on the common femoral artery into which the 24F femoral cannula is inserted. This latter technique not only prevents lower extremity ischemia but may reduce risk of arterial injury and retrograde dissection. Use of a coated Dacron graft may be associated with less bleeding (100). If distal limb perfusion is used and the ipsilateral femoral vein is also cannulated, then a method to provide better venous drainage of that limb is suggested to reduce edema. This can be minimized either by not taping the vein around the cannula (96) or placing a second (12F) venous cannula through the saphenous vein into the distal femoral vein (101). If limb ischemia does occur, Beyersdorf et al. (102) described a method of controlled limb reperfusion to improve outcome.

Femoral perfusion may lead to cerebral and coronary atheroembolism if there is extensive atheroma in the distal arch or descending aorta; ideally, this should be assessed by TEE before selecting the femoral route. If severe atherosclerosis is present, an alternate route should be used if possible. Femoral perfusion may also aggravate preexisting aortic dissections, and an alternate site for cannulation (see below) is recommended by some authors (103).

The most serious complication of femoral cannulation is retrograde arterial dissection, which may lead to retroperitoneal hemorrhage or retrograde extension all the way to the aortic root. The incidence of this complication has been reported at between 0.2% to 1.3% (104–108), although rates as high as 1 in 30 (3%) (109) and as low as 0 in 702 (91) have been reported. Kay et al. (107) noted a rate of 3% in 378 patients over 40 years old.

Femoral cannulation is being frequently used during limited access surgery and has been complicated by fatal dissection (110). Galloway et al. (111) reported a rate of about 0.8% in 1,063 patients undergoing retrograde femoral cannulation and CPB with a port-access system (HeartPort, Inc., Redwood City, CA). Retrograde arterial dissection is thought to be caused by either direct (cannula) or indirect (jet) trauma and to be more likely in the presence of atherosclerosis or cystic medial necrosis and in patients over 40 years old (53,107). Retrograde aortic dissection may present like antegrade aortic dissection already described but may be more difficult to recognize if it does not extend into the ascending aorta. In these cases it may present only by a sudden decrease in venous return and arterial pressure, excessive loss of perfusate, increased systemic line pressure, and oliguria. In this situation, TEE is extremely helpful in making the correct diagnosis.

Because of the nature of the dissection and the flap, discontinuation of retrograde perfusion and resumption of antegrade flow (via a cannula in the ascending aorta or by normal cardiac function) may resolve the problem. This may permit different management from antegrade dissections. If the dissection occurs early in the procedure, simply discontinuing CPB immediately (and hence retrograde femoral perfusion) and restoring intravascular volume (which can be facilitated by attaching the arterial line to the venous cannula and infusing perfusate from the pump) and then aborting the planned operation and doing nothing to the ascending aorta, even if it is affected by the dissection, can be successful (104). If the dissection occurs later when it is not possible or desirable to come off CPB, retrograde perfusion is immediately discontinued and the arterial cannula is introduced into the true lumen in the ascending aorta (often through the false lumen). Bypass is then resumed with antegrade perfusion and the planned operation may be completed without repair of the dissection itself or the ascending aorta. Carey et al. (106) reported long-term success in six of seven patients using this approach. Others recommend graft replacement of the ascending aorta (105).

A test infusion with the systemic pump through the arterial line (regardless of location of the arterial cannula) before initiating CPB accompanied by increased line pressure may warn of possible dissection and may avoid extensive dissection.

Axillary artery

Use of the axillary artery instead of the femoral artery when ascending aortic cannulation is not feasible or is undesirable is being increasingly advocated, either by direct cannulation or through an attached 8-mm graft (112–114). During a left thoracotomy, the intrathoracic subclavian artery may be cannulated (115). Advantages of the axillary artery over the femoral artery include: it is less likely to be involved by atherosclerosis; it has good collateral flow, decreasing the risk of ischemic complications; healing is better; and wound complications are less likely. By providing antegrade flow, it also is less likely to cause cerebral atheroembolization. Its use has also been advocated for establishing deep hypothermia before repair of type I aortic dissections because it is less likely to result in malperfusion and further expansion of the dissection as may occur with femoral perfusion (103).

Use of the right axillary artery is favored. The artery is approached through a 4- to 10-cm incision below and parallel to the lateral two thirds of the clavicle. Care must be taken to avoid traction on the brachial plexus. The axillary vein is retracted away from the artery (but may be used for venous cannulation) (113) and a purse-string suture is placed in the axillary artery and a 20F to 22F right-angled or flexible arterial cannula is inserted 2 to 3 cm. Of course, the contralateral radial or brachial artery (usually the left) must be used for intraarterial pressure monitoring. During lateral thoracotomies, the axillary artery may be approached through the axilla through a vertical incision along the lateral border of the pectoralis major (113) or the subclavian artery may be cannulated intrathoracically (115). Brachial plexus injury and axillary artery thrombosis have been reported as complications of this technique (112).

Other sites for Arterial Cannulation

Antegrade aortic perfusion can be accomplished by cannulating through the left ventricular apex and passing the cannula (20F to 22F) across the aortic valve into the aortic root (116–118). A 10F wire-reinforced arterial cannula has been used in a similar manner in an infant (119). In this circumstance, Robicsek (118) used a special padded vascular clamp (Heinrich Ulrich Co., Ulm, Germany) that allows clamping of the ascending aorta around the perfusion cannula. Both Robicsek (118) and Norman (117) described the use of special double-lumen or double-barreled cannulas that allow for both aortic perfusion and venting of the left ventricle. Coselli and Crawford (120) described retrograde perfusion through a graft sewn onto the abdominal aorta when distal occlusive disease prevents femoral cannulation and when ascending aortic cannulation is not feasible.


Back to Quick Links

Minimizing blood trauma, prime volume, resistance to flow, and avoiding leaks (either outward flow of blood or aspiration of air) are considerations in selection of tubing and connectors. To minimize blood trauma, one should strive to have smooth nonwettable inside walls of nontoxic materials, to avoid velocities above 100 cm/s, and to avoid exceeding a critical Reynolds' number above 1,000 (Table 5.4). The gradient necessary to propel the blood through the tubing should also be minimized (Table 5.4). The selection of large ID tubing aids in achieving these objectives. On the other hand, the larger the tubing, the greater the priming volume. Keeping tubing as short as possible will reduce prime volume, pressure gradients (resistance to flow), and blood trauma.

Desirable tubing characteristics include: transparency; resilience (reexpands after compression); flexibility; kink resistance; hardness (resists collapse); toughness (resists cracking and rupture); low spallation rate (the release of particles from the inner surface of tubing); inertness; smooth and nonwettable inner surface; toleration for heat sterilization; and blood compatibility. Medical-grade polyvinyl chloride seems to meet these standards. Silicone rubber and latex rubber tubing sometimes were used in roller pumps in the past; however, spallation and blood incompatibility are respective problems. New formulations of polyvinyl chloride are being developed for use in roller pumps to minimize spallation.

Disposable clear polycarbonate connectors with smooth nonwettable inner surfaces that make smooth junctions with plastic tubing (to minimize turbulence) are desirable. Smooth curves rather than sharp-angled bends will minimize turbulence. Connections must be tight enough to prevent leakage of blood when exposed to positive pressures (up to 500 mm Hg beyond the systemic flow pump) and aspiration of air on the venous side. The friction of fluted connectors with a larger OD than the ID of the plastic tubing or cannula into which the connector is inserted may provide sufficient tightness; otherwise, plastic bands may be applied tightly around all such connections at time of use. Most tubing and connectors are prepackaged and preassembled for convenience and safety. Heparin bonding onto the inner surface of the tubing and other components of the circuit may improve biocompatibility (see Chap. 9).


Back to Quick Links

These are discussed in detail in Chapter 3, and only a few aspects are mentioned here. Roller pumps are used for providing flow through the systemic line back into the patient ("arterial" pump) but also for administering cardioplegia and providing suction for vents and field ("cardiotomy") suctions and augmented venous drainage. They consist of a length of tubing located inside a curved raceway at the perimeter of the travel of rollers (usually two) mounted on the ends of rotating arms (usually two, 180 degrees apart), arranged so that one roller is compressing the tubing at all times. Flow of blood is induced by compressing the tubing, thereby pushing the blood ahead of the moving roller. Flow rate depends on the size of the tubing, length of the raceway, and rpm of the rollers. For a given pump and type and size of tubing, flow is proportional to pump speed in rpm. In vitro calibration curves should be constructed and checked periodically. This is done by measuring the output of the pump over a measured period of time at various pump settings in a benchtop circulation, preferably using blood or blood analogue.

The degree of compression, or occlusiveness, of the tubing by the rollers can be adjusted and appears to be important. Excessive compression aggravates hemolysis and tubing wear, whereas too little occlusion may also aggravate hemolysis but, more important, may reduce forward output and invalidate flow assumptions or readings based on rpm. Although there is some disagreement, most authorities believe that the least hemolysis occurs when compression is adjusted to be barely nonocclusive (39). This is accomplished by holding the outflow line vertically so the top of the fluid (blood or asanguinous) is 60 to 75 or 100 cm (24 to 30 or 39 inches) above the pump and then gradually decreasing the occlusiveness until the fluid level falls at a rate of 1 cm every 5 seconds (39) or 1 inch/min (121) or 1 cm/min (122). Mongero et al. (123) described another method of setting occlusion. The tighter setting ensures more accurate forward flow, makes the roller pump relatively insensitive to afterload, and may not be associated with more hemolysis.

Complications associated with roller pumps include: malocclusion (with the consequences noted above), miscalibration, or miscalculation including setting the wrong tube size into the pump controller (124); fracture of the pump tubing; "run away" (125); loss of power; spallation (126–128); and pumping of large amounts of air. If the outflow becomes occluded, pressure in the line will progressively rise until the tubing in the pump ruptures or connectors and tubing separate. This can be avoided by use of a pressure-regulated shunt between the outflow and inflow lines of the roller pump (123) or use of servoregulation of the pump to arterial line pressure so that it turns off when excessive pressures are detected. If inflow becomes limited, the roller pumps will develop high negative pressures producing cavitation, microbubbles, and hemolysis.

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. Unlike roller pumps, they are totally nonocclusive and afterload dependent (i.e., an increase in downstream resistance or pressure decreases forward flow delivered to the patient if no adjustment is made in the rpm). 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 quantitate 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. This can cause exsanguination of the patient or aspiration of air into the arterial line (e.g., from around the purse-string sutures) (129). Thus, whenever the centrifugal pump is not running, the arterial line must be clamped. Kolff et al. (129) described a check valve to prevent this problem. On the other hand, if the arterial line becomes occluded, these pumps will not generate excessive pressure (the maximum is only about 700 to 900 mm Hg) 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 because the maximum is only about –500 mm Hg if the inflow becomes occluded.

A reputed advantage of centrifugal pumps over roller pumps is less risk of passing massive air emboli into the arterial line. This is because they 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.


Back to Quick Links

Although numerous types of oxygenators have been used in the past, currently only two varieties are in use, the bubble and membrane oxygenators, and bubble oxygenators are rapidly disappearing from use in most parts of the world. The details concerning the function of these two types of oxygenators and the debate over whether the membrane oxygenator is superior is covered in Chapter 4. The oxygenator used does influence the configuration of the extracorporeal circuit, and often the oxygenator includes other components of the circuit (Fig. 5.6). The heat exchanger is usually an integral part of the oxygenator and usually is situated just proximal to the gas exchanging section or sometimes within the bubble chamber of a bubble oxygenator. Bubble oxygenators and some membrane oxygenators (e.g., model Capiox E, Terumo Medical, Inc., Tokyo, Japan) are positioned proximal to the pump and include an arterial reservoir, which is located distal to the oxygenating column and defoaming area and proximal to the systemic pump for which it serves as an "atrium." No additional venous reservoir is included in this circuit because the venous return empties directly into the oxygenating column of the oxygenator, as does the cardiotomy reservoir (Fig. 5.6A).

FIG 5.6. Schematic of various arrangements of cardiopulmonary bypass components. A: Sequence of components when a bubble oxygenator (BO) is used. B: Sequence when a membrane oxygenator (MO) is used. C: Older arrangement for pulsatile perfusion with MO and two pumps: one removes blood from a venous reservoir and transfers it through the MO to an arterial reservoir where it can be pumped by a pulsatile pump back into the patient's arterial system.  (Modified from Kirson LE, Laurnen ME, Tornbene MA. Position of oxygenators in the bypass circuit [Letter]. J Cardiothorac Anesth 1989;3:817–818, with permission.)

Most membrane oxygenators are positioned after the pump because the resistance in most requires blood to be pumped through them (Fig. 5.6B). A venous reservoir receives the venous return and the drainage from the cardiotomy reservoir and serves as the atrium for the systemic pump, which pumps the blood through the oxygenator and into the patient. Some low-resistance membrane oxygenators (e.g., Capiox E) may function adequately by gravity drainage of venous blood and hence are positioned like a bubble oxygenator. For pulsatile bypass with some membrane oxygenator circuits, a second pulsatile pump is placed beyond the membrane oxygenator to avoid the damping that would occur if it were placed proximal to the membrane oxygenator (Fig. 5.6C). This requires inclusion of a second (arterial) reservoir and a bypass line to handle the excess flow of the first (venous) pump, which must run slightly faster than the arterial pump. However, pulsatile flow can be achieved with use of relatively noncompliant membrane oxygenators configured as in Figure 5.6B. If there is no arterial reservoir between the membrane oxygenator and the arterial pump, there is a risk of drawing gas bubbles across the membrane and into the CPB circuit due to pressure changes as the roller rapidly decelerates and accelerates.

Oxygenators require a gas supply system. This requires at least a source of oxygen but usually also air (via an oxygen–air blender for membrane oxygenators) and sometimes carbon dioxide, a flow regulator, and flowmeter. An oxygen analyzer should be incorporated in the gas supply line after the blender and a gas filter and moisture trap. An anesthetic vaporizer may be incorporated in the gas supply line to the oxygenator. In this regard, one must be aware that volatile anesthetic liquids may be destructive to the plastic components of extracorporeal circuits, and hence one must consider the location of these vaporizers and use extreme care when filling them with anesthetic liquid so as not to contaminate any plastic (including tubing) component. When a vaporizer is used, a method of scavenging waste gas from the oxygenator outlet should be provided. When bubble oxygenators are used, gas flow must be initiated before the oxygenator is primed and continued thereafter to avoid back leakage of fluid through the bubble disperser plate, which may degrade its efficiency (130).


Back to Quick Links

Heat exchangers are designed to add or remove heat from the blood, thereby controlling the patient's body temperature. During its flow in the CPB circuit, the blood cools and hence heat must be added to avoid patient cooling. In addition, the patient's temperature is often deliberately lowered and then restored to normothermia before discontinuing CPB.

Although in the past separate heat exchangers were used in extracorporeal circuits, currently they are invariably included as an integral part of the disposable oxygenator. The details concerning the function and performance of blood heat exchangers are discussed in Chapter 4. They are usually located proximal to the gas exchanging section of the circuit to minimize the risk of releasing microbubbles of gas from the blood, which could occur if the blood is warmed after being saturated with gas. An additional risk of heat exchangers is water leakage into the blood path. Although this incident is rare, when it occurs it most often is manifested by the appearance of hemolysis and elevated serum potassium.

A source of hot and cold water, a regulator/blender, and temperature sensors are supplemental requirements of heat exchangers. Although hospital water supply may provide such a source, more frequently a stand-alone water cooler and heater is used. Malfunction of these cooler–heaters is one of the more common incidents during CPB (131). Separate heat exchangers are needed for administration of cardioplegic solution and/or blood for coronary perfusion.


Back to Quick Links

A reservoir is placed immediately before the systemic pump to serve as its "holding tank" or atrium 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. Additionally, it is a place to store excess blood when the heart and lungs are exsanguinated. 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 extracorporeal circuit when full CPB begins, especially in patients who have been in congestive heart failure or have long-standing valvular disease.

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, into which the cardiotomy reservoir empties, and as a ready source of blood for transfusion into the patient. One of its most important functions, however, is to provide time for the perfusionist to act if venous drainage is sharply reduced or stopped, to avoid pumping the CPB system dry and risking massive air embolism.

When a bubble oxygenator is used, the reservoir is placed beyond the oxygenating and defoaming chambers and is usually included as an integral part of the oxygenator. This may be referred to as an "arterial reservoir" (Fig. 5.6A). In this case, venous return and cardiotomy drainage blood enter directly into the oxygenating chamber of the bubble oxygenator; hence, this inlet must be as low as possible to facilitate venous return. With membrane oxygenators, the reservoir is the first component of the extracorporeal circuit, directly receiving the venous drainage and the cardiotomy drainage (Fig. 5.6B). Blood then passes through the systemic pump and then through the membrane oxygenator. However, this reservoir (if hard-shelled and open) may be physically attached to the membrane oxygenator housing.

Reservoirs may be rigid (hard-shell) plastic canisters (referred to as "open") or soft collapsible plastic bags (referred to as "closed"). Hard-shell reservoirs have the advantages of making it easier to measure volume, handling venous air more effectively, often having a larger capacity and being easier to prime, and permitting application of suction for vacuum-assisted venous return. Some hard-shell venous reservoirs incorporate macro- and microfilters and can also serve as the cardiotomy reservoir by directly receiving suctioned and vent blood. The soft bag 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. Schonberger et al. (132) observed more blood activation, blood loss, and blood administration with use of hard-shell reservoirs compared with collapsible venous reservoirs, which they attributed to additional exposure to integral filters and the blood–air interface. Another limitation of hard-shell reservoirs is that their defoaming elements are coated with silicone compounds (antifoam) that may cause systemic microembolization (133).


Back to Quick Links

These topics are discussed in Chapter 6.


Back to Quick Links

Gross and microembolic material are ever present during CPB (134,135) and are fully discussed in Chapter 16. Multiple strategies have been suggested to reduce the hazards of embolization, but the most obvious is the use of micropore filters (136–138). Two types of micropore filters are available. A depth filter consists of packed fibers or porous foam. The predominant example of this is the Swank Dacron wool filter. It has no defined pore size, but presents a tortuous large wetted surface that filters by impaction and absorption. Screen filters are usually made of woven polymer thread that has a defined pore size and filters by interception, although the smallest pore screen filters (0.2 to 5.0 m, used for prebypass filtration) are made of membranes. Screen filters vary in pore size and configuration. They not only block particulate emboli but also gross and microscopic air emboli. The latter is accomplished because the pores are filled with liquid that is maintained by surface-active forces. Excessive pressure can overcome this barrier by exceeding the so-called bubble point pressure (139).

Several studies have compared the performance of various micropore filters designed for cardiotomy (140,141) or arterial lines (142–147), and most have found the Dacron wool (depth) filter to be the most effective (140,141,144,145,147). Gourlay et al. (143–145) studied 13 commercially available arterial line filters and found all to have a similar pressure drop (24 to 34 mm Hg at a flow of 5 L/min) but to exhibit variable degrees of hemolysis and platelet loss and handling of gross and microscopic air. These findings did not appear to be related to pore size or type of material, except that again the Dacron wool (Swank) was best at removal of both microscopic and gross air. However, all filters tested were vastly superior to no filter in regards to interdiction of systemic line microemboli. Other authors expressed concern that the Dacron wool filter might cause significantly more hemolysis and thrombocytopenia and develop channeling and saturation breakthrough (136). Gourlay et al. (143) did not note excessive hemolysis, and Ware et al. (141) determined that although platelet counts were lower after the use of Dacron wool cardiotomy filters, the number of functional platelets were the same (i.e., the screen filter allowed more dysfunctional platelets to pass). Some concern has also been expressed that nylon screen filters may activate complement (136,137). Heparin-coated arterial line micropore filters have been introduced to reduce platelet aggregation and loss and facilitate debubbling and priming (148). However, studies have shown that the heparin-benzalkonium coating may leach off the screen during priming, rendering it ineffective or increasing the risk of passage of microscopic air (149).

Micropore filters may be used in several locations in the extracorporeal circuit. A survey conducted in 1993 found their frequency of use in various locations as follows: arterial line, 92%; cardiotomy reservoir, 89%; gas flow line to the oxygenator, 83%; blood product administration sets, 80%; prebypass (i.e., after priming but before connection of the circuit to the patient), 78%; and cardioplegia delivery line, 46% (150).

Most commercial cardiotomy reservoirs now contain an integrated micropore filter. Because the cardiotomy suction is a major source of microemboli (151) and because a micropore filter is more effective if placed in the cardiotomy reservoir line than arterial line (142), this would appear to be a reasonable practice. With the demonstrated presence of various foreign particulates in the disposables used for extracorporeal circuits, it also seems reasonable to use a prebypass filter during priming. The need for micropore filters on the cardioplegia delivery system has been questioned (152).

Limitations of arterial line microfilters include that they add to the cost, may obstruct, are harder to de-air (and therefore may be a source of gaseous microemboli), may generate microemboli, and cause hemolysis and platelet loss and complement activation. However, their wide use and available studies suggest little adverse effect from their use (153). Micropore arterial filters (Fig. 5.7) are excellent gross bubble traps, and on this basis alone, their routine use can probably be justified. If they are used for this purpose, however, they must have (unless they are self-venting) a continuously open purge line, which includes a one-way valve, that goes from the filter to the cardiotomy or venous reservoir to allow escape of trapped air. It is also recommended by filter manufacturers that a bypass line should be incorporated around the filter, which is clamped but can be opened in case the filter becomes obstructed.

FIG 5.7. Arterial line filters and bubble traps. A: Conventional adult arterial line microfilter and bubble trap. Blood enters tangentially at the top (left), which encourages any possibly entrained bubbles to rise to the top where they are vented out through a continuous purge line connecting the three-way stopcock to the cardiotomy or venous reservoir. Blood then passes through a screen microfilter (20- to 40-m pore size), which also serves as a barrier to the passage of gaseous microemboli. B: Arterial line bubble trap. The design and flow dynamics are similar to the arterial filter, but blood only passes through a coarse screen strainer (approximately 170-m pore size).

The use of leukocyte-depleting filters is receiving much attention. Activation of leukocytes is thought to be a major contributor to the inflammatory response to CPB and postischemic injury. Active removal of leukocytes and platelets with a separate cell separator throughout CPB seems to have beneficial clinical effects (154,155). The benefits of using leukocyte-depleting filters during CPB are less clear. Animal studies have been encouraging (156), but most clinical studies have failed to document significant leukocyte depletion or clinical benefits (157,158). In contrast, Suzuki et al. (159) noted reduced cardiac enzyme levels when a leukocyte removal filter was used in the blood cardioplegia delivery circuit. Gu et al. (160) also observed clinical benefits from leukocyte filtration of residual pump blood, and Gott et al. (161) reported reduced hospital length of stay and cost savings with aggressive use of multiple leukocyte filters (arterial line, cardioplegia delivery line, autotransfused blood, and all blood product administration lines). However, this benefit was only noted in low-risk patients and despite the fact that no difference was observed in morbidity or mortality. Thus, further research is required to define the role, if any, of leukocyte filtration during CPB.


Back to Quick Links

Hemoconcentrators (also referred to as hemofilters or ultrafiltration devices) contain semipermeable membranes (typically hollow fibers) that permit passage of water and electrolytes out of the blood. They are used in lieu of diuretics to remove excess fluid or electrolytes (e.g., potassium) and to raise the hematocrit of the perfusate. They can be connected to the CPB circuit in several different configurations. Blood may be drawn from the venous line, the arterial or venous reservoirs, or the systemic flow line, and filtered blood may be returned to the venous line or the cardiotomy or venous reservoirs. Except when blood is taken from a high pressure port or line (e.g., the systemic flow line), a pump must be used to propel blood through the device and may be used to control flow even if blood comes from the systemic flow line. Pressure is generated within the blood channels by resistance to flow through the hollow fibers or by placing a partially occluding clamp downstream. Suction may or may not be applied to the plasma water side of the membrane to facilitate filtration.

Fluid removal can be as great as 180 mL/min (at a flow of 500 mL/min) (162) but more often is in the range 30 to 50 mL/min (163). Various types of membranes are used. Molecules up to a molecular weight of 20,000 Da are re-moved. At least some heparin is removed, and thus adequacy of heparinization must be monitored. Because of blood viscosity considerations, caution should be exercised in raising the hematocrit excessively during hypothermia.

The hemoconcentrator can be used after CPB to concentrate the pump blood before it is given back to the patient either via a bag or pumped directly into a venous line. Advantages of their use compared with centrifugal cell washers (e.g., Cell Saver) are that they conserve platelets, albumin, and coagulation factors (162) and are cheaper. However, hemoconcentrated blood still contains heparin that may need to be neutralized after infusion. Other potential adverse consequences are additional entrance into the circuit, potential for complement activation, retention of free hemoglobin and proteolytic enzymes (e.g., polymorphonuclear elastase), and loss of heparin, and excessive rise in hematocrit. However, no adverse effects were encountered by Boldt et al. (162). These authors compared six different hemofiltration devices with a conventional cell salvage device and identified some significant differences. Compared with diuretics, hemoconcentrators are more easily controlled and do not cause excessive potassium loss (164). See Chapter 7 for a further discussion of these systems.

Modified ultrafiltration (MUF) refers to a practice of withdrawing blood from the patient (usually out of the arterial cannula), after weaning from CPB, and passing it through a hemoconcentrator and pumping it back into the patient via the venous cannulas. It is primarily used in pediatric cases where, because of smaller circulating volumes, the benefits are increased (165). In a survey conducted in 1996, about 44% of pediatric cardiac surgery programs in North America used MUF (166). About 41% used it on all pediatric cases, whereas 45% based it on the patient's weight (about one-half of the respondents chose a weight of less than 10 or 15 kg). In most centers, the arteriovenous configuration is used. Venovenous MUF carries less risk of air cavitation and hemodynamic instability but may be associated with recirculation of filtered blood and does not deliver oxygenated blood to the pulmonary vasculature.

Some groups advocate use of the blood cardioplegia system as the MUF circuit, removing the crystalloid cardioplegia tubing from the cardioplegia roller head, flushing the cardioplegic solution out with CPB circuit blood, and inserting a primed hemoconcentrator distal to the blood cardioplegia pump and proximal to the blood cardioplegia system (167). The advantage of using this system is that it includes a roller pump, bubble trap, heat exchanger, and line pressure monitor and is already connected to the arterial line. It does carry the risk of accidental infusion of cardioplegic solution if not adequately flushed before MUF. Naik et al. (168) and Sutton (169) described other methods of setting up MUF circuits.

In the aforementioned survey (166), technical complications that were encountered in 84% of centers included air cavitation, patient cooling, circuit disruption, unintended infusion of cardioplegic solution, clotted circuit, and transient exsanguination. Because blood is typically aspirated from the arterial cannula by the MUF pump, air cavitation and air aspiration (from around the arterial cannulation site or through the membrane oxygenator) may occur if excessive negative pressure develops due to occlusion of the arterial cannula tip or kinking of the withdrawal line. Because of the risk of air entering the arterial circuit, antegrade flow should not be permitted once MUF has begun. For this reason, the arterial pump flow rate (which may be used to add volume from the pump oxygenator to the MUF circuit) should never exceed the MUF pump flow rate. Strategies to reduce air entry and cavitation developing in the arterial line (due to developing negative pressure) include extra vigilant monitoring of the CPB systemic line pressure, use of pressure servoregulation to stop the MUF pump if negative pressure develops in the arterial line, bypassing the arterial line filter, and having the surgeon manually maintain optimal aortic cannula position. Other safety measures include use of a bubble trap and bubble detector, inverting the hemoconcentrator so that blood enters the top, and monitoring the pressure in the MUF return line beyond the MUF pump. This topic is discussed further in Chapter 30.


Back to Quick Links

Overall monitoring of the patient during CPB is covered in Chapter 27. This discussion focuses on ancillary devices used to monitor CPB circuit performance.

In-line blood-gas monitors

Noninvasive flow-through devices are available to measure blood gases and other electrolytes and the hemoglobin/hematocrit in the arterial and venous lines (170–172). The arterial monitor is similar to a pulse oximeter by providing continuous assessment of arterial oxygenation and permits more rapid and precise control of blood gases (173,174). Venous oximetry permits rapid assessment of the balance of oxygen supply and demand (175–177). Rubsamen (178), a medicolegal scholar, asserted that in-line blood-gas monitoring is a standard of care essential both from the standpoint of patient safety and prevention of devastating malpractice suits. Unfortunately, little scientific data document that this approach is superior or more safe to the visual monitoring of arterial and venous lines plus periodic discrete blood-gas sampling. Further, currently available devices are imperfect and may provide misleading information (179). For example, an apparently adequate mixed venous oxygen saturation may give a false sense of security (180). Finally, economic considerations cannot be ignored because in-line sensors add significant cost to the setup of the CPB circuit.

Another approach at more timely monitoring of blood gases and various electrolytes is the use of automated operating room analyzers (181,182), but the perfusion team or other individuals in the operating room must then accept responsibility for quality control and recognize their limitations (171,181). One must also be aware that blood drawn from the sampling port of some oxygenators may give misleading information (183). Stammers (184) summarized the clinical evaluation studies of various in-line and off-line blood-gas and electrolyte monitoring devices.

Systemic line pressure

The pressure in the CPB systemic flow line, measured after the pump but before the arterial line filter, should be monitored continuously to detect obstruction in the arterial line, malposition of the arterial cannula, arterial dissection, or obstruction of the filter. This pressure must be interpreted in the context of the expected pressure drop across the arterial cannula at the indicated flow rate and the patient's monitored intraarterial pressure (4). It is also a useful guide to verify proper initial arterial cannulation and as a monitor of central arterial pressure in the early postbypass period, when the radial artery pressure may be misleading. It is desirable to include an audible alarm in this systemic pressure monitoring system to alert the perfusionist if pressures are higher than expected. Many also connect this pressure site to the systemic pump so that it is servocontrolled to stop the pump if excessive pressures develop.

Systemic line flowmeter

A systemic line flowmeter is necessary if a centrifugal pump is being used and is a helpful adjunct to confirm proper occlusiveness and calibration when a roller pump is used. Akers et al. (185) compared four different extracorporeal blood flowmeters and the influence of changing hematocrit on measured values. The differences were small and do not appear to be clinically significant.

Oxygen concentration

An oxygen analyzer should be placed in the gas supply line to the oxygenator. This is essential if oxygen is being diluted with another gas. The analyzer should be placed downstream from the blender (186). Monitoring gas flow into the oxygenator is also desirable. Kirson and Goldman (187) described a comprehensive system for monitoring the delivery of ventilating gas to the oxygenator.

Expired gas

Monitoring the concentration of oxygen, carbon dioxide, and anesthetic vapors exiting the oxygenator ensures that oxygen is passing through the oxygenator and provides information regarding metabolic activity and depth of anesthesia (187). Zia et al. (188) found a fair correlation of exhaust gas CO2 from bubble oxygenators with arterial blood CO2 during cooling and hypothermia but a poor correlation during warming. With membrane oxygenators, another group found no correlation between exhaust gas and arterial blood CO2 during all phases of CPB (189).

Pressure gradient across membrane oxygenators

Some manufacturers recommend monitoring the pressure gradient across membrane oxygenators because development of excessive pressure gradients has been reported to be the most frequent manifestation of oxygenator dysfunction (190,191) and may be an early indicator of possible oxygenator failure. If the gradient is not being monitored but a centrifugal systemic pump is being used, a high pressure gradient through the membrane should be suspected when systemic flow rate is unexpectedly low for the rpm setting and the CPB systemic line or patient arterial pressure.

Temperature monitoring

Monitoring water temperature entering the heat exchangers (oxygenator and cardioplegia delivery system) ensures adequate cooling and avoids excess heating and potentially hazardous cooling and warming gradients (192). Knowledge of the temperature of the venous blood draining from the patient provides information regarding the adequacy of cooling and rewarming when inducing hypothermia and restoring normothermia, respectively. With renewed interest in the possible role of cerebral hyperthermia as a contributor to cerebral dysfunction after CPB, monitoring the temperature of the perfusate being delivered to the patient from the CPB circuit is considered by many to be essential for the safe conduct of CPB.

Low-level sensor and bubble detector

A low-level sensor with alarms on the CPB reservoir and a bubble detector on the systemic flow line are considered desirable safety devices. Whether they should be connected to the systemic pump to automatically turn off the pump is the subject of debate (faster response time versus risk of false alarms). Furthermore, some reservoirs and oxygenators in current use preclude the application of the low-level alarms, emphasizing the essential role of a vigilant perfusionist.


Epiaortic and TEE can be extremely helpful in conducting CPB. Besides their role in evaluating the aorta before arterial cannulation, TEE is accurate in detecting atherosclerosis in the descending aorta that might influence the use of femoral artery cannulation. It can also detect clots or tumors in the left atrium, left atrial appendage, or left ventricle and in the right-sided chambers that could influence venous cannulation and left-sided venting. Detection of a patent foramen ovale using two-dimensional images, color flow Doppler, and agitated saline echocontrast may anticipate a source of venous air if the left heart is to be opened, and detection of a patent ductus arteriosus may explain excessive return of blood to left heart during CPB (193).

Evaluations of the degree of aortic regurgitation (194) and the presence of a dilated coronary sinus (29) have impact on the administration of antegrade and retrograde cardioplegia. TEE may provide information regarding the presence of a persistent LSVC, which has an obvious impact on venous cannulation and administration of retrograde cardioplegia.

During introduction of various cannulas, TEE can evaluate positioning of peripherally introduced systemic venous cannulas, long arterial cannulas introduced into ascending aorta, retrograde coronary sinus cannulas, and left-sided venting cannulas. This is particularly helpful when the surgeon cannot directly inspect or palpate the heart due to limited access or adhesions. Use of TEE is essential to placement of various cannulas during port-access surgery and is helpful in proper placement and assessing proper function of intraaortic balloon pumps, especially when they are placed antegrade through the ascending aorta.

During bypass, TEE also can be used to assess the degree of left ventricular distension and decompression and assist with the differential diagnosis of ascending aortic hematoma and the diagnosis of aortic dissection. This is particularly important during femoral artery cannulation for systemic perfusion. During partial left heart bypass (femoral vein-to-femoral artery and left atrium-to-distal aorta or femoral artery) it can be helpful in assessing the balance of flow between the upper and lower body and blood volume regulation. Finally, it is useful in detecting intracardiac air and assessing adequacy of de-airing after cardiac surgery (195,196).

Automated data collection

Automatic data collection systems are available to assist with the preoperative calculations and to process and store data during CPB, which should free up the perfusionist to attend to more important tasks (197). Whether this will indeed improve outcome remains to be demonstrated (174,198).


Back to Quick Links

Traditionally, CPB is conducted with the heart approached through a median sternotomy, which provides excellent exposure for venous and arterial cannulation. Approaching the heart through a right or left thoracotomy, a limited sternotomy (limited to upper portion or lower portion), or initiating CPB before exposing the heart presents special cannulation problems. These alternative approaches may be used to avoid sternotomy due to previous sternotomy, especially if complicated by prior infection or aneurysms of the ascending aorta or right ventricle and to avoid patent coronary grafts that might pass beneath the sternum (e.g., right internal mammary artery to left coronary circulation). Alternative approaches also may be appropriate for focused access to specific coronary arteries (e.g., right or circumflex) (199), for surgery involving the descending aorta, or when circulatory support is required before entering the chest (e.g., severe cardiac failure or problems encountered during attempted reentry) (20). These alternative approaches are also being used for minimal access or so-called minimally invasive cardiac surgery (Fig. 5.8)

FIG 5.8. Cannulas used for minimally invasive cardiopulmonary bypass. A: Smaller (typically, 29 to 29 French), nontapered, two-stage, wire-wound venous cannula that is inserted into the right atrium (RA) and inferior vena cava (IVC). B: Two-stage wire-wound venous cannula with a flattened section in the central portion of the cannula. This is inserted into the RA and IVC through a small chest incision. The cross-sectional area of the flattened portion of the cannula permits the same venous drainage as a round-shaped venous cannula but allows the surgeon to optimally position the cannula and avoid kinking as it exits the chest. C: Long, wire-wound, diffusion-tipped arterial cannula placed in the aorta. The fixation ring around the circumference near the tip is movable, and the wires in the wall allow the surgeon to position the cannula away from the surgical field.

Cannulation through a right thoracotomy

This approach poses little problem with venous (atrial) cannulation but provides poor access to the ascending aorta and no access to the left ventricle. Peters et al. (24) evaluated various options for bicaval cannulation during minimal access right thoracotomies. Aortic cannulation and cross-clamping can be difficult as is de-airing the left ventricle. Arterial cannulation may be more conveniently accomplished using the femoral or axillary arteries as discussed earlier. Femoral-femoral bypass with deep hypothermia has been advocated by others (200). Placing external defibrillator pads on the left chest (front and back over the heart) may facilitate defibrillation if needed.

Cannulation through left thoracotomy

One can conduct partial or complete left heart bypass and partial or complete CPB through the left chest.

Partial left heart bypass

This method relies on the patient's right heart to pump blood through the lungs to the left heart and the patient's lungs to provide gas exchange. Blood is usually removed from the left heart with a large venous cannula placed directly into the left atrium. This carries a risk of systemic air embolism if one is not careful during insertion and removal of this cannula (the left heart should be well filled and ventilation momentarily interrupted during these procedures). Also, if excessive suction is applied to the venous line, especially if the atrial purse-string is not secure, air can be entrained into the left heart and venous cannula (and hence pumped into the systemic circulation). Another problem is that improper positioning or movement of the tip of the left atrial cannula can impair drainage of the left heart and flow into the left heart bypass circuit. Another way of draining blood from the left heart is by placing a cannula into the apex of the left ventricle. This may provide excellent flow of blood into the left heart circuit but adds the risks of injury to the left ventricle, bleeding, and late aneurysm formation. Arterial return may be into the descending aorta or femoral artery (usually the left, because the patient is in the right lateral decubitus position).

The extracorporeal circuit typically consists only of tubing and a centrifugal pump and does not include a reservoir, heat exchanger, or bubble trap. This minimizes the need for heparinization but precludes the ability to add or sequester fluid, adjust temperature, or prevent systemic air embolization. Some have advocated including a heat exchanger in the circuit to restore normothermia (201).

The anesthesiologist must be prepared to administer intravascular volume expanders as needed through other sites of the vascular access. Using a rapid infusion system connected to a large-bore intravenous site can facilitate volume administration.

Often this technique is used during surgery on the descending thoracic aorta. The patient's left ventricle supplies blood to the aorta proximal to the clamps and the left heart bypass circuit supplies blood to the body distal to the clamps. Typically, about two thirds of normal basal cardiac output (i.e., two thirds of 2.4 L/min/m2 or about 1.6 L/min/m2) is pumped to the lower body and what is left is pumped by the left ventricle to the upper body. Arterial pressure should be measured in the proximal (right radial or brachial artery) and in the distal aorta via the right femoral artery. This will permit assessment of adequacy of flow to each part of the body, whereas the central venous and pulmonary artery diastolic or occlusion pressure will assess filling of the ventricles.

Use of TEE is particularly helpful in assessing filling and function of the left heart and left atrial cannula placement and function. If the arterial pressure is too high proximally and too low distally, the left heart bypass circuit should pump more blood. Conversely, if the arterial pressure is too low proximally and too high distally, the left heart bypass circuit should pump less blood. If both pressures are low, blood volume, and possibly a vasoconstrictor, may need to be added. If both pressures are high, more anesthesia and possibly a vasodilator should be added. O'Connor and Rothenberg (202) reviewed this topic in greater detail.

Full left heart bypass

This can be accomplished with the just described cannulation technique for left-sided coronary artery surgery. If the heart fibrillates, blood can still passively pass through the right heart and lungs, but often an elevated central venous pressure is required.

Partial cardiopulmonary bypass

This technique is used to facilitate descending aortic surgery. The major issue is how to obtain systemic venous drainage. Several options are available: through the pulmonary artery with the cannula tip left in the main pulmonary artery or threaded retrograde into the right ventricle (203), through the right ventricular outflow tract, directly into the RA through the right atrial appendage or via the femoral vein. The problem with the latter is achieving adequate venous drainage through long small catheters, but this is usually solved by use of thin-walled cannulas threaded up into the RA with proper positioning facilitated by TEE. Augmented venous return can be used if gravity drainage is inadequate. Direct cannulation of the RA is quite difficult with risk of compression of the heart, tearing of the atrium, and bleeding. Cannulation of the pulmonary artery is somewhat easier.

The extracorporeal circuit resembles that previously described for full CPB and includes a pump, reservoir, oxygenator, heat exchanger, and bubble trap. It is important to remember that the left ventricle is still the source of blood supply to the upper part of the body, and the heart and the patient's own lungs are providing gas exchange for that blood. Thus, the principles of relative flows to the lower part of the body and the upper part of the body described under partial left heart bypass apply. Furthermore, the adequacy of gas exchange of blood going to the upper body must be assessed separately (from samples drawn from the radial or brachial artery) from that coming out of the oxygenator and going to the lower body.

Full cardiopulmonary bypass

The only difference from what has just been discussed is that virtually all systemic venous return must be drained from the right heart, which accentuates the demand on venous cannulation and drainage. The CPB circuit supplies all systemic flow and gas exchange. Ascending aortic cross-clamping and administration of antegrade cardioplegia can be problematic. Sasaguri et al. (204) described a special double-lumen balloon catheter that is inserted through the left ventricular apex into the ascending aorta with TEE guidance for occlusion of the ascending aorta and administration of cardioplegic solution during thoracic aortic surgery via a left thoracotomy.


Back to Quick Links

Heparin coating of circuits is used in an effort to improve biocompatibility and hopefully to reduce inflammatory reactions and heparinization requirements and thus improve outcome (205–207). Other surface modification methods (e.g., surface modifying additive) are being introduced (208). This topic is discussed in Chapter 9.

Circuits for CPB in children, and especially small infants, pose great challenges in terms of cannulation and miniaturization (209,210). These issues are discussed in Chapter 30.

Retrograde cerebral perfusion was first introduced as a method to treat massive air embolism (211) but is now widely used as a cerebroprotective strategy during aortic surgery and deep hypothermic circulatory arrest. This presents important issues in cannulation, monitoring, and setup of the extracorporeal circuit, which are discussed in Chapter 34.

Limited access minimally invasive surgery and, particularly, use of the HeartPort system imposes unique problems in cannulation (212). These have been covered partially in the sections on venous and arterial cannulation, augmented venous drainage, and unusual cannulation but are discussed additionally in Chapter 35. Computerized control of CPB has been described and used clinically (213).


Back to Quick Links

We gratefully acknowledge the artwork of Norman P. Pregent, C.C.P., and Steve Schuenke.


Back to Quick Links

  • The primary purpose of CPB is to facilitate cardiac surgery.

  • Major components include the membrane oxygenator (for gas exchange) with integral heat exchanger (for temperature regulation) and the systemic blood flow pump (roller or centrifugal type) for whole body perfusion.

  • Other functions of the CPB circuit include suctioning of blood and air from the operative site and cardioplegia delivery.

  • Venous drainage is most often accomplished by gravity siphonage via large-bore cannulas inserted in the RA/IVC or SVC and IVC.

  • Cavoatrial cannulation is simpler and provides good drainage of both cavae and the right heart.

  • Peripheral venous cannulation is facilitated by use of cannulas advanced to the RA and augmented venous assistance (kinetic or vacuum).

  • Use of augmented venous drainage requires proper cannula placement and careful regulation of the degree of vacuum applied to the venous line.

  • A persistent LSVC is seen infrequently but when present can negatively affect placement of a pulmonary artery catheter, delivery of cardioplegia, and right heart decompression.

  • Arterial cannulas should be chosen so that the pressure drop is less than 100 mm Hg at full CPB flow.

  • The risk of complications with femoral arterial cannulation is much higher than with ascending aortic cannulation.

  • Use of either long or diffusion-tipped arterial cannulas can minimize the risk of dislodgment of atheroma in the ascending or transverse aorta.

  • Dissection of the aorta (either retrograde or antegrade) must be promptly recognized and surgically corrected to decrease the risk of patient morbidity or mortality.

  • Tubing sizes and lengths and connectors should be chosen to minimize blood velocity and priming volume.

  • CPB reservoirs may be either hard-shell ("open") or soft collapsible ("closed") types, but there appears to be no advantage of one over the other unless vacuum-assisted venous return is used, which requires a hard-shell type.

  • Hemoconcentrators can be used during and after CPB to remove plasma water and raise the hematocrit and are more cost effective than cell salvage and washing devices.

  • Alternative thoracotomies and cannulation sites are being used for minimally invasive cardiac surgery, and newer cannulas have become available to facilitate this approach.

  • TEE is a useful adjunct to CPB for assessing the aorta before cannulation, cannula placement, determining cardiac and valvular function, and evaluating the effectiveness of cardiac decompression and de-airing maneuvers.


Back to Quick Links

    1. Peirce EC II. Extracorporeal circulation for open-heart surgery . Springfield, IL: Charles C Thomas, 1969.

    2. Kirklin JWBarratt-Boyes BE. Cardiac surgery , 2nd ed. New York: Churchill-Livingstone, 1993.

    3. Riley JB, Hardin SB, Winn BA, et al. In vitro comparison of cavoatrial (dual stage) cannulae for use during cardiopulmonary bypass. Perfusion 1986;1:197–204.

    4. Delius RF, Montoya JP, Merz SJ, et al. A new method for describing the performance of cardiac surgery cannulas. Ann Thorac Surg 1992;33:278–281.

    5. Bloom DF, Cornhill JF, Malchesky PS, et al. Technical and economic feasibility of reusing disposable perfusion cannulas. J Thorac Cardiovasc Surg 1997;114:448–460.

    6. Sadeghi AM, Rose EA, Michler RE, et al. A simplified method for the occlusion of the venae cavae during cardiopulmonary bypass. Ann Thorac Surg 1986;41:678.

    7. Cooley DA. Caval occlusion clamps for temporary cardiopulmonary bypass. J Thorac Cardiovasc Surg 1970;59:292.

    8. Al-Ebrahim KE, El-Shafei H. Cuffed venous return cannulas in minimally invasive cardiac operation [Letter]. Ann Thorac Surg 1998;65:1509.

    9. Phillips SJRomanowski E. A new designed venous cannula for cardiopulmonary bypass. J Thorac Cardiovasc Surg 1972;63:769–770.

    10. Kirsh MM, Lemer JHZwischenberger JB. Rapid technique of occlusion of the venae cavae for total cardiopulmonary bypass during repeat cardiac operations. Ann Thorac Surg 1987;43:566–567.

    11. Morritt GN, Holden MP. The cuffed endotracheal tube in emergency cardiopulmonary bypass operations. Ann Thorac Surg 1981;31:287–288.

    12. Arom KV, Ellestad C, Grover FL, et al. Objective evaluation of the efficacy of various venous cannulas. J Thorac Cardiovasc Surg 1981;81:464–469.

    13. Bennett EV Jr, Fewel JG, Ybarra J, et al. Comparison of flow differences among venous cannulas. Ann Thorac Surg 1983;36:59–65.

    14. Bennett EV Jr, Fewel JG, Grover FL, et al. Myocardial preservation: effect of venous drainage. Ann Thorac Surg 1983;36:132–142.

    15. Lake CL. Controversies in the management of cardiopulmonary bypass. In: Kaplan JA ed. Cardiothoracic and vascular anesthesia update . Vol. 1. Philadelphia: W.B. Saunders Co., 1990:1–21.

    16.  Casthely PA , Bregman D eds. Cardiopulmonary bypass: physiology, related complications, and pharmacology . Mount Kisco, NY: Futura Publishing Co., 1991.

    17. Rosenfeldt FL, Watson DA. Interference with local myocardial cooling by heat gain during aortic cross-clamping. Ann Thorac Surg 1979;27:13–16.

    18. Lawrence DR, Desai JB. Forty-five degree, two-stage cannula: advantages over standard two-stage venous cannula. Ann Thorac Surg 1997;63:253–254.

    19. Taylor PC, Effler DB. Management of cannulation for cardiopulmonary bypass in patients with adult-acquired heart disease. Surg Clin North Am 1975;55:1205–1215.

    20. Merin O, Silberman S, Brauner R, et al. Femoro-femoral bypass for repeat open-heart surgery. Perfusion 1998;13:455–459.

    21. Jones RE, Fitzgerald D, Cohn LH. Reoperative cardiac surgery using a new femoral venous right atrial cannula. J Card Surg 1990;5:170–173.

    22. Westaby S. Extrathoracic cannulation for urgent cardiopulmonary bypass in cardiac tamponade: use of internal jugular vein. J Cardiovasc Surg 1988;29:103–105.

    23. Flege JB Jr, Wolf RK. Venous drainage to the heart-lung machine via the internal jugular vein. Ann Thorac Surg 1997;63:861.

    24. Peters WS, Stevens JH, Smith JA, et al. Minimally invasive right heart operations: techniques for bicaval occlusion and cardioplegia. Ann Thorac Surg 1997;64:1843–1845.

    25. Choudhry AK, Conacher ID, Hilton CJ, et al. Persistent left superior vena cava. J Cardiothorac Anesth 1989;3:616–619.

    26. Harris AM, Shawkat S, Bailey JS. The use of an endotracheal tube for cannulation of left superior vena cava via coronary sinus for repair of a sinus venosus atrial septal defect. Br Heart J 1987;58:676–677.

    27. Horrow JC, Lingaraju N. Unexpected persistent left superior vena cava: diagnostic clues during monitoring. J Cardiothorac Anesth 1989;3:611–615.

    28. Winter FS. Persistent left superior vena cava: survey of world literature and report of thirty additional cases. Angiology 1954;5:90–132.

    29. Hasel R, Barash PG. Dilated coronary sinus on pre-bypass echocardiography. J Cardiothorac Vasc Anesth 1996;10:430–435.

    30. Shahian DM. Retrograde coronary sinus cardioplegia in the presence of persistent left superior vena cava. Ann Thorac Surg 1992;54:1214–1215.

    31. Yokota M, Kyoku I, Kitano M, et al. Atresia of the coronary sinus orifice: fatal outcome after intraoperative division of the drainage left superior vena cava. J Thorac Cardiovasc Surg 1989;98:30–32.

    32. Babka RM. A comparison of the use of venous pumping to gravity return of blood to the oxygenator during cardioplegic arrest. Proc Am Acad Cardiovasc Perfus 1988;9:47–50.

    33. Toomasian JM, McCarthy JP. Total extrathoracic cardiopulmonary support with kinetic assisted venous drainage: experience in 50 patients. Perfusion 1998;13:137–143.

    34. Taketani S, Sawa Y, Massai T, et al. A novel technique for cardiopulmonary bypass using vacuum system for venous drainage with pressure relief value: an experimental study. Artif Organs 1998;22:337–341.

    35. Darling E, Kaemmer D, Lawson S, et al. Experimental use of an ultra-low prime neonatal cardiopulmonary bypass circuit utilizing vacuum-assisted venous drainage. J Extra-Corp Tech 1998;30:184–189.

    36. Fried DW, Zombolas TL, Weiss SJ. Single pump mechanically aspirated venous drainage (SPMAVD) for cardiac reoperation. Perfusion 1995;10:327–332.

    37. Wallock M, Kuehn B, Hoff W. Single pump mechanically aspirated venous drainage (SPMAVD) for cardiac surgery [Letter]. Perfusion 1996;11:351–352.

    38. Fried DW, Zombolas TL, Weiss SJ. Authors' response. Perfusion 1996;11:352–353.

    39. Bernstein EF, Gleason LR. Factors influencing hemolysis with roller pumps. Surgery 1967;61:432–442.

    40. Indeglia RA, Shea MA, Varco RE, et al. Mechanical and biologic considerations in erythrocyte damage. Surgery 1967;114:126–138.

    41. Ratnaraj J, Manohar G. Pulmonary artery catheter displacement during cannulation for CPB [Letter]. J Cardiothorac Vasc Anesth 1991;5:648.

    42. Troianos CA. Transesophageal echocardiographic diagnosis of pulmonary artery catheter entrapment and coiling. Anesthesiology 1993;79:602–604.

    43. Ambesh SP, Singh SK, Dubey DK, et al. Inadvertent closure of the superior vena cava after decannulation: a potentially catastrophic complication after termination of bypass [Letter]. J Cardiothorac Vasc Anesth 1998;12:723–724.

    44. Herrema IHWinsser LJA. Flow directed pulmonary artery catheter obstructs venous drainage cannula of cardiopulmonary bypass machine [Letter]. Anaesthesia 1988;43:799.

    45. Brodman R, Siegel H, Lesser M, et al. A comparison of flow gradients across disposable arterial perfusion cannulas. Ann Thorac Surg 1985;39:225–233.

    46. Galletti PMBrecher GA. Heart-lung bypass, principles and techniques of extracorporeal circulation . New York: Grune &Stratton, 1962:184–188.

    47. Drews JA, Cleveland RJ, Nelson RJ. An approach to aortic cannulation with a caution on hemolysis associated with angled cannulas. Rev Surg 1974;31:57–59.

    48. Groom RC, Hill AG, Kuban B, et al. Aortic cannula velocimetry. Perfusion 1995;10:183–188.

    49. Muehrcke DD, Cornhill JF, Thomas JD, et al. Flow characteristics of aortic cannulae. J Card Surg 1995;10:514–519.

    50. Ringgaard S, Madsen T, Pedersen EM, et al. Quantitative evaluation of flow patterns in perfusion cannulae by a new magnetic resonance imaging method. Perfusion 1997;12:411–416.

    51. Benaroia M, Baker AJ, Mazer D, et al. Effect of aortic cannula characteristics and blood velocity on transcranial Doppler-detected microemboli during cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1998;12:266–269.

    52. Magner JB. Complications of aortic cannulation for open-heart surgery. Thorax 1971;26:172–173.

    53. McAlpine WA, Selman MW, Kawakami T. Routine use of aortic cannulation in open heart operations. Am J Surg 1967;114:831–834.

    54. Davidson KG. Cannulation for cardiopulmonary bypass. In: Taylor KM ed. Cardiopulmonary bypass: principles and management . Baltimore: Williams &Wilkins, 1987:55–89.

    55. Taylor PC, Groves LK, Loop FD, et al. Cannulation of the ascending aorta for cardiopulmonary bypass. J Thorac Cardiovasc Surg 1976;71:255–258.

    56. Mills NL, Everson CT. Atherosclerosis of the ascending aorta and coronary artery bypass: pathology, clinical correlates, and operative management. J Thorac Cardiovasc Surg 1991;102:546–553.

    57. Blauth CI, Cosgrove DM, Webb BW, et al. Atheroembolism from the ascending aorta. J Thorac Cardiovasc Surg 1992;103:1104–1112.

    58. Barbut D, Grassineau D, Lis E, et al. Posterior distribution of infarcts in strokes related to cardiac operation. Ann Thorac Surg 1998;65:1656–1659.

    59. Beique FA, Joffe D, Tousignant G, et al. Echocardiographic-based assessment and management of atherosclerotic disease of the thoracic aorta. J Cardiothorac Vasc Anesth 1998;12:206–220.

    60. Murphy DA, Craver JM, Jones EL, et al. Recognition and management of ascending aortic dissection complicating cardiac surgical operations. J Thorac Cardiovasc Surg 1983;85:247–256.

    61. Davila-Roman VG, Kouchoukos NT, Schechtman KB, et al. Atherosclerosis of the ascending aorta is a predictor of renal dysfunction after cardiac operations. J Thorac Cardiovasc Surg 1999;117:111–116.

    62. Bar-El Y, Goor DA. Clamping of the atherosclerotic ascending aorta during coronary artery bypass operations. J Thorac Cardiovasc Surg 1992;104:469–474.

    63. Ohteki H, Itoh T, Natsuaki M, et al. Intraoperative ultrasonic imaging of the ascending aorta in ischemic heart disease. Ann Thorac Surg 1990;50:539–542.

    64. Wareing THDavilla-Roman VG, Barzilai B, et al. Management of the severely atherosclerotic ascending aorta during cardiac operations: a strategy for detection and treatment. J Thorac Cardiovasc Surg 1992;103:453–462.

    65. Sylivris S, Calafiore P, Matalanis G, et al. The intraoperative assessment of ascending aortic atheroma: epiaortic imaging is superior to both transesophageal and direct palpation. J Cardiothorac Vasc Anesth 1997;11:704–707.

    66. Davila-Roman V, Phillips K, Davila R, et al. Intraoperative transesophageal echocardiography and epiaortic ultrasound for assessment of atherosclerosis of the thoracic aorta. J Am Coll Cardiol 1996;28:942–947.

    67. Konstadt SN, Reich DL, Quintana C, et al. The ascending aorta: how much does transesophageal echocardiography see? Anesth Analg 1994;78:240–244.

    68. Konstadt SN, Reich DL, Kahn R, et al. Transesophageal echocardiography can be used to screen for ascending aortic atherosclerosis. Anesth Analg 1995;81:225–228.

    69. Guzzetta NA, Lee E, Sadel SM, et al. Does the degree of atheromatous disease in the descending aorta correlate with atheromatous disease in the ascending aorta? Anesth Analg 1995;80:SCA 80(abst).

    70. Rokkos CF, Kouchoukos NT. Surgical management of the severely atherosclerotic ascending aorta during cardiac operations. Semin Thorac Cardiovasc Surg 1998;10:240–246.

    71. Grossi EA, Kanchuger MS, Schwartz DS, et al. Effect of cannula length on aortic arch flow: protection of the atheromatous aortic arch. Ann Thorac Surg 1995;59:710–712.

    72. Ribakov GH, Katz ES, Galloway AC, et al. Surgical implications of transesophageal echocardiography to grade atheromatous aortic arch. Ann Thorac Surg 1992;53:758–763.

    73. Stern A, Tunick PA, Culliford AT, et al. Aortic arch endarterectomy increases the risk of stroke during heart surgery in patients with protruding aortic arch atheromas. Circulation 1997;96:I–185(abst 1024).

    74. Byrne JG, Aranki SF, Cohn LH. Aortic valve operations under deep hypothermic circulatory arrest for the porcelain aorta: "no-touch" technique. Ann Thorac Surg 1998;65:1313–1315.

    75. Svensson LG, Sun J, Cruz HA, et al. Endarterectomy for calcified porcelain aorta associated with aortic value stenosis. Ann Thorac Surg 1996;61:149–152.

    76. King RC, Kanithanon RC, Shockley KS, et al. Replacing the atherosclerotic ascending aorta is a high-risk procedure. Ann Thorac Surg 1998;66:396–401.

    77. Liddicoat JR, Doty JR, Stuart RS. Management of the atherosclerotic ascending aorta with endoaortic occlusion. Ann Thorac Surg 1998;65:1133–1135.

    78. Paul D, Hartman GS. Foley balloon occlusion of the atheromatous ascending aorta: the role of transesophageal echocardiography. J Cardiothorac Vasc Anesth 1998;12:61–64.

    79. Duda AM, Letwin LB, Sutter FP, et al. Does routine use of aortic ultrasonography decrease the stroke rate in coronary artery bypass surgery? J Vasc Surg 1995;21:98–107.

    80. Still RJ, Hilgenberg AD, Akins CW, et al. Intraoperative aortic dissection. Ann Thorac Surg 1992;53:374–380.

    81. Garcia-Rinaldi R, Vaughan GD III, Revuelta JM, et al. Simplified aortic cannulation. Ann Thorac Surg 1983;36:226–227.

    82. Salama FD, Blesovsky A. Complications of cannulation of the ascending aorta for open heart surgery. Thorax 1970;25:604–607.

    83. Parker R. Aortic cannulation. Thorax 1969;24:742–745.

    84. Krous HF, Mansfield PB, Sauvage LR. Carotid artery hyperperfusion during open-heart surgery. J Thorac Cardiovasc Surg 1973;66:118–121.

    85. Kulkarni MG. A complication of aortic cannulation. J Cardiovasc Surg 1968;9:207–208.

    86. Dalal FY, Patel KD. Another sign of inadvertent carotid cannulation [Letter]. Anesthesiology 1981;55:487.

    87. Ross WT Jr, Lake CL, Wallons HA. Cardiopulmonary bypass complicated by inadvertent carotid cannulation. Anesthesiology 1981;54:85–86.

    88. Watson BG. Unilateral cold neck. Anaesthesia 1983;38:659–661.

    89. McLeskey CH, Cheney FW. A correctable complication of cardiopulmonary bypass. Anesthesiology 1982;56:214–216.

    90. Magilligan DJ Jr, Eastland MW, Lell WA, et al. Decreased carotid flow with ascending aortic cannulation. Circulation 1972;45[Suppl I]:I-130–I-133.

    91. Salerno TA, Lince DP, White DN, et al. Arch versus femoral artery perfusion during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1978;78:681–684.

    92. Gott JP, Cohen CL, Jones EL. Management of ascending aortic dissections and aneurysms early and late following cardiac operations. J Card Surg 1990;5:2–13.

    93. Troianos CA, Savino JS, Weiss RL. Transesophageal echocardiographic diagnosis of aortic dissection during cardiac surgery. Anesthesiology 1991;75:149–153.

    94. Lees MH, Herr RH, Hill JD, et al. Distribution of systemic blood flow of the rhesus monkey during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1971;61:570–586.

    95. Herman BE, Wallace HW, Gadboys HL, et al. Anterior crural syndrome as a complication of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1966;52:755–758.

    96. Hendrickson SC, Glower DD. A method for perfusion of the leg during cardiopulmonary bypass via femoral cannulation. Ann Thorac Surg 1998;65:1807–1808.

    97. Gates JD, Bichell DP, Rizzu RJ, et al. Thigh ischemia complicating femoral vessel cannulation for cardiopulmonary bypass. Ann Thorac Surg 1996;61:730–733.

    98. Greason KL, Hemp JR, Maxwell JM, et al. Prevention of distal leg ischemia during cardiopulmonary support via femoral cannulation. Ann Thorac Surg 1995;60:209–210.

    99. VanderSalm TJ. Prevention of lower extremity ischemia during cardiopulmonary bypass via femoral cannulation. Ann Thorac Surg 1997;63:251–252.

    100. Sata J, Rimpilainen J, Rainio P, et al. A feasible femoral cannulation method during cardiopulmonary bypass [Letter]. Ann Thorac Surg 1998;65:1194.

    101. Utoh J, Gotto H, Ashimura K, et al. A simple switching technique from cardiopulmonary bypass to a long-term extracorporeal life support system. J Thorac Cardiovasc Surg 1996;112:206–207.

    102. Beyersdorf F, Mitrev Z, Ihnken K, et al. Controlled limb reperfusion in patients having cardiac operations. J Thorac Cardiovasc Surg 1996;111:873–881.

    103. Svensson LG. Editorial comment: autopsies in acute type A aortic dissection, surgical implications. Circulation 1998;98:II-302–II-304.

    104. Benedict JS, Buhl TL, Henney RP. Acute aortic dissection during cardiopulmonary bypass. Arch Surg 1974;108:810–813.

    105. Bigutay AM, Garamella JJ, Danyluk M, et al. Retrograde aortic dissection occurring during cardiopulmonary bypass. JAMA 1976;236:465–468.

    106. Carey JS, Skow JR, Scott C. Retrograde aortic dissection during cardiopulmonary bypass: "nonoperative" management. Ann Thorac Surg 1977;24:44–48.

    107. Kay JH, Dykstra DC, Tsuji HK. Retrograde ilio-aortic dissection. A complication of common femoral artery perfusion during open-heart surgery. Am J Surg 1966;111:464–468.

    108. Matar AF, Ross DN. Traumatic arterial dissection in open-heart surgery. Thorax 1967;22:82–87.

    109. Jones TW, Vetto RR, Winterscheid LC, et al. Arterial complications incident to cannulation in open-heart surgery. Ann Surg 1960;152:969–974.

    110. Reichenspurner H, Gulielmos V, Wunderlich J, et al. Port-access coronary artery bypass grafting with the use of cardiopulmonary bypass and cardioplegic arrest. Ann Thorac Surg 1998;65:413–419.

    111. Galloway AC, Shemin RJ, Glower DD, et al. First report of the Port-Access International Registry. Ann Thorac Surg 1999;67:51–58.

    112. Sabik JF, Lytle BW, McCarthy PM, et al. Axillary artery: an alternative site of arterial cannulation for patients with extensive aortic and peripheral vascular disease. J Thorac Cardiovasc Surg 1995;109:885–891.

    113. Bichell DP, Balaguer JM, Aranki SF, et al. Axilloaxillary cardiopulmonary bypass: A practical alternative to femorofemoral bypass. Ann Thorac Surg 1997;64:702–705.

    114. Baribeau YR, Westerbrook BM, Charlesworth DC, et al. Arterial inflow via an axillary artery graft for the severely atherosclerotic aorta. Ann Thorac Surg 1998;66:33–37.

    115. Whitlark JD, Sutter FP. Intrathoracic subclavian artery cannulation as an alternative to femoral or axillary artery cannulation [Letter]. Ann Thorac Surg 1998;66:303.

    116. Golding LAR. New cannulation technique for the severely calcified ascending aorta. J Thorac Cardiovasc Surg 1985;90:626–627.

    117. Norman JC. A single cannula for aortic perfusion and left ventricular decompression. Chest 1970;58:378–379.

    118. Robicsek F. Apical aortic cannulation: application of an old method with new paraphernalia. Ann Thorac Surg 1991;51:320–322.

    119. Watanabe H, Eguchi S, Miyamura H, et al. Transapical aortic cannulation in pediatric patients. Ann Thorac Surg 1997;63:1149–1150.

    120. Coselli JS, Crawford ES. Femoral artery perfusion for cardiopulmonary bypass in patients with aortoiliac artery obstruction. Ann Thorac Surg 1987;43:437–439.

    121. Stammers AH. Extracorporeal devices and related technologies. In: Kaplan JA ed. Cardiac anesthesia , 4th ed. Philadelphia: W.B. Saunders, 1999:1017–1060.

    122. Reed CCStafford TB. Cardiopulmonary bypass , 2nd ed. Houston: Medical Press, 1985.

    123. Mongero LB, Beck JR, Orr TR, et al. Clinical evaluation of setting pump occlusion by the dynamic method: effect on flow. Perfusion 1998;13:360–368.

    124. Tempe DK, Khanna SK. Accidental hyperperfusion during cardiopulmonary bypass: suggested safety features. Ann Thorac Surg 1998;65:306.

    125. Kurusz M, Shaffer CW, Christman EW, et al. Runaway pump head. J Thorac Cardiovasc Surg 1979;77:792–795.

    126. Uretzky G, Landsburg G, Cohn D, et al. Analysis of microembolic particles originating in extracorporeal circuits. Perfusion 1987;2:9–17.

    127. Hubbard LC, Kletchka HD, Olsen DA, et al. Spallation using roller pumps and its clinical implications. Am SECT Proc 1975;3:27–32.

    128. Kurusz M, Christman EW, Williams EH, et al. Roller pump induced tubing wear: another argument in favor of arterial line filtration. J Extra-Corp Technol 1980;12:49–59.

    129. Kolff J, McClurken JB, Alpern JB. Beware centrifugal pumps: not a one-way street, but a dangerous siphon! [Letter]. Perfusion 1990;5:225–226.

    130. Dickson TA. Hypoxemia after intraluminal oxygen line obstruction during cardiopulmonary bypass [Letter]. Ann Thorac Surg 1990;49:512.

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

    132. Schonberger JPAM, Everts PAM, Hoffman JJ. Systemic blood activation with open and closed venous reservoirs. Ann Thorac Surg 1995;59:1549–1555.

    133. Orenstein JM, Sato N, Arron B, et al. Microemboli observed in deaths following cardiopulmonary bypass surgery: silicone antifoam agents and polyvinyl chloride tubing as source of emboli. Hum Pathol 1982;13:1082–1090.

    134. Pearson DT. Micro-emboli: gaseous and particulate. In: Taylor KM ed. Cardiopulmonary bypass: principles and management. Baltimore: Williams &Wilkins, 1986:313–354.

    135. Butler BD, Kurusz M. Gaseous microemboli: a review. Perfusion 1990;5:81–99.

    136. Berman L, Marin F. Micropore filtration during cardiopulmonary bypass. In: Taylor KM ed. Cardiopulmonary bypass: principles and management . Baltimore: Williams & Wilkins, 1986:355–374.

    137. Marshall L. Filtration in cardiopulmonary bypass: past, present and future. Perfusion 1988;3:135–147.

    138. Joffe D, Silvay G. The use of microfilters in cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1994;8:685–692.

    139. Pascale F. Removal of gaseous microemboli from extracorporeal circulation. Med Instrument 1985;19:70–72.

    140. Solis RT, Horak J. Evaluation of a new cardiotomy blood filter. Ann Thorac Surg 1979;28:487–488.

    141. Ware JA, Scott MA, Horak JK, et al. Platelet aggregation during and after cardiopulmonary bypass: effect of two different cardiotomy filters. Ann Thorac Surg 1982;34:204–206.

    142. Pearson DT, Watson BG, Waterhouse PS. An ultrasonic analysis of the comparative efficiency of various cardiotomy reservoirs and micropore blood filters. Thorax 1978;33:352–358.

    143. Gourlay T, Gibbons M, Fleming J, et al. Evaluation of a range of arterial line filters. Part I. Perfusion 1987;2:297–302.

    144. Gourlay T, Gibbons M, Taylor KM. Evaluation of a range of arterial line filters. Part II. Perfusion 1988;3:29–35.

    145. Gourlay T. The role of arterial line filters in perfusion safety. Perfusion 1988;3:195–204.

    146. Patterson RH Jr. , Wasser JS, Porro RS. The effect of various filters on microembolic cerebrovascular blockade following cardiopulmonary bypass. Ann Thorac Surg 1974;17:464–473.

    147. Pedersen T, Hatteland KSemb BKH. Bubble extraction by various arterial filters measured in vitro with doppler ultrasound techniques. Ultrasound Med Biol 1982;8:71–81.

    148. Hsu LC. Principles of heparin-coating techniques. Perfusion 1991;6:209–219.

    149. Palanzo DA, Kurusz M, Butler BD. Surface tension effects of heparin coating on arterial line filters. Perfusion 1990;5:277–284.

    150. Silvay G, Ammar T, Reich DL, et al. Cardiopulmonary bypass for adult patients: a survey of equipment and techniques. J Cardiothorac Vasc Anesth 1995;9:420–424.

    151. Liu J-F, Su Z-K, Ding W-X. Quantitation of particulate micro-emboli during cardiopulmonary bypass: experimental and clinical studies. Ann Thorac Surg 1992;54:1196–1202.

    152. Munsch C, Rosenfeldt F, Chang V. Absence of particle-induced coronary vasoconstriction during cardioplegic infusion: is it desirable to use a microfilter in the infusion line? J Thorac Cardiovasc Surg 1991;101:473–480.

    153. Kurusz M. Arterial line filtration during cardiopulmonary bypass: history and controversy. Proc Am Acad Cardiovasc Perfus 1983;4:76–86.

    154. Morioka K, Muraoka R, Chiba Y, et al. Leukocyte and platelet depletion with a blood separator: effects on lung injury after cardiac surgery with cardiopulmonary bypass. J Thorac Cardiovasc Surg 1996;111:45–54.

    155. Chiba Y, Morioka K, Muraoka R, et al. Effect of depletion of leukocytes and platelets on cardiac dysfunction after cardiopulmonary bypass. Ann Thorac Surg 1998;65:107–114.

    156. Lazar HL, Zhang X, Hamasaki T, et al. Role of leukocyte depletion during cardiopulmonary bypass and cardioplegic arrest. Ann Thorac Surg 1995;60:1745–1748.

    157. Baksaas ST, Videm V, Mollnes TE, et al. Leukocyte filtration during cardiopulmonary bypass hardly changed leukocyte counts and did not influence myeloperoxidase, complement, cytokinin or platelets. Perfusion 1998;13:429–436.

    158. Hurst T, Johnson D, Cujec B, et al. Depletion of activated neutrophils by a filter during cardiac valve surgery. Can J Anaesth 1997;44:131–139.

    159. Suzuki I, Ogoshi N, Chiba M, et al. Clinical evaluation of a leukocyte-depleting blood cardioplegia filter (BC1B) for elective open-heart surgery. Perfusion 1998;13:205–210.

    160. Gu YJ, deVries AJ, Boonstra PW, et al. Leukocyte depletion results in improved lung function and reduced inflammatory response after cardiac surgery. J Thorac Cardiovasc Surg 1996;112:494–500.

    161. Gott JP, Cooper WA, Schmidt FE, et al. Modifying risk for extracorporeal circulation: trial of four anti-inflammatory strategies. Ann Thorac Surg 1998;66:747–754.

    162. Boldt J, Zickmann B, Fedderson B, et al. Six different hemofiltration devices for blood conservation in cardiac surgery. Ann Thorac Surg 1991;51:747–753.

    163. Faulkner SC, Kurusz MManning JV Jr, et al. Clinical experience with the Amicon Diafilter during cardiopulmonary bypass. Proc Am Acad Cardiovasc Perfus 1987;8:66–69.

    164. High KM, Williams DR, Kurusz M. Cardiopulmonary bypass circuits and design. In: Hensley FA Jr , Martin DE eds. A practical approach to cardiac anesthesia 2nd edition . Boston: Little, Brown and Co, 1995:465–481.

    165. Naik SK, Knight A, Elliott MJ. A successful modification of ultrafiltration for cardiopulmonary bypass in children. Perfusion 1991;6:41–50.

    166. Darling E, Nanry K, Shearer I, et al. Techniques of paediatric modified ultrafiltration. Perfusion 1998;13:93–103.

    167. Groom RC, Akl BF, Albus RA, et al. Alternative method of ultrafiltration after cardiopulmonary bypass. Ann Thorac Surg 1994;58:573–574.

    168. Naik S, Elliott M. Ultrafiltration. In: Jonas RA , Elliott MJ eds. Cardiopulmonary bypass in neonates, infants and young children . Boston: Butterworth-Heineman, 1994:158–172.

    169. Sutton RG. Renal considerations, dialysis, and ultrafiltration during cardiopulmonary bypass. Intern Anesth Clin 1996;34:165–176.

    170. Alston RP, Trew A. An in vitro assessment of a monitor for continuous in-line measurement of PO2, PCO2 and pH during cardiopulmonary bypass. Perfusion 1987;2:139–147.

    171. Bashein G, Pino JA, Nessly ML, et al. Clinical assessment of a flow-through fluorometric blood gas monitor. J Clin Monit 1988;4:195–203.

    172. Pino JA, Bashein G, Kenny MA. In vitro assessment of a flow-through fluorometric blood gas monitor. J Clin Monit 1989;4:186–194.

    173. Pearson DT. Blood gas control during cardiopulmonary bypass. Perfusion 1988;3:113–133.

    174. Justison GA, Parsons S. Improved quality control utilizing continuous blood gas monitoring and computerized perfusion systems. In Proceedings of the 27th International Conference of the American Society of Extra-Corporeal Technology . Reston, VA: American Society of Extra-Corporeal Technology, 1989:83–87.

    175. Baraka A, Barody M, Harous S, et al. Continuous venous oximetry during cardiopulmonary bypass: influence of temperature changes, perfusion flow and hematocrit level. J Cardiothorac Anesth 1990;4:35–38.

    176. Philbin DM, Inada E, Sims N, et al. Oxygen consumption and on-line blood gas determinations during rewarming on cardiopulmonary bypass. Perfusion 1987;2:127–129.

    177. Swan H, Sanchez M, Tyndall M, et al. Quality control of perfusion: monitoring venous blood oxygen tension to prevent hypoxic acidosis. J Thorac Cardiovasc Surg 1990;99:868–872.

    178. Rubsamen DS. Continuous blood gas monitoring during cardiopulmonary bypass: how soon will it be the standard of care [Editorial]? J Cardiothorac Anesth 1990;4:1–4.

    179. Mark JB, Fitzgerald D, Fenton T, et al. Continuous arterial and venous blood gas monitoring during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1991;102:431–439.

    180. McDaniel LB, Zwischenberger JM, Vertrees RA, et al. Mixed venous oxygen saturation during cardiopulmonary bypass poorly predicts regional venous saturation. Anesth Analg 1994;80:466–472.

    181. Bashein G, Greydanus WK, Kenny MA. Evaluation of a blood gas and chemistry monitor for use during surgery. Anesthesiology 1989;70:123–127.

    182. Nicolson SC, Jobes DR, Steven JM, et al. Evaluation of a user-operated patient-side blood gas and chemistry monitor in children undergoing cardiac surgery. J Cardiothorac Anesth 1989;3:741–744.

    183. Kent AP, Tarr JT, Fox MA. Where to sample during cardiopulmonary bypass [Letter]. J Cardiothorac Anesth 1989;3:136.

    184. Stammers AH. Monitoring controversies during cardiopulmonary bypass: how far have we come? Perfusion 1998;13:35–43.

    185. Akers T, Bolen G, Gomez J, et al. In vitro comparison of ECC blood flow measurement techniques. In Proceedings of the 28th International Conference of the American Society of Extra-Corporeal Technology . Reston, VA: American Society of Extra-Corporeal Technology, 1990:17–22.

    186. Kurusz M, Conti VR, Arens JF. Oxygenator failure [Letter]. Ann Thorac Surg 1990;49:511.

    187. Kirson LE, Goldman JM. A system for monitoring the delivery of ventilating gas to the oxygenator during cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1994;8:51–57.

    188. Zia M, Davies EW, Alston RP. Oxygenator exhaust capnography. A method of estimating arterial carbon dioxide tension during cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1992;6:42–45.

    189. O'Leary MJ, MacDonnell SP, Ferguson CN. Oxygenator exhaust capnography as an index of arterial carbon dioxide during cardiopulmonary bypass when using a membrane oxygenator. J Anaesth 1999;82:843–846.

    190. Svenmarker S, Haggmark S, Jansson E, et al. The relative safety of an oxygenator. Perfusion 1997;12:289–292.

    191. Wahba A, Philipp A, Behr R, et al. Heparin-coated equipment reduces the risk of oxygenator failure. Ann Thorac Surg 1998;65:1310–1312.

    192. Geissler HJ, Allen JS, Mehlhorn U, et al. Cooling gradients and formation of gaseous microemboli with cardiopulmonary bypass: an echocardiographic study. Ann Thorac Surg 1997;64:100–104.

    193. Amin IM, Maranets I, Barash P. Transesophageal echocardiography of the distal aortic arch. J Cardiothorac Vasc Anesth 1998;12:599–560.

    194. Moisa RB, Zeldis SM, Alper SA, et al. Aortic regurgitation in coronary artery bypass grafting: implication for cardioplegia administration. Ann Thorac Surg 1995;60:665–668.

    195. Orihashi K, Matsuura Y, Hamanaka Y, et al. Retained intracardiac air in open heart operations examined by transesophageal echocardiography. Ann Thorac Surg 1993;55:1467–1471.

    196. Tingleff J, Joyce FS, Pettersson G. Intraoperative echocardiographic study of air embolism during cardiac operations. Ann Thorac Surg 1995;60:673–677.

    197. Berg E, Knudsen N. Automatic data collection for cardiopulmonary bypass. Perfusion 1988;3:263–270.

    198. Gourlay T. Computers in perfusion practice. Perfusion 1987;2:79–85.

    199. Uppal R, Mills NL, Wechsler AS, et al. 1993 Update: left thoracotomy for reoperative coronary artery bypass procedures. Ann Thorac Surg 1993;55:1575–1576.

    200. Cohn LH. Update of right thoracotomy, femoro-femoral bypass, and deep hypothermia for re-replacement of mitral valve. Ann Thorac Surg 1997;64:578–579.

    201. Ireland KW, Follette DM, Iguidbashian J, et al. Use of a heat exchanger to prevent hypothermia during thoracic and thoracoabdominal aneurysm repairs. Ann Thorac Surg 1993;55:534–537.

    202. O'Connor CJ, Rothenberg DM. Anesthetic considerations for descending thoracic aortic surgery. Part II. J Cardiothorac Vasc Anesth 1995;9:734–747.

    203. Westaby S, Katsumata T. Proximal aortic perfusion for complex arch and descending aortic disease. J Thorac Cardiovasc Surg 1998;115:162–167.

    204. Sasaguri S, Fukuda T, Yamamoto T, et al. Transapical aortic occlusion for cardioplegic delivery during reconstruction of thoracoabdominal aortic aneurysm with deep hypothermic circulatory arrest. J Thorac Cardiovasc Surg 1999;117:186–188.

    205. Aldea GSO'Gara P, Sharpira OM, et al. Effects of anticoagulation protocol on outcome in patients undergoing CABG with heparin-bonded cardiopulmonary bypass circuits. Ann Thorac Surg 1998;65:425–433.

    206. Mahoney CB. Heparin-bonded circuits: clinical outcome and costs. Perfusion 1998;13:1892–1204.

    207. Shore-Lesserson L. Pro: heparin-bonded circuits represent a desirable option for cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1998;12:705–709.

    208. Gu YJ, Boonstra PW, Rijnsburger AA, et al. Cardiopulmonary bypass circuit treated with surface-modifying additives: a clinical evaluation of blood compatibility. Ann Thorac Surg 1998;65:1343–1347.

    209. Groom RC, Akl BF, Albus R, et al. Pediatric cardiopulmonary bypass: A review of current practice. Intern Anesth Clin 1996;42:141–163.

    210. Elliott M. Cannulation for cardiopulmonary bypass for repair of congenital heart disease. In Jonas RA Elliott MJ eds. Cardiopulmonary bypass in neonates, infants and young children . Boston: Butterworth-Heineman, 1994.

    211. Mills NL, Ochsner JL. Massive air embolism during cardiopulmonary bypass: causes, prevention, and management. J Thorac Cardiovasc Surg 1980;80:708–717.

    212. Peters WS, Fann JI, Burdon TA, et al. Port-access cardiac surgery: a system analysis. Perfusion 1998;13:253–258.

    213. Beppu T, Imai Y, Fukui Y. A computerized control system for cardiopulmonary bypass. J Thorac Cardiovasc Surg 1995;109:428–438.