Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.


PART THREE – Clinical Management of Special Surgical Problems

Chapter 17 – Anesthesia for Cardiovascular Surgery

Frank H. Kern,Richard J. Ing,
William J. Greeley



Unique Aspects of Pediatric Cardiac Anesthesia, 571



Developmental Physiology, 572



Developmental Pharmacology,575



Congenital Heart Disease, 576



Physiologic Approach to Congenital Heart Disease, 577



Physiologic Consequences of Congenital Heart Disease,579



Preoperative Evaluation, 579



Intraoperative Management, 581



Cardiopulmonary Bypass, 588



Physiologic Differences Between Pediatric and Adult Patients, 588



Pathophysiology of Hypothermic Cardiopulmonary Bypass in Children,590



Pathophysiology of Deep Hypothermic Circulatory Arrest, 593



Specific Organ Effects: Myocardium, 595



Specific Organ Effects: Brain, 596



Low Flow Cardiopulmonary Bypass, 597



Deep Hypothermic Circulatory Arrest, 598



Cardiopulmonary Bypass Management, 600



Initiation of Cardiopulmonary Bypass, 600



Discontinuation of Cardiopulmonary Bypass,600



Management Strategies After Weaning from Cardiopulmonary Bypass, 601



Pulmonary Artery Hypertension,602



Right Ventricular Dysfunction, 604



Left Ventricular Dysfunction, 605



Surgical Procedures and Special Techniques, 607



Complete Anatomic and Physiological Repairs,609



Palliation Surgery, 620



Single Ventricle Procedures, 624



Anticoagulation, Hemostasis, and Blood Conservation, 629



Anesthesia for Closed Heart Operations, 632



Patent Ductus Arteriosus, 632



Coarctation of the Aorta, 632



Extracardiac Shunts, 632



Pulmonary Artery Banding,633



Atrial Septectomy, 633



Vascular Rings and Slings, 633



Cardiac Pacing, 633



Anesthesia for Interventional Cardiac Procedures, 635



Transcatheter Closure for Atrial Septal Defect, 636



Transcatheter Closure for Ventricular Septal Defect,637



Balloon Angioplasty and Valvotomy, 637



Radiofrequency Ablation of Accessory Pathways, 638



Anesthesia for Noncardiac Surgery in Patients with Congenital Heart Disease, 638



Anesthesia for Nonsurgical Cardiac Disease, 639



Cardiomyopathy, 639



Pericarditis, Pericardial Effusions, and Cardiac Tamponade,640



Summary, 641

Creativity, innovation, hard work, and a bit of luck hallmark many of the scientific breakthroughs that have had a major impact on patient care. Nowhere is this more evident than the development of the heart-lung machine and the application of hypothermia for cardiac surgery. In the early 1950s in Minnesota, Bigelow's insights regarding hypothermia, Gibbon's first heart-lung pump, Lillihei's cross-circulation techniques, and Kirklin's refinements of the heart-lung machine and successful application formed the cornerstones for decades of success in cardiac surgery. This period of innovation marked dramatic breakthroughs for pediatric cardiology and surgery.

Stimulated by the earlier development of the Blalock-Taussig (BT) shunt for palliation of cyanotic heart disease and the development of extracorporeal circulation, a whole new era began in the management of congenital heart defects. Today, cardiac surgery has become an effective treatment for children with congenital heart defects. Through a cooperative effort between those in pediatric cardiology and cardiac surgery, significant progress in medical diagnosis and surgical treatment has been achieved. These advances encouraged the development of anesthesiologists with pediatric cardiovascular interests who understood the pathophysiology of congenital heart defects, the surgical procedures, and the required technology. This area of anesthetic expertise is distinguished by broad knowledge requirements in anesthesiology, pediatric medicine, and cardiology, along with an in-depth knowledge of cardiac and pediatric anesthesiology, intensive care medicine, congenital heart disease (CHD), and its surgical treatment. In addition, working with a multidisciplinary team with special interests and skills in managing patients with CHD and their families is fundamental to anesthetic practice. Once viewed primarily as a technical challenge, pediatric cardiac anesthesia has evolved into an exciting and demanding area based on sound physiologic principles and specialized skills.


The process of repairing CHD pushes human physiology to its limits. Nowhere else in medicine are patients exposed to such biologic extremes as during congenital heart surgery. Patients may be cooled to 15° to 18° C, be acutely hemodiluted to upward of 50% of their extracellular fluid volume, and undergo periods of total circulatory arrest. Managing patients with abnormal blood flow patterns exposed to these physiologic extremes is the challenge facing the anesthesiologist. Knowledge generated from operating rooms, intensive care units, catheterization and echocardiography laboratories, and animal experiments has been invaluable in improving care of these patients.

Clearly, the successful perioperative management of these patients cannot occur without a skilled group of multispecialty physicians, perfusionists, nurses, and respiratory therapists dedicated to the congenital cardiac patient. This team-oriented approach has evolved from an idealized construct to clear guidelines set forth by the American Academy of Pediatrics (AAP) (2002) to establish pediatric cardiovascular centers. Paramount to the AAP guidelines are dedicated facilities, personnel, and patient volume. The basis for these guidelines is provided by studies that have demonstrated a reduced mortality rate with increased hospital and surgical volume. Impact studies have demonstrated that regionalization reduces pediatric cardiac surgical deaths, with evidence-based studies outlining the benefits of referral to larger institutions and studies demonstrating a strong relationship between operator volume and outcomes with adult interventional catheterization procedures ( Hannan et al., 1998 ; Malenka et al., 1999 ; Smith et al., 2002 ; Allen et al., 2003 ). Perhaps the most often quoted study of this group is the New York State review demonstrating that mortality for both high- and low-complexity pediatric cardiovascular procedures is significantly affected by case volume. Sixteen New York hospitals were evaluated. In hospitals performing fewer than 100 congenital cardiac surgeries a year, the mortality rate was 8.3% compared with a mortality rate of 6.0% for hospitals performing more than 100 cases per year. Mortality also correlated with the individual surgeon's annual case volume. Surgeons performing 75 or fewer operations a year had an in-hospital mortality rate of 8.8% versus 6.0% for surgeons with annual volumes of more than 75 operations per year ( Hannan et al., 1998 ).

The increased expectation for better outcomes has been a driving force for the dedication of pediatric cardiac centers and the increasing interest in pediatric cardiac anesthesia support. This has had a dramatic impact on the volume of cases undergoing anesthesia in the operating room, catheterization laboratory, and echocardiography suite. It has become imperative for the anesthesia team to understand the principles underlying the management of patients with CHD and to apply them to the field of clinical anesthesia. This chapter provides an overview of some of the unique features of the pediatric patient, of CHD, and of the surgical procedures; addresses perioperative anesthetic management for procedures requiring cardiopulmonary bypass (CPB); and discusses closed heart procedures and anesthesia for interventional procedures.


Although many of the principles that govern modern pediatric cardiovascular anesthesia are similar to the principles applied to any pediatric patient, several important differences do exist ( Box 17-1 ). Broadly speaking, certain characteristics of the cardiac system, CHD, and the surgical procedures account for the differences and are reviewed here. These differences result from the maturational effects of the developing heart, differing pathophysiologic conditions in CHD, the diversity of surgical repairs, and the use of specialized CPB techniques, such as deep hypothermia with or without total circulatory arrest.

BOX 17-1 

Unique Characteristics of Pediatric Cardiac Anesthesia


Normal organ system development and maturational changes of infancy



Cardiovascular: blood flow patterns of circulation at birth, myocardial compliance, systemic and pulmonary vasculature, β-adrenergic receptors



Pulmonary: respiratory quotient, closing capacity, chest compliance



Central nervous system: brain growth, cerebral blood flow, autonomic regulation



Renal: glomerular filtration rate, creatinine clearance



Hepatic: liver blood flow, microsomal enzyme activity

Disease-growth interrelationship



Effects of systemic disease alter somatic and organ growth



Compensatory ability of developing organs for recovery from injury

Immunologic immaturity of the infant

Obligatory miniaturization, that is, small patient size and body surface area

Congenital heart disease

Diverse anatomic defects and physiologic changes

Altered ventricular remodeling owing to myocardial hypertrophy and ischemia

Chronic sequelae of congenital cardiac disease

Surgical procedures

Diversity of operations

Frequent intracardiac and right ventricular procedures

Use of deep hypothermia and circulatory arrest during repair

Trend toward repair in early infancy

Evolution of surgical techniques to avoid residua and sequelae

Trend toward wider application of certain operations

Fetal Development

Organogenesis of the cardiovascular system is essentially complete by the sixth to eighth week of fetal life. All subsequent in utero cardiovascular growth and development are determined by flow patterns and intravascular/intracavitary pressures (Rudolph, 1985, 2001 [391] [392]). When organogenesis is normal, early fetal flow patterns result in a pressure load on right ventricular ejection. The myocardium responds by developing a predominant right ventricular mass. In late fetal life, pressure differences between the right ventricle (RV) and the left ventricle (LV) are small, and at term the right and left ventricular chambers are of a similar mass. In utero, as well as postnatally, the ventricles undergo a modeling process that depends on pressure and flow characteristics of blood circulating through the heart ( Perloff et al., 1982 ; Rudolph, 1985, 2001 [391] [392]) ( Fig. 17-1 ).


FIGURE 17-1  Comparison of the pressure and saturation data (circled values) for a normal neonatal (A) and adult (B) heart. Diagram of a postnatal heart is during transition to normal, mature type of circulation. Note the retrograde cardiac output through the ductus arteriosus and the pressure difference between the right atrium (RA) and left atrium (LA) and the right ventricle (RV) and left ventricle (LV).  (Adapted from Rudolph AM: Congenital diseases of the heart. 1974, Year Book Medical Publishers, Chicago.)




When cardiac morphogenesis is abnormal, loading conditions differ and the normal ventricular modeling process is altered ( Perloff, 1982 ). Marked changes occur in ventricular size and function. The fetus is generally protected from the consequences of these changes because the lungs are not used for gas exchange, and in the parallel fetal circulation, both ventricles are recruited to pump blood to the systemic circulation. Also, the fetus is more adaptive in that ventricular muscle can increase cell number through cellular hyperplasia and cell mass through hypertrophy. Postnatally, separation of the circulations requires the ventricles to act independently or, if single-ventricle physiology exists, to suddenly increase pressure and volume workload by at least twofold, meeting the circulatory requirements of both pulmonary and systemic circulations (Rudolph, 1985, 2001 [391] [392]).

Another factor complicating adaptation of the neonatal circulation at birth is the limited ability of the neonatal myocyte to augment its intrinsic contractility ( Teital et al., 1991 ). The reduction in myocardial reserve is a function of the ultrastructure of the neonatal myocardium, β-adrenergic receptor density, and reactive pulmonary vascular physiology.

Immature Myocardium

The ultrastructure of the neonatal myocyte is dominated by organelles necessary for cellular synthetic function, resulting in a 50% reduction in the number of myofibrils and a nonlin ear or chaotic arrangement to contractile elements ( Legato, 1975 ). The neonatal myocytes lack a transverse tubular system, possess an immature sarcoplasmic reticulum (SR), and are rounder and more globular than mature myocytes. Intracellular calcium stores are maintained in the SR, a closed intracellular membranous network surrounding the myofibrils. Calcium is released from the SR and initiates cross-bridging between actin and myosin that occurs in a process known as excitation-contraction coupling. In the neonatal myocyte, intracellular calcium stores are reduced and the transport process is impaired. The ability of calcium release from the SR to affect actinmyosin binding is further weakened by the greater distance between the SR and the contractile unit of the myocyte (the sarcomere [actin, myosin], and the regulatory proteins [troponin and tropomyosin]) ( Becker and Caruso, 1981 ; Humpherys and Cummings, 1984 ; Vetter et al., 1986 ; Nasser et al., 1987 ). The myocytic intracellular calcium stores are quantitatively less and calcium release is less effective than in a mature myocyte. The neonatal myocyte is more dependent on extracellular calcium and less dependent on intracellular calcium for contraction, making ionized calcium levels an important consideration in the neonate with myocardial dysfunction or complex congenital cardiac disease.

These ultrastructural changes reduce myocardial contractility and contribute to reduced ventricular compliance. In vitro experiments confirm that neonatal myocardium is both less compliant and less capable of stretch, a function of both the relative increase in noncontractile elements and the globular shape of neonatal myocytes ( Becker and Caruso, 1981 ). Furthermore, the reduced myocardial compliance, together with an LV and an RV of equal mass and wall thickness, creates a greater ventricular interdependence so that failure of one ventricle results in shifting of the septum into the other ventricular chamber. The other low-compliance ventricle cannot readily compensate for the geometric changes, and in the neonate, univentricular heart failure rapidly becomes biventricular.

It has widely been accepted that immature myocardial contractile function is relatively poor in the fetus. Many texts and articles have stated over decades that fetal myocardium is unable to increase ventricular stroke volume when an increased myocardial preload is administered ( Rudolph, 2001 ). However, work in fetal lambs demonstrated that at any level of mean arterial pressure, an increase in left atrial pressure does increase left ventricular stroke volume ( Rudolph, 2001 ). The Starling curve for neonates is not flat and does not differ significantly from adults. This is more consistent with observed clinical practice and is an important correction of previously accepted dogma on neonatal cardiac physiology. What remains unresolved is whether the performances of the fetal and adult myocardium are comparable ( Hawkins et al., 1989 ).

The neonatal heart appears to be more resistant to ischemia-reperfusion injury than is the normal adult heart. Much of this protection is due to a resistance to calcium influx during and after an ischemic event and to the larger energy stores in immature myocytes. Calcium influx, as discussed earlier, is reduced by the immature SR, thereby reducing calcium entry and subsequent cell injury ( Chizzonite and Zak, 1981 ). In addition, the immature myocardium, by virtue of its growth requirements, has increased glycogen and amino acid stores, which increase cellular anaerobic capacity when nutrient delivery is reduced by ischemia ( Hoerter, 1976 ; Julia et al., 1990, 1991 [217] [218]). These factors appear to protect the normal immature myocardium from ischemic injury.

The cyanotic neonate or the neonate who presents in congestive heart failure appears to be less tolerant of ischemia than is the normal neonate ( Jarmakani et al., 1978 ). Experimental evidence suggests that this is most likely due to impaired substrate delivery and marginal myocardial energy reserves ( Jarmakani et al., 1978 ; Julia et al., 1991 ).

Normal Blood Flow Patterns

The cardiovascular system changes markedly at birth due to dramatic alteration in blood flow patterns (Rudolph, 1985, 2001 [391] [392]) ( Fig. 17-2 ). During fetal life, blood flow returning to the heart bypasses the nonventilated, fluid-filled lungs. Blood is then preferentially shunted across the patent foramen ovale (PFO) into the left atrium (LA) or passes from the RV across the patent ductus arteriosus (PDA) to the systemic circulation. At the time of birth, physiologic closure of the PDA and the foramen ovale brings about the normal adult circulatory pattern. The presence of certain congenital heart defects or pulmonary disease can disrupt this normal adaptive process, creating a transitional circulation where right-to-left shunting across the foramen ovale or the PDA persists (Rudolph, 1985, 2001 [391] [392]). Under such circumstances, the continued presence of a transitional circulation leads to severe hypoxemia, acidosis, and hemodynamic instability, which are poorly tolerated in the neonate. On the other hand, when initially treating some forms of cyanotic CHD, the prolongation of this transitional circulation is actually beneficial, permitting pulmonary blood flow (PBF) and postnatal viability. An example of the latter is pulmonary atresia, where PBF is supplied via the PDA. Closure of the PDA results in a cessation of PBF with severe hypoxemia and death if ductal flow is not rapidly restored. Ductal patency can be maintained with the administration of prostaglandin E1. Importantly, the transitional circulation can be manipulated by pharmacologic or mechanical ventilatory techniques to promote hemodynamic stability in the young patient.


FIGURE 17-2  Course of the fetal circulation in late gestation. Note the selective blood flow patterns across the patent foramen ovale and the ductus arteriosus.



Altered Blood Flow Patterns

The combination of low compliance, poor response to exogenously administered catecholamines, and a reduced response to β-adrenergic drugs places the neonate at a distinct disadvantage. When combined with the altered circulatory patterns common in CHD, neonates face an increased likelihood of acute and severe heart failure.

Altered blood flow patterns become apparent at birth when the neonate must suddenly alter its loading conditions. Flow patterns change from a parallel circuit to a series circuit where each ventricle must supply flow independently in series to the systemic and pulmonic circulations ( Baylen et al., 1977 ). Early ventricular failure may ensue, or, if the circulatory system can compensate for abnormal flow patterns, a persistent stimulus for further pathologic ventricular modeling occurs in the neonate or infant. An RV or LV with increased muscle mass and increased end-diastolic pressures causing abnormal systolic and diastolic functions is the net result (Jarmakani et al., 1971, 1972 [205] [206]). Tetralogy of Fallot is a classic example of how persistent abnormal flow patterns affect cardiac function. Children in whom repair is delayed until 1 year of age demonstrate a propensity toward increased ventricular dysrhythmias in later life and have an increased incidence of late sudden death. This is presumed to be due to chronic cyanosis and hypoxemia in the hypertrophied right ventricular myocardium, resulting in multiple small ischemic areas ( Franciosi and Blanc, 1968 ). When repair is delayed until the age of 3 years, a significant reduction in subsequent ventricular function can also be demonstrated ( Borow et al., 1981 ; Graham, 1982 ). The sooner that normal flow patterns can be restored, the less stimuli there are for abnormal ventricular remodeling, and normal cardiac function can be preserved. Children with tetralogy of Fallot repaired earlier than 3 months of age have very good long-term outcomes but may have a transient, early postoperative period of greater inotropic support and fluid requirements ( van Dongen et al., 2003 ).

The anesthesiologist needs to understand the impact of blood flow patterns on the neonatal circulatory system, because altered blood flow patterns at the time of surgery may have a negative effect on ventricular function. The extreme example is an arterial switch operation for transposition of the great arteries (TGA) in a child with an “unprepared” LV. Such a child cannot be weaned from CPB due to acute failure of the LV. This LV has insufficient muscle mass to handle systemic afterload because of a prolonged period of pumping against low pulmonary artery (PA) pressures before repair. The presence of an altered myocardial matrix and dysfunctional contractility should be given strong consideration in the perioperative management of the neonate and young child with CHD.


For many years, intravenous β-adrenergic receptor (β AR) agonist agents have been the mainstay of treatment for the pediatric patient with low cardiac output syndrome. However β AR agonists are less effective during the newborn period compared with their use in older children or adults (Teitel et al., 1985, 1991 [455] [456]; Hausdorf et al., 1992 ). Dobutamine in doses of 6 to 20 mcg/kg per minute results in only a modest increase in load-independent measures of contractility. When propranolol, a β-adrenergic blocker, is administered, a severe reduction in load-independent measures of contractility ensues. In addition, a reduced β AR number and high circulating catecholamines are present in the normal neonate. At rest, neonatal animals have maximal binding of β ARs, resulting in a reduced sensitivity to exogenous catecholamine administration.

Advances in molecular biology techniques have lead to improved understanding of the molecular mechanisms by which drugs and hormones work. With knowledge about the signal transmission pathways that influence myocardial contractility, it is clear that the neonatal heart differs significantly from myocardial responses at an older age. These findings have not only explained the less effective inotropic response to βAR agonist agents seen in neonates but also revealed new lines of investigation and the use of novel agents with improved therapeutic efficacy ( Teitel et al., 1985 ; Anderson, 1989 ).

Although the basic mechanisms that control myocardial contractility appear to be the same in the fetal, neonatal, and adult heart, the maturational process of most myocardial signal transmission pathways is far from complete at birth. Variation in receptor systems with age have been measured at the level of function, quantity of signal transmission proteins within cells, and even differing relative expression of isoforms of the same signaling protein (protein isoforms often possess subtle functional differences) ( Table 17-1 ).

TABLE 17-1   -- Comparison of molecular biology of neonatal and adult myocardial signal transmission pathways





 Myocardial NE levels



 NE release with stress (for example, CPB)



Receptor number




 Angiotensin II



Second messenger c AMP (βAR stimulation)



Desensitization to catecholamines



Contractile elements Troponin I

80% Troponin Is isoform (TnIs)

100% Troponin Ic isoform (TnIc)


20% Troponin Ic isoform (TnIc)


NE, Norepinephrine; CPB, cardiopulmonary bypass; βAR, β-adrenergic receptor; TnIs, troponin Is slow skeletal muscle fiber isoform; TnIc, troponin I cardiac muscle fiber isoform.




Descriptions of the developmental pharmacology of the heart require knowledge of the signal transmission pathways mediating myocardial contractility. Of the many types of excitable transmembrane proteins found in the sarcolemma of myocytes, the majority belong to the G protein-coupled receptor family (Schwinn et al., 1991, 1994 [417] [418]; Brodde et al., 1992 ). This group of receptors mediates several important signal pathways involved with cardiovascular homeostasis (e.g., adrenergic and muscarinic cholinergic receptor systems). Activated G protein-coupled receptors are linked via guanine nucleotide regulatory proteins (G proteins) to specific membrane effectors, which are responsible for the generation of cytoplasmic second messenger molecules. For example, stimulation of βARs in the myocardium causes activation of a stimulatory G protein (Gs) that in turn activates adenyl cyclase to increase generation of the second messenger cyclic adenosine monophosphate (cAMP). cAMP activates protein kinases in the sarcoplasm, which then phosphorylate target structures (e.g., myosin, ATPase, calcium channels) to cause increased myocardial contractility ( Fig 17-3 ). A second regulatory feedback process known as desensitization is simultaneously initiated with receptor activation in some receptor systems, which with continued receptor stimulation leads to blunted responsiveness and a decrease in receptor number ( Schwinn et al., 1991 ; Schwinn, 1994 ). Prolonged βAR activation results in diminished cAMP generation for a given stimulus (uncoupling) and ultimately disappearance of receptors from the cell surface (sequestration and downregulation).


FIGURE 17-3  Schematic diagram of the 12-surface G protein-coupled receptors, β1 and β2, histamine 2 (H2), vasoactive intestinal peptide (VIP), 5-hydroxytryptophan (5HT), and prostaglandin E1 (PGE1). All activate the Gs (or G stimulatory) protein and increase cAMP through stimulation of adenylyl cyclase, A1, M1, and somatostatin (SS). All activate the Gi (or G inhibitory) protein, which inhibits adenylyl cyclase, reducing cytosolic levels of cAMP. The A1, endothelin (ET), and angiotensin II (Ang II) receptors activate the Gq protein, which activates the second messenger phospholipase C and hydrolyses membrane lipids to produce diacylglycerol (DAG) and inositol triphosphate (IP3). IP3 and DAG activate protein kinases and mobilize calcium in a similar fashion as cAMP.



α1ARs are present in human myocardium, where their stimulation has a more modest inotropic effect compared with βARs. α1ARs are linked via Gq proteins and result in the hydrolysis of membrane phospholipids, resulting in diacylglycerol (DAG) and inositol triphosphate (IP3). IP3 and DAG activate protein kinases and mobilize calcium in a similar fashion as cAMP, resulting in a positive inotropic effect. The sarcolemma then becomes an important site of signal convergence for myocardial contractility, because several different receptor systems influence the generation of each “inotropic” second messenger at this site ( Fig 17-4 ).


FIGURE 17-4  Schematic of G protein-coupled receptor. Drug binding to membrane receptors activates guanine nucleotide regulatory proteins, which activate a cytoplasmic second messenger through a specific membrane effector. An example is the Gs (G stimulatory protein), which in the presence of GTP activates adenylyl cyclase, a membrane effector molecule that generates cAMP. cAMP activates protein kinases and increases the intracellular calcium concentration necessary for excitation-contraction coupling.



Although βAR agonist agents mediate the most potent inotropic effect in adults ( Brodde et al., 1992 ), they have not been as thoroughly investigated in the neonate. Because myocardial receptor systems that mediate contractility mature at different rates, it is possible that a more potent combination of receptor system and cardiotonic agents may be necessary for the treatment of low cardiac output in the neonate. For example, there are 12 known membrane receptors located on myocytes that possess inotropic properties. Current therapeutic approaches use only the βAR membrane receptor pharmacologically to augment inotropy.

In addition to membrane receptors, phosphodiesterase III inhibitors such as milrinone, enoximone, and amrinone appear to have impressive effects in some neonates who are refractory to high doses of βAR agonist agents ( Hausdorf et al., 1992 ). These drugs improve myocardial inotropy via a mechanism that differs from β-adrenergic agonists. Phosphodiesterase III inhibitors decrease the rate of cAMP breakdown by inhibiting the action of the enzyme phosphodiesterase in the cytoplasm of the myocyte. By augmenting the length of action of cAMP, contractility is enhanced. Phosphodiesterase inhibitors have inotropic potential by augmenting the efficacy of endogenous catecholamines, and when used in conjunction with β-adrenergic agents, improved contractility should be anticipated ( Hausdorf et al., 1992 ). Milrinone lactate (1,6-dihydro-2-methyl-6-oxo-[3,4′-bipyridine]-5-carbonitrile lactate) is a bipyridine inotrope/vasodilator. Milrinone selectively inhibits peak III cAMP phosphodiesterase and increases circulating cGMP, which facilitates vasodilation in the systemic and pulmonary bed. Peak III cAMP phosphodiesterase is the predominant isoform in cardiac and vascular smooth muscle. Inhibition of this enzyme results in increased cAMP, producing an increase in intracellular ionized calcium in cardiac muscle cells and relaxation of vascular smooth muscle. The effects of milrinone include increased myocardial lusitropy (energy-dependent relaxation), cAMP-mediated sarcolemmal calcium influx, inotropy, and cGMP-mediated smooth muscle vasodilation. Milrinone enhances cardiac contractility and cardiac relaxation and provides systemic and pulmonary afterload reduction.

In addition to conventional inotropic agents, clinical trials have demonstrated the inotropic potential of agents such as thyroxin, which works via several noncatecholamine mechanisms to augment contractility ( Portman et al., 2000 ). The role of adrenal steroid depletion, particularly in stressed neonates, can be a factor in myocardial and vascular insensitivity to exogenously administered catecholamines. Although steroid depletion is uncommon, a brief trial of steroids pending the results of a cortisol level should be considered in neonates refractory to inotropes ( Ng et al., 2001 ).

Still more theoretical is the potential of other receptor systems as angiotensin II receptors, which have only 30% to 50% of the inotropic potency of βARs in adult myocardium ( Brodde et al., 1992 ) but are expressed at a 10-fold higher level than in the adult myocardium ( Urata et al., 1989 ) ( Table 17-2 ). Alternative approaches to inotropic support have been limited by drugs that are nonselective and therefore have additional untoward systemic effects. However, the recent interest in drugs such as vasopressin for cardiovascular support continues to raise interest as alternative options for patients with CHD ( Mann et al., 2002 ).

TABLE 17-2   -- Inotropic potency of myocardial membrane G protein–coupled receptors


Percent Inotropic Effect

Tissue Sample

β1 + β2















Angiotensin II






5-HT, 5-Hydroxytryptamine; VIP, vasoactive intestinal peptide; ET, endothelin.




Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


The marked spectrum of intracardiac shunts, valve stenosis, disrupted great artery connections and the absence of cardiac chambers complicates a uniform anesthetic approach to patients with CHD. Moreover, there are myocardial changes resulting from the hemodynamic stress and increased cardiac work incurred by these defects. Functionally, these myocardial changes place the ventricles at great risk for the development of intraoperative ischemia and failure. An understanding of the isolated defect, associated myocardial changes, and hemodynamic consequences is fundamental to planning an appropriate anesthetic ( Becker and Caruso, 1981 ). Too often, because of the complexity and diversity of the defects, the attentions of the anesthesiologist, cardiologist, and surgeon are on the specific anatomic defect and the physiologic changes are ignored. It is imperative for the anesthesiologist not only to understand the anatomy but also to understand its hemodynamic and functional consequences on the cardiovascular system. Although an isolated heart defect may be identified, the entire cardiopulmonary system is usually affected.

Data compiled in the report of the New England Regional Infant Cardiac Program ( Fyler, 1980 ) concerning the frequency of major cardiac lesions seen in the first year of life are given ( Table 17-3 ). The first four lesions comprise almost 50% of all defects, whereas the first eight comprise almost 70%. Extracardiac anomalies are found in one fourth of patients with CHD. The frequency of extracardiac anomalies in infants with symptomatic cardiac disease is presented ( Table 17-4 ).

TABLE 17-3   -- Frequency of major cardiac lesions in first year of life among 2251 infants with heart disease


Frequency (%)

Ventricular septal defect


D-Transposition of great arteries


Tetralogy of Fallot


Coarctation of aorta


Hypoplastic left ventricle


Patent ductus arteriosus


Endocardial cushion defect


Heterotaxias (dextro, meso-, levo-, asplenia)


Pulmonary stenosis


Pulmonary atresia


Atrial septal defect


Total anomalous pulmonary venous return


Tricuspid atresia


Single ventricle


Aortic stenosis


Double-outlet right ventricle


Truncus arteriosus


Modified from Flyer DC: Pediatrics 65(Suppl):377, 1980. Reproduced by permission of Pediatrics © 1980. Copyright © 1980 by AAP.




TABLE 17-4   -- Frequency of extracardiac anomalies in symptomatic cardiac disease according to afflicted system







Specific syndromes



Central nervous
























From Greenwood RD, Rosenthal A, Parisi L, et al. Pediatrics 55:485, 1975. Reproduced by permission of Pediatrics © 1975. Copyright © 1975 by the AAP.


Percent of total number of 1566 infants with cardiac abnormalities.



The structural alterations seen in CHD comprise an encyclopedic list of malformations and prevent the development of a single, uniform anesthetic plan. A general physiologic classification is listed ( Table 17-5 ). Fortunately, although structurally complex, these defects can be understood in a physiologic framework. Identification and classification on the basis of physiology provide an organized framework for the intraoperative anesthetic management and postoperative care of children with complex congenital cardiac defects. In general, congenital heart lesions fit into one of four categories: shunts, mixing lesions, outflow obstruction, and regurgitant lesions (see Table 17-5 ). Each lesion has a specific effect on PBF, systemic blood flow and ventricular function, which manifests as cyanosis, congestive heart failure, or problems with PA hypertension.

TABLE 17-5   -- Classification of congenital heart defects

Physiologic Classification

Pulmonary Blood Flow



Left to right




Volume-overloaded ventricle






Develop CHF

 AV canal



Right to left Tetralogy of Fallot

Pressure-overloaded ventricle

Pulmonary atresia/VSD



Eisenmenger's complex



Mixing Lesions


Generally ↓ but variable    p/   s

Variable pressure versus volume loaded

Tricuspid atresia



Anomalous venous return


Usually cyanotic

Univentricular heart



Obstructive Lesions

Interrupted aortic arch


Ventricular dysfunction

Critical aortic stenosis



Critical pulmonic stenosis


Pressure-overloaded ventricle

Hypoplastic left heart syndrome



Coarctation of the aorta


Ductal dependence

Mitral stenosis



Regurgitant Lesions

Ebstein's anomaly


Volume overloaded

Other secondary causes


Develop CHF

ASD, atrial septal defect; AV, atrioventricular; CHF, congestive heart failure; PDA, patent ductus arteriosus;    p, pulmonary blood flow;    s, systemic blood flow; VSD, ventricular septal defect.




Shunt Lesions

Shunts are intracardiac connections between chambers or extracardiac connections between the aorta and PA, such as an atrial septal defect (ASD) or PDA. The direction of blood flow through the shunt is dependent on the relative pressures on either side of the shunt and the size of the shunt orifice ( Berman, 1985 ). If the shunt is nonrestrictive and does not impede blood flowing freely in each direction, then the main determinant of blood flow is the resistance of the pulmonary and systemic vascular beds. The effect that a shunt lesion has on the cardiovascular system depends on its size and its direction, either right-to-left or left-to-right. Left-to-right shunts occur when the pulmonary vascular resistance (PVR) is less than the systemic vascular resistance (SVR) and blood flow is preferentially directed toward thelungs, resulting in increased PBF. In patients with large left-to-right shunts and low PVR, PBF can be significantly increased. This results in three pathophysiologic problems: (1) volume overload of the pulmonary circulation, (2) increased cardiac work for the LV that is required to increase stroke volume and heart rate to ensure adequate systemic perfusion, and (3) excessive PBF resulting in progressive elevation in PVR. The demand on the LV for increased cardiac output is limited in the infant, so that a large left-to-right shunt may outstrip the capacity of the left heart to maintain adequate systemic perfusion, and congestive heart failure results. Surgical closure of a hemodynamically significant defect such as a ventricular septal defect (VSD) with a large left-to-right shunt can produce worsening ventricular failure in the early postoperative period due to lack of a “pop off” because the VSD is now closed, requiring the LV and RV to pump blood solely against SVR and PVR, respectively. If the left-to-right shunt is left unrepaired, prolonged exposure to increased PBF results in progressive elevation in PVR. Fixed changes in pulmonary arterioles may develop, leading to pulmonary vascular obstructive disease. Common left-to-right shunt lesions are listed (see Table 17-5 ).

Right-to-left shunts occur when pulmonary vascular or outflow tract resistance exceeds SVR, thereby reducing PBF. The systemic circulation receives an admixture of deoxygenated blood via the shunt and results in cyanosis and hypoxemia. Pure right-to-left shunting due to raised PVR is seen in both Eisenmenger's complex and persistent pulmonary hypertension of the newborn. More commonly, PVR is low and a more complex lesion with obstruction to pulmonary outflow, proximal to the pulmonary vasculature, produces right-to-left shunting. Defects such as tetralogy of Fallot or pulmonary atresia with a VSD are classic right-to-left shunts. The shunting occurs through the VSD because of pulmonary outflow obstruction. Cardiac output is generally normal with right-to-left shunting lesions unless hypoxemia is severe and oxygen delivery to tissue is inadequate. There are two pathophysiologic problems: (1) reduced PBF resulting in systemic hypoxemia and cyanosis, and (2) an increased impedance to right ventricular ejection resulting in a pressure-overloaded RV and right ventricular dysfunction.

Mixing Lesions

Mixing lesions comprise the largest group of cyanotic congenital heart defects (see Table 17-5 ). In these defects, the mixing between the pulmonary and systemic circulation is so complete that the receiving chambers are considered a common chamber. The pulmonary-to-systemic blood flow ratio (   p/   s) is independent of shunt size and totally dependent on vascular resistance or outflow obstruction. The pulmonary and systemic circulations tend to be in parallel with one another rather than in series. In patients with no outflow obstruction, flow to the systemic or pulmonary circulation is dependent on the relative vascular resistances of both circuits, as in a univentricular heart or double-outlet RV. If SVR exceeds PVR in these defects, the tendency is toward excessive PBF and left-to-right shunting predominates. These patients have increased PBF, increased cardiac output, and a gradual elevation of PVR over time. If PVR exceeds SVR, as may occur episodically, in ductal-dependent lesions such as hypoplastic left heart syndrome (HLHS), systemic blood flow predominates, and PBF dramati cally decreases, worsening hypoxemia ( Table 17-6 ). In patients with a mixing lesion and left ventricular outflow obstruction, PBF may be so excessive that systemic perfusion is impaired. Patients with mixing lesions and right ventricular outflow obstruction such as single ventricle with subpulmonic stenosis, systemic-to-pulmonary flow can vary from balanced flow to significantly decreased PBF and hypoxemia depending on the degree of obstruction. Typical mixing lesions include truncus arteriosus, univentricular heart, total anomalous pulmonary venous return, pulmonary atresia with large VSD, and single atrium.

TABLE 17-6   -- Ductal dependent lesions

PDA Provides Systemic Flow

PDA Provides Pulmonary Flow

Coarctation of the aorta

Pulmonary atresia

Interrupted aortic arch

Critical pulmonary stenosis

Hypoplastic left heart syndrome

Severe subpulmonic stenosis with VSD

Critical aortic stenosis

Tricuspid atresia with pulmonic stenosis

PDA, patent ductus arteriosus; VSD, ventricular septal defect.




Obstructive Lesions

Obstructive lesions usually occur across the ventricular outflow tracts and range from mild to severe. Severe lesions present in the newborn period with a pressure-overloaded, diminutive, and profoundly dysfunctional ventricle proximal to the obstruction. These lesions include critical aortic stenosis (AS), critical pulmonary stenosis (PS), coarctation of the aorta, interrupted aortic arch, and HLHS. In the left-sided obstructive defects, systemic perfusion is dependent on blood flow (desaturated) from the RV via the PDA and coronary perfusion is supplied by retrograde flow from the descending aorta. In right-sided obstructive lesions, PBF is supplied from the aorta via the PDA and right ventricular function is impaired. Pathophysiologic problems in left heart obstructive lesions include (1) profound left ventricular failure, (2) impaired coronary perfusion and ventricular ectopy, (3) hypotension, (4) PDA-dependent systemic circulation, and (5) systemic hypoxemia. The pathophysiologic problems of right heart obstructive lesions include (1) right ventricular dysfunction, (2) decreased PBF, (3) systemic hypoxemia, and (4) PDA-dependent PBF. Milder forms of outflow obstruction remain clinically asymptomatic for many years, such as with mild to moderate AS or PS, or asymptomatic coarctation of the aorta.

Regurgitant Lesions

Regurgitant lesions are uncommon as primary congenital defects. Ebstein's malformation of the tricuspid valve is the only pure regurgitant defect presenting in the newborn period. However, regurgitant lesions are frequently associated with an abnormality of valve structure, such as incomplete or partial atrioventricular canal defect, truncus arteriosus, or tetralogy of Fallot with an absent pulmonary valve. The pathophysiology of regurgitant lesions are (1) volume-overloaded circulation and (2) progression toward ventricular dilation and failure.

When considering the incidence of all of the congenital heart defects, three uncomplicated left-to-right shunts (VSD, ASD, and PDA) and two obstructive lesions (PS and coarctation) comprise 60% of all congenital cardiac defects. Mixing lesions, complicated obstructive defects, and right-to-left shunting lesions account for the vast majority of the remaining 40%. Interestingly, it is the latter group of defects that are more difficult to manage, are more labor intensive, and have a significantly higher morbidity and mortality rate. This observation is directly attributed to the complexity of the cardiovascular abnormalities seen in this group of patients where there is usually an absence of a chamber or a major ventricle-artery connection.


The chronic effects of CHD are a consequence of the imposed hemodynamic stress of the defect or the residua and sequelae after cardiac repair or palliation. These effects continue to alter normal growth and development of the cardiovascular system as well as of other organ systems throughout life ( Graham, 1982 ).

Because complete surgical cures are rarely achieved and some repairs are palliative rather than corrective, abnormalities before and after repair produce long-term effects that affect the care of patients with CHD ( Stark, 1989 ). Although the overall outlook for these patients is good in most instances, every defect has associated myocardial changes and every repair leaves certain obligatory abnormalities. The effects of altered myocardial loading conditions after cardiac surgery early in life require close follow-up. For example, after cardiac surgery for critical AS in infancy, subsequent operative or catheterization procedures often become necessary due to residual AS or insufficiency. Intervention limits volume and pressure overload and subsequent myocardial damage later in life ( Brown et al., 2003).

Although many of the abnormalities are trivial and have no major import, others affect major organ system processes such as ventricular function, central nervous system growth, the conduction system of the heart, and PBF. Under these circumstances, the long-term quality of life is affected. Whether anesthetizing these patients for their primary cardiac repair or for noncardiac surgery, these chronic changes should be ascertained and be reflected in the anesthetic plan.

The myocardium is continually remodeled by specific hemodynamic stresses throughout life. Right ventricular growth and development are influenced by the low-resistance afterload of the pulmonary circulation. The LV is coupled to the high-resistance systemic circulation, which accelerates its rate of growth and development. This situation gives rise to the adult heart where left ventricular dominance of myocardial muscle mass occurs. This entire developmental process is referred to as dynamic ventricular modeling ( Becker and Caruso, 1981 ). Abnormal hemodynamic loading conditions associated with CHD interrupt the normal ventricular modeling process ( Perloff, 1982 ).

Abnormalities of ventricular performance at rest and with exercise can be detected in patients with chronic hemodynamic overload and complex cyanotic lesions. These abnormalities in ventricular function are the consequences of chronic ventricular pressure and/or volume overload, repeated episodes of myocardial ischemia, and residua or sequelae of surgical treatment (e.g., ventriculotomy, altered coronary artery supply, inadequate myocardial protection) ( Graham, 1982 ). The modified Fontan procedure, for example, has a 40% increase in hydraulic power cost to move blood through the heart compared with the normal two-ventricle circulation. Power is the rate at which work is done. So although Fontan physiology is a significant improvement over shunt physiology, it still requires a 40% increase in myocardial work attributable to the lack of a pulmonary ventricle ( Senzaki et al., 2002 ). The physiologic adaptive responses to chronic hypoxemia and ventricular pressure or volume overload are the primary stimuli producing the long-term ventricular dysfunction. For example, chronic volume overload of the LV as seen with left-to-right shunts or a chronic pressure-loaded LV due to left-sided obstructive lesions results in congestive heart failure. Chronic right ventricular volume overload as seen in pulmonic insufficiency after tetralogy of Fallot repair or a pressure-loaded RV such as with residual PS is also associated with chronic ventricular dysfunction and failure ( Engle and Perloff, 1983 ). The mechanism for the dysfunction and failure in pressure-loaded ventricles is probably related to the development of myocardial hypertrophy as an adaptive response to chronic hemodynamic overload. The resultant myocardial hypertrophy outgrows vascular supply and results in ischemia and fibroblast proliferation. Permanent changes in myocardial structure and function are the end result. In patients with volume-overloaded ventricles, the heart contraction is inefficient due to extended sarcomere length and impaired actin-myosin cross-bridging.

In patients with cyanotic conditions, the long-term compensation for chronic cyanosis shows a major redistribution of organ perfusion with selected blood flow to the heart, brain, lung, and kidney and decreased flow to the splanchnic circulation, skin, muscle, and bone. Chronic cyanosis is associated with increased work of breathing in an attempt to increase oxygen uptake and delivery. The most dramatic complications are the decreased rate of somatic growth, increased metabolic rate, and increased hemoglobin concentration seen in cyanotic children with unrepaired defects.

Airway concerns are a major issue for children with congenital heart lesions. Airway pathology related to heart disorders fall into two broad categories: (1) disorders related to anomalous relationships between vascular structures and the tracheobronchial tree (e.g., vascular rings); this is discussed further in Anesthesia for Closed Heart Operations; and (2) disorders related to enlarged cardiac structures (e.g., dilated PAs, enlarged LA, ventricular dilation, and hypertrophy) ( Kussman et al., 2004 ). Enlarged cardiac structures can result from increased left-to-right shunting, left atrial enlargement, right and left ventricular hypertrophy or dilation, and dilated PAs. The proximity of the PAs to the bronchi and the relationship of the LA to the distal trachea are shown ( Fig 17-5 ). Increases in left-to-right shunting can dilate PAs and result in airway compression at a number of sites: (1) PA enlargement can cause the aorta to compress the left lateral trachea. (2) PA dilation can cause compression of the left main bronchus at the origin of the upper lobe bronchus and at the junction of the right intermediate and right middle lobe bronchi. (3) Left atrial enlargement can affect the distal trachea and main stem bronchus. (4) Ventriculomegaly can compress the left main bronchus.


FIGURE 17-5  Relationship of the pulmonary arteries to the tracheobronchial tree. (RUL, right upper lobe; LUL, left upper lobe; LLL, left lower lobe; PT, pulmonary trunk; LA, left atrium; RLL, right lower lobe; RML, right middle lobe; RPA, right pulmonary artery.) (Reproduced with permission from Berlinger et al.: Tracheobronchial compression in acyanotic congenital heart disease. Ann Otol Rhinol Laryngol 92:387-390, 1983.)




Symptoms of airway involvement include wheezing, stridor (expiratory or both inspiratory and expiratory), respiratory distress, and apnea. Incomplete obstruction can result in air trapping, atelectasis, pneumonia, and aspiration. Prolonged airway compression can cause tracheomalacia and/or bronchomalacia.


The anesthesiologist who cares for children with CHD is presented with a broad spectrum of anatomic and physiologic abnormalities. Patients range from young, healthy, asympto matic children undergoing closure of a small ASD to the newborn infant with HLHS requiring aggressive perioperative hemodynamic and ventilatory support. Intertwined with the medical diversity of these patients are the psychological issues of both the patient and the parents. Preparation of the patient and the family is time consuming, but omitting or compromis ing this aspect of patient care is a major deterrent to a successful outcome and patient/parental satisfaction. Cardiac surgeons, cardiologists, anesthesiologists, intensivists, and nurses must work as a team in preparing the patient and the family for surgery and postoperative recovery. This team-oriented approach also serves as a checkpoint to prevent errors and omissions in preoperative, intraoperative, and postoperative care necessitated by the complexity of cardiac surgery for CHD. The preoperative visit offers the family the opportunity to meet the surgeon and anesthesiologist and to begin preparing the patient and family for surgery.

The preoperative evaluation should always start with a careful history and physical examination. The history should concentrate on the cardiopulmonary system. Parents should be questioned about the general health and activity of their child. Fundamentally, a child's general health and activity reflect his or her cardiorespiratory reserve. Abnormalities may point toward cardiovascular or other organ system dysfunction that may pose anesthetic or surgical risk. Does the child have normal or impaired exercise tolerance? Is he or she gaining weight appropriately or exhibiting signs of failure to thrive on the basis of cardiac cachexia? Does the child exhibit signs of congestive heart failure (diaphoresis, tachypnea, poor feeding, or recurrent respiratory infections)? Is there progressive cyanosis or new onset of cyanotic spells? Any intercurrent illness such as a recent upper respiratory infection (URI) or pneumonia must also be ascertained. This may require delaying surgery, because of the negative impact airway reactivity and elevation of PVR may have on surgical outcome. It is becoming clear that a URI is not an innocuous problem when elective cardiac surgery is planned. A retrospective study of 713 children scheduled for elective cardiac surgery found that 96 had symptoms of a URI preoperatively. It was found that if symptomatic, they had a higher incidence of respiratory and multiple postoperative complications compared with children without a URI (29.2% versus 17.3% and 25% versus 10.3%, respectively; P < .01) and a higher incidence of postoperative bacterial infections (5.2% versus 1.0%; P = .01). These children with a URI also stay an average of 25 hours longer in the intensive care unit, although total hospital stay may not be prolonged. Parental confirmation of a URI is an important diagnostic indicator ( Malviya et al., 2003 ).

Recurrent pneumonia, as mentioned previously, is a frequent finding in patients with congestive heart failure. In particular, patients with shunt physiology or mixing lesions with excessive PBF and altered lung compliance are at risk for viral and bacterial lung infections. Respiratory syncytial virus (RSV) is a particularly common and poorly tolerated lung infection in these patients.

A good history must delineate previous surgical interventions. The presence of shunts, patches, and conduits has an impact on the selected surgical and anesthetic approach. The presence of a left BT shunt is simply ligated intraoperatively, but accessing a left radial artery for invasive monitoring may be difficult and may not provide accurate or useful information. Current medications, previous anesthetic problems, and family history of anesthetic difficulties are equally important.

In the modern era of echocardiography and cardiac catheterization, physical examination rarely contributes additional information about the underlying cardiac lesion. However, the absence of a previous shunt murmur or the presence of a new murmur suggesting mitral regurgitation could suggest partial occlusion of a shunt or endocarditis, respectively.

It is extremely useful to assess the child's overall clinical condition. For example, an ill-appearing, cachectic child in respiratory distress has limited cardiorespiratory reserve, so the use of excessive premedication or a prolonged inhalational induction could result in significant hemodynamic instability and even cardiac arrest.

Laboratory evaluation should include hemoglobin, hematocrit, and serum electrolyte measurements if the patient is taking diuretics. An elevated hematocrit indicates the chronicity of a relative hypoxemia. Levels above 60% may predispose to capillary sludging and secondary end-organ damage, including stroke ( Richardson and Clark, 1976 ). Harvesting patient blood just before the beginning of the procedure may reduce the need for transfusion after the procedure and provide “fresh whole blood” to treat postoperative coagulopathy in patients with a high preoperative hematocrit.

Echocardiography with Doppler color flow imaging (echo-Doppler) is invaluable, providing a noninvasive means of assessing intracardiac anatomy, blood flow patterns, and estimates of physiologic status (Sahn, 1985 ). For many cardiac defects, more invasive studies are generally not required if a good echocardiographic assessment is made. Echo-Doppler is especially helpful for defining intracardiac abnormalities. Extracardiac abnormalities, such as PA or vein stenosis, are sometimes more difficult to definitively define by echo-Doppler and may require cardiac catheterization or cardiac computed tomography or magnetic resonance imaging. The ability to accurately interpret anatomy and physiology requires a skilled echocardiographer and again points out the need for a well-integrated and interactive team. As intraoperative transesophageal echocardiography (TEE) is becoming an increasingly relied-on operative technique, the anesthesiologist must understand the anatomy and views offered by TEE and assist in the decisions based on the information available. In adult cardiovascular surgery, trained anesthesiologists have accurately assessed the intraoperative images in at least 95% of intraoperative echo studies after the surgical correction ( Bettex et al., 2003 ). Although the complexity and variety of clinical defects are greater in children, the anesthesiologist needs to be involved in the interpretation, medical management, and additional operative interventions based on intraoperative echocardiogram.

Cardiac catheterization remains the gold standard for assessing anatomy and physiologic function in CHD (Rudolph, 1985, 2001 [391] [392]). Important catheterization data for the anesthesiologist include the following:



Location, size, and direction of intracardiac shunting (   p/   s)



Pulmonary and systemic arterial pressures



Ventricular and atrial pressures with specific attention to left and right ventricular end-diastolic pressure



Oxygen saturation data



Intracardiac chamber size



Pulmonary vascular resistance



Valvular anatomy and function



Anatomy, location, and function of previously created shunts



Anatomic distortion of systemic or pulmonary arterial vessels, especially as it relates to previously placed shunts



Coronary artery anatomy

A careful review of the cardiac catheterization data and an understanding of how this information affects the operative and anesthetic plans are essential. Not all of the medical problems can be evaluated and corrected preoperatively; the surgeon and anesthesiologist must discuss potential management problems and any needs for further evaluation or intervention before operative intervention is considered. Appropriate communication and cooperation between the two physicians maximize patient care and facilitate perioperative clinical management. Typically, most institutions have a regularly sched uled, combined cardiology/cardiac surgery/anesthesiology/intensive care unit meeting to discuss candidates for surgery, during which all of the essential information regarding the previous list is displayed and discussed. Such a meeting is invaluable for learning about particular patients for surgery as well as providing a continuing educational opportunity to understand CHD and its medical, surgical, and interventional treatment options.


The goal of premedication is to achieve adequate sedation in a nontraumatic fashion and to maintain respiratory and hemodynamic stability. In children with complex CHD, premedication is advocated. This improves oxygen saturation, diminishes myocardial oxygen consumption, and promotes a more satisfactory induction. Many premedication combinations have been used, including intramuscular opioids and sedatives, rectal barbiturates, a combination of a rectal barbiturate and an intramuscular opioid, or a combination of oral medications, including an opioid, a barbiturate, and/or a benzodiazepine. Rectal barbiturates are effective, but reports of apnea, bradycardia, and respiratory depression in 1% to 2% of patients mandate close observation by skilled personnel ( Laishley et al., 1986 ). Bradycardia and apnea in particular are poorly tolerated in the child with CHD and may result in significant morbidity and mortality. Intramuscular medications cause pain and agitation, which generally result in increased myocardial oxygen consumption and arterial desaturation at the time of injection. Because this usually occurs before arrival into the operating suite and in the presence of the parents, most institutions have abandoned intramuscular injections for children entirely. Oral administration of premedication is effective and is the most widely accepted premedication for children with heart disease. In general, children younger than 6 months do not require a premedication agent. In children between the age of 6 and 9 months, midazolam (0.3 to 0.7 mg/kg) may be administered orally. In older children, 0.5 to 1.0 mg/kg (maximum dose, 20 mg) is effective. A calm, cooperative, sedated child is the usual result. In patients with known paradoxical responses to benzodiazepines, alternative agents such as pentobarbital (2 to 4 mg/kg) may be considered. Barbiturates have greater myocardial depressant effects, and dosages may need to be reduced in selective patients.


Operating Room Preparation

Advanced, careful preparation of the operating room is essential. The anesthesia machine must have the capacity to provide air as well as oxygen and nitrous oxide to help balance pulmonary and systemic blood flow. Intravenous tubing must be free from air bubbles to prevent air embolism to the left side of the circulation in patients with open communication, such as an ASD. Resuscitative drugs, labeled and ready for administration, should include calcium gluconate or calcium chloride, sodium bicarbonate, atropine, phenylephrine, lidocaine, and epinephrine. An inotropic infusion, usually dopamine, should be premixed and ready for administration for most cases, and additional infusions are made available if there is a strong suspicion for their need (epinephrine and milrinone). For all pediatric cases, certain anesthetic drugs are made available for use on an emergency basis (succinylcholine and atropine). These drugs are selected because of the potential for airway reactivity, hypotension, and bradycardia during anesthetic induction. In pediatric cardiac anesthesia, many of the patients have high endogenous catecholamines as an adaptive response to their underlying cardiac disease and have limited cardiovascular reserve. Thus, resuscitative drugs should be drawn up prior to anesthetic induction.

For congenital heart surgery, the ability to rapidly alter body temperature for cooling and rewarming is essential. During deep hypothermic CPB patients are cooled to 15° to 18°C. Surface cooling with a heating/cooling water mattress, warm air convection device, and an efficient room cooling/heating system are important in the operative management of these patients. The use of ice packs to the head is generally applied if circulatory arrest is part of the operative plan.

Physiologic Monitoring

The monitoring used for any particular patient should be dependent on the condition of the patient and the type and extent of the surgical procedure. The perioperative monitoring techniques available are listed ( Box 17-2 ). Noninvasive monitoring devices are placed before the induction of anesthesia. In the crying pediatric patient, monitoring devices can be applied immediately after the induction of anesthesia, except for precordial stethoscope and pulse oximetry. Standard monitoring includes a precordial stethoscope five-lead electrocardiographic system, pulse oximetry, an appropriate-sized noninvasive blood pressure cuff, end-tidal CO2 monitoring, and end-tidal gas monitoring. Additional monitoring includes an indwelling arterial catheter, central venous catheter, temperature probes, and TEE. Foley catheters are used in neonates and infants undergoing hypothermic circulatory arrest or reoperations and may be electively withheld in older children for less complex procedures unless dictated by renal insufficiency, prolonged procedures, significant fluid intake, or surgeon preference.

BOX 17-2 

Intraoperative Monitoring

Cardiopulmonary System


Standard seven-lead system, ST-T wave analysis

Pulse oximetry

Noninvasive blood pressure cuff

Capnograph, inhalational gas monitoring

Respiratory mechanics monitoring (peak inspiratory pressure, mean arterial pressure, peak end-expiratory pressure, tidal volume)

Indwelling arterial catheter

Central venous pressure catheter

Transthoracic pressure monitor

Left or right atrium, pulmonary artery

Transesophageal echocardiography with Doppler color flow imaging

Activated coagulation time monitor, blood gas monitoring, including lactate, pH, hematocrit

Central Nervous System

Peripheral nerve stimulator

Jugular venous bulb saturation

Transcranial Doppler

Processed electroencephalography, bispectral analysis

Near infrared spectroscopy, cerebral oxygenation index


Nasopharyngeal, rectal, esophageal, tympanic

Renal Function

Foley catheter

Continuous monitoring of arterial pressure is possible only through the use of an indwelling intra-arterial catheter. In young children, cannulation of the radial artery with a 22-gauge catheter is preferred. In older children and adolescents, a 20-gauge catheter may be substituted. Care must be taken to ensure that previous or currently planned operative procedures such as a radial artery cutdown, subclavian flap for coarctation repair, or a classic BT shunt do not interfere with the selected site of arterial pressure monitoring. Other sites available for cannulation include the ulnar, femoral, or axillary artery. Cannulation of the posterior tibial or dorsalis pedis artery is not usually performed for complex operative procedures. Peripheral arterial catheters of the distal lower extremities function poorly after CPB and do not reflect central aortic pressure when distal extremity temperature remains low ( Stern et al., 1985 ). Alternatively, in patients with poor arterial access, a radial or an ulnar cutdown should be considered.

Myocardial and cerebral preservation is principally maintained through hypothermia, so the accurate and continuous monitoring of body temperature is crucial. Rectal or urinary and nasopharyngeal temperatures are monitored because they reflect core and brain temperatures, respectively. Monitoring of esophageal temperature is a good reflection of cardiac and thoracic temperatures. Tympanic probes are used successfully in some centers and are a useful reflection of cerebral temperature ( Kern et al., 1992) . They rarely cause tympanic membrane injury.

Pulse oximetry and capnography provide instantaneous feedback concerning adequacy of ventilation and oxygenation. They are useful in balancing shunt flow and providing data about surgically created shunts and PA bands, especially after the surgical procedure is completed. Peripheral vasoconstriction in patients undergoing deep hypothermia circulatory arrest (DHCA) sometimes renders digital oxygen saturation probes less reliable. Alternative sites such as the ear lobe or the tongue sensor have been used successfully in the newborn to provide a more central measure of oxygen saturation, with less temperature-related variability ( Jobes and Nicolson, 1988 ).

The use of transthoracic (right atrium [RA], LA, or PA) or transvenous PA catheters is determined on an individual basis based on the disease process, surgical procedure, and needs of postoperative monitoring. For example, in neonates with PA hypertension or in children undergoing a Fontan procedure for tricuspid atresia or univentricular heart, these measurements can be especially useful. In the Fontan operation, no ventricle pumps blood to the lungs; adequacy of flow through the pulmonary bed is dependent on maintaining a gradient from the superior vena cava to the common atrium. Failure to maintain this gradient results in no forward flow, low cardiac output, and death. Monitoring of intracardiac common atrial pressure is useful in the intraoperative and postoperative management of these patients. In newborns, infants, and young children, transvenous PA catheters are more difficult to place.

As a general guideline, a transvenous PA catheter may be placed using the internal jugular approach in children weighing more than 7 kg. A 5.0 F PA catheter is used for patients with a body weight of 7 to 25 kg, and a 7.0 F PA catheter for children weighing more than 25 kg. For infants weighing less than 7.0 kg, percutaneous placement of a PA catheter can be performed from the femoral vein. The smaller transvenous catheters are difficult to float to the PA because of the extreme flexibility of the catheter, especially when they warm to body temperature. They therefore may require fluoroscopic guidance to place successfully. The use of intraoperative echo-Doppler has markedly reduced the need for placement of indwelling intracardiac catheters or transvenous PA catheters.

Cardiac Monitoring


In the late 1980s, echocardiography was introduced. The most promising was echocardiography with Doppler color flow imaging. Several reports from the late 1980s and early 1990s described the usefulness of intraoperative echo-Doppler during congenital heart surgery ( Hagler et al., 1988 ; Ungerleider et al., 1989b ; Muhiudeen et al., 1990, 1991 [318] [319]). Two-dimensional echocardiography combined with pulsed-wave Doppler ultrasonography and color flow mapping demonstrates detailed morphologic as well as physiologic information in most cases.

The availability of biplane and omniplane TEE probes in smaller sizes has enabled TEE to become the standard modality for intraoperative echocardiography. The increased viewing angles available with these multiplane imaging probes have significantly improved the ability to evaluate the entire heart both before and after the repair. Epicardial echocardiography, which requires direct placement of the probe on the heart by the operative surgeon or cardiologist, is now infrequently used. Issues such as probe-induced arrhythmias, hemodynamic changes, and the risk of infection have always been considerations with this technique ( Smallhorn, 2002 ). Further, the ability to obtain a variety of useful imaging views is limited by the amount of heart exposed and the probe angles available from an epicardial technique. In small neonates or when the surgeon attempts to provide smaller, more cosmetically appealing incisions, the exposed surface area of the heart is quite limited. Epicardial imaging is generally reserved for neonates weighing 2 to 2.5 kg or less or for a child with esophageal anomalies (tracheoesophageal fistula repair). Although monoplane probes are capable of being placed in infants weighing less than 2 kg, the available views remain limited, particularly in the more complex repairs performed in neonates.

With the use of TEE in the operating room, anatomic and physiologic data can be obtained before CPB. Occasionally, the preoperative evaluation may result in a revision of the initial diagnosis or identify an additional defect not previously recognized. This evaluation may refine the anesthetic and operative plans ( Ungerleider et al., 1989a ; Muhiudeen et al., 1991 ; Smallhorne, 2002; Bettex et al., 2003 ). Because of the unrestricted TEE approaches in anesthetized patients, new anatomic findings may be discovered and management plans changed accordingly. Postbypass echo-Doppler evaluation is able to immediately assess the quality of the surgical repair and to assess cardiac function by examining ventricular wall motion and systolic thickening ( Ungerleider et al., 1989 ; Muhiudeen et al., 1992 ;Smallhorne, 2002 ; Bettex et al., 2003 ). This technique can show residual structural defects after bypass, which can be immediately repaired in the same operative setting and prevent the patient from leaving the operating room with significant residual structural defects that later require reoperation ( Fig 17-6 ). The ability to identify patients with new right and left ventricular contraction abnormalities after bypass, as determined by a change in wall motion or systolic thickening, allows for immediate and more thoughtful pharmacologic interventions when guided by TEE evaluation. Importantly, postbypass ventricular dysfunction and residual structural defects are identified by echo-Doppler assessment; left uncorrected, these are associated with an increased incidence of reoperation and greater morbidity and mortality ( Ungerleider et al., 1992 ). This monitoring tool helps assess surgical outcome and identify operative risk factors.


FIGURE 17-6  (A) Echocardiogram with a Doppler flow map in the long-axis view illustrating a residual ventricular septal defect (VSD) resulting from patch dehiscence after initial repair. Turbulent flow through the VSD appears as a mosaic of white particles (arrow).This finding necessitated immediate reinstitution of CPB and rerepair. (B) Repeat Doppler flow map in the long-axis view illustrates patch closure (arrow) of the VSD after rerepair. Note the absence of turbulent flow with the loss of the mosaic of white. AO, aorta; LA, left atrium; LV, left ventricle; RV, right ventricle.



Surgeons can demonstrate an operative learning curve with a reduced incidence of residual defects with experience. However, even when experienced surgeons perform the procedures, the use of an intraoperative echocardiogram can detect a 3% to 4% incidence of clinically significant residual disease that requires further surgical repair ( Ungerleider et al., 1995 ). Patients leaving the operating room with residual disease have a considerable increase in hospital cost, length of stay, and need for further operative or interventional procedures. Cost and outcome benefits exist if residual anatomic disease is minimized by ensuring the most complete repair possible through the use of intraoperative TEE ( Ungerleider et al., 1995 ).

The TEE probe is usually placed in the esophagus after the induction of anesthesia and intubation, so it is then available for monitoring. In a few centers, a prebypass image is not obtained and probe placement commences after weaning from CPB. The advantages of placing the probe at the beginning of the case are availability of a continuous monitor of cardiac structure and function and ability to monitor function at the time of weaning from CPB ( Cyran et al., 1991 ; Muhiudeen et al., 1992 ). In addition, pharyngeal or esophageal bleeding is less likely when the probe is placed before heparin administration. Because of its ideal imaging location, biplane and multiplane TEE has been especially helpful in evaluating pulmonary venous return and the integrity of the left atrioventricular valve (AVV) after mitral valvuloplasty, complete AVV repair, and correction of complex CHD ( Smallhorne, 2002 ). With the advent of smaller biplane and omniplane probes, enhanced viewing profiles have been achieved for most newborn infants (2 to 2.5 kg) requiring cardiac surgery ( Decoodt et al., 1992 ). Potential complications of TEE in neonates include failure to insert the TEE probe, descending aorta and airway compression due to relatively large probe size or during probe flexion, accidental extubations particularly in orally intubated patients, and there has been one case of gastric incision described (Stevenson, 1995 ). In this case, the surgeon mistook the TEE probe for the end of the sternum. Many of the complications described are considered avoidable; with increased experience and more careful monitoring, these events should become increasingly rare ( Stevenson, 1995 ).

The technique for intraoperative echocardiographic analysis in children using an epicardial approach is less commonly applied (Ungerleider et al., 1989a, 1992 [464] [465]; Shah et al., 1992 ; Bengur et al., 1998 ). This approach requires that a clean, short-focused 5.0- or 7.0-MHz transducer be passed over the anesthesia screen, into a sterile sheath, where it then can be placed on the epicardial surface of the heart. This technique allows probe manipulations to optimally evaluate the major structures and dynamic function of the heart. The disadvantages of this approach include that sufficient operator skill and experience are required for probe manipulations ( Ungerleider et al., 1992 ), views may be limited by the ability to place the probe on only selective regions of the heart, and the plane of the probe may provide off-angle and infrequently viewed cuts through the heart ( Mochizuki et al., 1999 ).

Since the initial reports in the late 1980s, TEE has become a standard intraoperative approach in most major centers. Stevenson (2003) surveyed 70 congenital heart centers in the United States and Canada about their intraoperative use of echocardiography. Sixty-five centers responded, with 100% of them using intraoperative echocardiography (98% via TEE). Seventy-two percent of the centers used echocardiography (TEE) for all cases. The average duration of TEE experience for these centers was 6.1 years, and the majority of the centers relied on pediatric cardiologists trained in TEE to perform and interpret the studies.

Central Nervous System Monitoring

The goals of brain monitoring are twofold. The first goal is to improve understanding of the cerebral function and dysfunction during cardiac surgery so that effective brain protection strategies can be developed. The second goal is to provide “online” cerebral monitoring to elucidate correctable cerebral perfusion abnormalities during CPB. Because many of the determinants of normal brain perfusion become externally controlled by the cardiac team during CPB (e.g., flow rate [cardiac output], perfusion pressure, temperature, hematocrit, arterial and venous cannula positions, and Paco2), knowledge of the effect of these factors on the brain in neonates, infants, and children is essential. Furthermore, examination of the brain under unusual bio logical circumstances, such as after total circulatory arrest or during continuous-flow CPB at deep hypothermia (15° to 18° C), permits a unique opportunity to describe cerebrovascular physiology and pathophysiology. Processed electroencephalography, transcranial Doppler (TCD), cerebral blood flow (CBF), jugular venous oxygen saturation, near infrared spectroscopy (NIRS), and cerebral metabolism measurements have provided important information during pediatric cardiovascular surgery ( Austin et al., 1997 ).


This is helpful in monitoring physiologic function of the central nervous system during deep hypothermic bypass and total circulatory arrest. For example, during deep hypothermia and before total circulatory arrest, the electroencephalogram can identify residual cerebral electrical activity ( Hickey and Wessel, 1987 ). Isoelectric silence can then be induced by further cooling. Because any residual electrical activity during arrest is associated with cerebral metabolism above basal activity, an isoelectric state may minimize ischemic injury to the brain during circulatory arrest. The use of drug-induced electrical silence does not have the same protective effect as hypothermia and may contribute to hemodynamic compromise in patients with postoperative myocardial dysfunction. In addition, the absence of electrical activity, particularly in newborns, does not necessarily correlate with optimal brain cooling. Case reports exist in which premature and near-term asphyxiated newborns have prolonged electrical silence with the eventual return of cortical electrical activity although with minimal neurologic recovery ( Ashwal, 1997 ). Nonetheless, these observations suggest that electroencephalographic monitoring may not be as useful in newborns as has been suggested in adults to ensure optimal cerebral protection from hypothermia.

The electroencephalogram may also be useful in detecting the level and depth of anesthesia. In particular, the bispectral index (BIS), a processed electroencephalogram, has proved to be an effective monitor in older children and adults ( Glass et al., 1997 ; Todd, 1998 ; Laussen et al., 2001 ). In newborns and infants, its reliability has been questioned, as processed electroencephalographic monitoring and its associated numerical correlation with anesthesia depth are based on adult electroencephalographic wave forms ( Davidson et al., 2001 ). Evidence from studies in newborns suggests a much poorer correlation with the BIS number and the depth of anesthesia. Stimulus-induced elevations to the BIS occurred at much lower BIS levels in infants than in children during emergence from anesthesia (Davidson et al., 2001 ). Similarly, children with cerebral palsy and mental retardation demonstrate lower BIS values than matched normal children when awake and at similar levels of inhalational anesthetic ( Choudhry et al., 2002 ).

After CPB, the presence of electroencephalogram-based seizures has been an indicator of significant neurologic injury ( Rappaport et al., 1998 ; Bellinger et al., 1999 ). A strong correlation between post-CPB seizures and measured reductions in intelligence quotient later in life was demonstrated by the Boston Circulatory Arrest Study Group ( Newburger et al., 1993 ; Rappaport et al., 1998 ). Possible seizure activity in the postoperative period should be suspected when physiologic parameters such as tachycardia or hypertension are seen. A low threshold for electroencephalographic evaluation or the use of antiepileptic agents as part of postoperative sedation (midazolam) should be strongly considered in the neonatal population. The etiology for post-CPB electroencephalographic seizures remains unclear. However, there is an increased risk with the presence of a VSD, suggesting that left-sided air and air embolism may be factors ( Newburger et al., 1993 ).

Transcranial Doppler.

TCD is one of a number of methods used to monitor CBF during pediatric cardiac surgery ( Lundar et al., 1987 ; Hillier et al., 1991 ). TCD technology uses the Doppler principle to detect shifts in the frequency of reflected signals from blood in the middle cerebral artery to calculate blood flow velocity ( Bishop et al., 1986 ). Because the diameter of this large cerebral artery is relatively constant, flow velocity should approximate CBF. The principal advantages of TCD include that it is noninvasive, it does not require radiation exposure, and it is a continuous monitor. An additional advantage of this technique is the capability of assessing rapid alterations in blood flow velocity due to temperature or perfusion changes, as commonly occur during cardiac surgery. The limitations of TCD monitoring include reproducibility, especially at low flows, where minute movement of the patient's head can dramatically alter signal intensity and alter baseline measurements; and the lack of validating studies of TCD during hypothermic CPB, where temperature, reduced flow rates, and the laminar flow characteristics of nonpulsatile perfusion may limit the accuracy of CBF velocity measurements. Although CBF velocity measurements by TCD have reasonable correlation with more standard measures of CBF during normothermia, there have been few studies examining its validity during hypothermic CPB.

TCD has been used to investigate the effect of CPB and deep hypothermic circulatory arrest (DHCA) on cerebral hemodynamics in children as well as to assess the incidence of cerebral emboli and the presence of flow reductions associated with cannula malplacement or perfusion abnormalities during bypass. Studies using TCD have enabled several investigative groups to provide important information regarding questions of normal and abnormal brain perfusion during cardiac surgery in children. Questions regarding cerebral perfusion pressure, autoregulation, and effect of Paco2 and temperature have been addressed using TCD in children and are discussed later in Cardiopulmonary Bypass ( Lundar et al., 1987 ; Hillier et al., 1991 ; Austin et al., 1997 ).

TCD has also provided qualitative and quantitative information regarding the presence of gaseous emboli in the middle cerebral artery during cardiac surgery ( Padayachee et al., 1987 ; van der Linden et al., 1991) . Quantification of this important mechanism of cerebral injury during cardiac surgery would be instructive, because it has been suggested to be a contributor to neurologic injury.

It also has been instructive as a method for determining optimal perfusion techniques. An example is a study in dogs ( Cook et al., 2001) that suggested low-flow bypass actually results in a higher degree of gaseous emboli to the brain compared with high-flow bypass. The authors suggest that with higher SVR during low-flow CPB, an increased proportion of emboli is sent to the cerebral circulation. If supported by other studies, low-flow CPB may not be viewed as favorably as a preferred method of CPB management ( Cook et al., 2001) . Future investigations using TCD should address this mechanism of injury.

CBF studies in which xenon clearance technology was used have improved understanding of cerebrovascular dynamics in young children during CPB, especially during deep hypothermia and after periods of circulatory arrest (Greeley et al., 1988, 1989 [156] [157]; Greeley and Ungerleider, 1991 ; Kern et al., 1991a, 1992b [222] [223]). In general, this investigational tool has been used to describe the effects of CPB, temperature, and various perfusion techniques on CBF and, indirectly, on brain metabolism. Studies in which this methodology was used have shown that some of the mechanisms of CBF autoregulation, such as pressure-flow regulation, are lost with deep hypothermia and that cerebral reperfusion is impaired after a period of total circulatory arrest ( Fig 17-7 ).


FIGURE 17-7  Bar chart of the changes in cerebral blood flow (CBF) before, during, and after cardiopulmonary bypass (CPB) in 67 infants and children (values are mean ± SD). Group (Frp) A underwent repair with moderate hypothermic CPB (MoCPB) at 28° to 32°C. Group B underwent deep hypothermic bypass (DHCPB) at 18° to 22°C. Group C underwent total circulatory arrest (TCA) at 18°C. Stage I, prebypass; stages II and III, during hypothermic bypass; stage IV, warming on bypass; stage V, after bypass. Note the impaired cerebral reperfusion after TCA (group C).  (From Greeley WJ, Ungerleider RM, Smith L, et al.: The effects of deep hypothermic cardiopulmonary bypass and total circulatory arrest on cerebral blood flow in infants and children. J Thorac Cardiovasc Surg 97:737-745, 1989a, with permission.)




The capability of measuring cerebral metabolic activity during cardiac surgery has been applied both in immature animals and in neonates, infants, and children. Methods for monitoring cerebral metabolic activity include determining the cerebral metabolic rate for oxygen (CMRO2) from CBF measurements, measuring jugular venous bulb saturation, and the use of NIRS. With CMRO2used as a metabolic index, the effects of temperature and DHCA on brain metabolism have been described ( Greeley et al., 1991b ). The primary effect of cooling during cardiac surgery is to reduce energy metabolism so that low-flow states and DHCA can be used. Monitoring the efficacy of brain cooling can be performed by measuring CMRO2or, more simply, by examining the venous oxygen saturation of the brain. The higher the saturation level during cooling, the greater are oxygen metabolic suppression and the protective cooling effects. A catheter can be placed in the right internal jugular vein, advanced retrograde to the venous bulb, and positioned to assess the cerebral venous effluent ( Kern et al., 1992a ). Using this technique, potential mechanisms for brain injury have been identified and effective protection strategies suggested, ensuring more complete brain cooling.

Near Infrared Spectroscopy.

NIRS has the capability of measuring regional brain tissue oxyhemoglobin and cytochrome aa 3, the terminal mitochondrial enzyme in the respiratory chain. With the use of NIRS, intracellular brain tissue oxygen delivery and utilization during CPB have been preliminarily observed ( Greeley et al., 1991) . After promising animal studies indicating that mixed venous oxygen levels measured from the jugular bulb correlated linearly with cerebral mixed venous saturations measured with the NIRS monitor, commercial devices measuring oxyhemoglobin saturation and desaturation were approved by the U.S. Food and Drug Administration and are clinically available ( Abdul-Khaliq et al., 2002 ). This device has two flexible pediatric disposable probes, which are easily applied to a patient's forehead. An oxyhemoglobin saturation index is measured in both hemispheres of the brain. Marked differences between perfusion to the right or the left side have suggested problems with adequate surgical arterial or venous cannulation placement. In addition, low cerebral oxygen delivery can be inferred by reductions in the measured oxyhemoglobin saturation index levels. Low oxyhemoglobin saturations are identified in patients undergoing hypoplastic left heart surgery during rewarming and after separation from CPB even when hemodynamics and arterial blood gases appear to be adequate. Efforts to increase cardiac output and oxygen-carrying capacity by raising the hematocrit and lowering the SVR generally improve cerebral oxygen saturation.

In addition to operative monitoring, there has been an increased interest in postoperative cerebral monitoring. The advantage of NIRS monitors is that the monitoring probes have longevity; the disposable flexible probes may be left in situ for up to 72 hours. Postoperatively in the intensive care unit, the data displayed may help determine adequate cerebral oxygenation trends. Clinically, NIRS may be used as an adjunctive continuous monitor of cerebral oxygen delivery, which at normothermia has a strong correlation with systemic oxygen delivery as measured by mixed venous saturations. Some congenital cardiac centers have begun to use a noninvasive cerebral oxygen saturation monitor as an adjunct for trends in effective cardiac output and oxygen delivery. The NIRS monitor is particularly useful in managing infants with single-ventricle anatomy after the Norwood procedure with or without the Sano modification in the intensive care unit when optimization of systemic cardiac output is crucial. However, the one disadvantage of the NIRS monitor is that it cannot be calibrated.

Induction and Mainteinance of Anesthesia

The principles of intraoperative management of cardiothoracic surgical procedures are based on an understanding of the pathophysiology of each disease process and a working knowledge of the effects of the various anesthetic and other pharmacologic interventions on the particular patient's condition. Selection of an induction technique is dependent on the degree of cardiac dysfunction, the cardiac defect, and the degree of sedation pro vided by the premedication. In children with good cardiac reserve, induction techniques can be quite varied as long as induction is careful and well monitored. The execution of induction is more important than the specific anesthetic technique in patients with reasonable cardiac reserve. A wide spectrum of anesthetic induction techniques with a variety of agents has been used safely and successfully; such as sevoflurane, sevoflurane and nitrous oxide, halothane, halothane and nitrous oxide, intravenous or intramuscular ketamine, or intravenous fentanyl, midazolam, propofol, or thiopental ( Laishley et al., 1986 ; Williams et al., 1999 ; Russell et al., 2001 ). In patients with more limited cardiac reserve, the choice of induction agent becomes more important. In a prospective double-blind randomized study of inhalational agents in children undergoing congenital heart surgery, the use of halothane was compared with the use of sevoflurane for both induction and maintenance of anesthesia ( Russell et al., 2001 ). Sevoflurane demonstrated a significant hemodynamic benefit compared with halothane. Primary outcome variables included bradycardia, severe hypotension (defined as a 30% decrease in the resting mean), and oxygen desaturation (>20% decrease) for at least 30 seconds. Patients receiving halothane experienced twice as many episodes of hypotension as did those receiving sevoflurane (P = .03) and more common use of therapeutic vasopressors to treat hypotension, particularly in patients with significantly altered cardiac physiology.

The most widely practiced intravenous induction techniques today include intravenous induction with a benzodiazepine such as midazolam and an opioid such as fentanyl or sufentanil. Alternative induction agents include etomidate, ketamine, and propofol. Etomidate provides hemodynamic stability and has been advocated for pediatric patients with limited cardiac reserve ( Tobias, 2000 ). Ketamine is also an effective induction agent in children and has been advocated for patients with tetralogy of Fallot and other cyanotic lesions because it increases SVR, maintains cardiac output, and promotes left-to-right shunting across a VSD or extracardiac shunt. Ketamine can be administered intravenously or intramuscularly. An intramuscular injection, however, may result in pain, agitation, and subsequent arterial desaturation.

Propofol is also an effective induction agent in congenital cardiac patients. It does need to be titrated in patients with limited cardiac reserve because it causes a decrease in mean arterial pressure and SVR. In patients with single-ventricle shunt-dependent physiology, propofol induction and maintenance have been found to cause an increased right-to-left shunt with significantly decreased PBF ( Williams et al., 1999 ). Another concern with propofol has been the association with severe metabolic acidosis after prolonged infusions described in children in the intensive care unit. This is rarely a problem if administered for less than 12 hours, but propofol kinetics are altered in infants recovering from cardiac surgery. An increased volume of distribution and reduced metabolic clearance after surgery cause prolonged elimination ( Rigby-Jones et al., 2002 ).

The application of EMLA cream at the site of intravenous cannula insertion facilitates cannulation and minimizes patient pain and stress. EMLA is commonly applied 60 minutes before intravenous catheter placement when an intravenous induction is believed to be most suitable.

An inhalation induction is generally well tolerated and is the preferred approach in children without intravenous access. It should also be noted that parental presence in the operating room at the induction of anesthesia is an increasingly common practice. The use of inhalation inductions with parents present is easily orchestrated and well tolerated by the patient and observing parent.

Differential anesthetic uptake among patients with cyanotic versus acyanotic defects is common. Patients with reduced PBF have a delay in anesthetic uptake (see Chapter 6 , Pharmacology of Pediatric Anesthesia). In extreme cyanosis, sevoflurane, because of its reduced solubility, may not achieve an adequate alveolar concentration to fully induce anesthesia. Halothane, because of its greater solubility, would be more efficacious in extreme cyanosis. In more conventional congenital cardiac patients, inhalation induction with halothane or sevoflurane can easily and safely be performed and the effect of right-to-left shunting on uptake and distribution is not clinically significant.

In patients who are at risk for right-to-left shunting and systemic desaturation, oxygenation is well maintained with a good airway and ventilation, even with halothane-induced hypotension ( Greeley et al., 1986) . Skilled airway management and effective ventilation are essential and take precedent over drug selection during anesthetic induction. It is essential to understand the complexities of shunts and vascular resistance changes, but airway, ventilation (CO2), and oxygen effects on the cardiovascular system are of primary importance during the induction of anesthesia.

After anesthetic induction, intravenous access is established or a larger, more appropriate-sized indwelling intravenous catheter is placed. A nondepolarizing muscle relaxant is usually administered, and an intravenous opioid and/or inhalation agent is chosen for maintenance anesthesia. The child is preoxygenated with 100% oxygen, and an endotracheal tube is carefully positioned. Preoxygenation is done even in the ductal-dependent patient with increased PBF; this avoids desaturation during intubation. If the child arrives in the operating room with an endotracheal tube in place, replacement should be considered because inspissated secretions in a tube with a small internal diameter can cause significant obstruction to gas flow. During CPB, when humidified ventilation is discontinued, airway secretions increase and endotracheal tube obstruction can occur. This effect can be minimized by starting with a new endotracheal tube.

Due to the diverse array of congenital heart defects and surgical procedures, an individualized anesthetic management plan is essential. The maintenance of anesthesia in these patients depends on the age and condition of the patient, the nature of the surgical procedure, the anticipated duration of CPB, and the need for postoperative ventilatory support. Choice of a particular anesthetic agent is less important when the appropriate monitors are used and adherence to the physiologic guidelines mentioned earlier are met. More important than the specific anesthetic techniques and drugs is the skilled execution of the anesthetic plan, taking into account patient response to drugs, changes associated with surgical manipulation, and early recognition of intraoperative complications. In patients with complex defects requiring preoperative inotropic and ventilatory support, a carefully controlled induction and maintenance anesthetic with a potent opioid is usually chosen. In patients with a simple ASD or small perimembranous VSD, a potent inhalation agent alone or in combination with moderate opioid dosages is preferred as the principal anesthetic agent. This allows for extubation in the operating room or shortly after arrival in the intensive care unit and a less prolonged period of intensive care monitoring ( Kloth and Baum, 2002 ). The reported changes in blood pressure and heart rate for the inhalation agents in normal children are observed in pediatric cardiac surgical patients as well.

Although sevoflurane, desflurane, halothane, and isoflurane decrease blood pressure in neonates, infants, and children, the vasodilatory properties of isoflurane and sevoflurane may improve overall cardiac output compared with those of halothane ( Murray et al., 1986 ; Russell et al., 2001 ). Despite improved cardiac reserve with isoflurane and desflurane, the incidence of laryngospasm, coughing, and desaturation during induction of anesthesia limits their use as an induction agent in children with congenital heart defects ( Friesen and Lichtor, 1983 ). Children with complex CHD and limited cardiac reserve require an anesthetic technique that provides hemodynamic stability. Inhalation agents are less well tolerated as a primary anesthetic in patients who have limited cardiac reserve, especially after CPB. Fentanyl and sufentanil are excellent induction and maintenance anesthetics for this group of patients.

Low to moderate doses of these opioids can be supplemented with incremental doses of inhalation anesthetics. The advantage of adding low concentrations of inhalation agents is a shortened period of postoperative mechanical ventilation while maintaining the advantage of intraoperative hemodynamic stability. Clearly, postoperative mechanical ventilation is required when a high-dose opioid technique is used. The hemodynamic effect of fentanyl at a dose of 25 mcg/kg with pancuronium given to infants in the postoperative period after operative repair of a congenital heart defect shows no change in left atrial pressure, PA pressure, PVR, and cardiac index and a small decrease in SVR and mean arterial pressure ( Hickey et al., 1985a ). Higher doses of fentanyl at 50 at 75 mcg/kg with pancuronium results in a slightly greater fall in arterial pressure and heart rate in infants undergoing repair for complex congenital heart defects ( Hickey and Hansen, 1984 ). Fentanyl has also been shown to block stimulus-induced pulmonary vasoconstriction and contributes to the stability of the pulmonary circulation in neonates after congenital diaphragmatic hernia repair ( Hickey et al., 1985a ). The use of fentanyl appears to stabilize the pulmonary vascular responsiveness in newborns and young infants with reactive pulmonary vascular beds and to be helpful in weaning from CPB and stabilizing shunt flow.

Sufentanil and pancuronium provide the same cardiovascular stability as fentanyl and pancuronium in pediatric cardiovascular patients. Children receiving a sufentanil induction as a single dose of 5 to 20 mcg have a stable preintubation period ( Moore et al., 1985 ; Greeley et al., 1987a ). Intubation and other stimuli such as sternotomy do not produce clinically significant alterations in hemodynamics, although changes are greater than with equipotent doses of fentanyl. Its use as an infusion produces fewer alterations in heart rate and blood pressure, which are particularly important in infants where marked hemodynamic changes are poorly tolerated. For neonates with critical CHD, a sufentanil anesthetic and continued postoperative infusion have been shown to reduce morbidity after cardiac surgery and to be superior to a halothane anesthetic and routine morphine postoperatively ( Anand and Hickey, 1992 ). The blunting of the stress response observed in this study was believed to account for the differences in morbidity. Although high-dose intraoperative opioids followed by postoperative infusion had been the preferred approach in the 1980s and 1990s, evidence suggests no real advantage to high-dose opioids in reducing the stress response compared with moderate-dose opioids. Lower doses facilitate early extubation and limit the need for inotropes in the postoperative period ( Gruber et al., 2001 ).

Alfentanil and remifentanil are short-acting potent opioids that have been used for cardiac surgery in children and show some promise in pediatric anesthesia cases because of their short elimination half-life and hemodynamic stability. A significantly slower heart rate has been observed in children anesthetized with remifentanil compared with fentanyl ( Friesen et al., 2003 ).

As a primary anesthetic in children undergoing CPB, remifentanil and alfentanil must be administered via continuous infusion due to their short half-lives. When the infusion is discontinued, the patient's plasma concentration falls rapidly, particularly with remifentanil, and patients require supplements of longer-acting opioids such as fentanyl or morphine. The use of remifentanil and alfentanil is generally limited to repairs where early extubation in the operating room is planned. Remifentanil is an ultrarapid-acting opioid. Its unique metabolism by plasma and tissue esterases elimination makes it a more predictable drug. CPB can dramatically alter the pharmacokinetic profile of a drug. In adults, Hug and others (1994) have shown that CPB prolonged the elimination half-life of alfentanil and increased its central volume of distribution and volume of distribution at steady state. In a study of children undergoing CPB, Davis and others (1999) noted that for remifentanil, its volume of distribution and elimination half-life were unaffected by CPB.

Because of the widespread use of opioids in pediatric cardiac surgery and the availability of invasive monitoring, the pharmacokinetics and pharmacodynamics of these drugs have been well studied ( Davis et al., 1987 ; Greeley et al., 1987a ). In general, the clinical pharmacology of fentanyl and that of sufentanil share the same age-related pharmacokinetic and pharmacodynamic features ( Table 17-7 ). Furthermore, sequential sufentanil anesthetics in neonates with CHD show marked increases in clearance and elimination between the first week and the third or fourth week of life ( Greeley et al., 1988 ). The latter observation is most likely due to maturational changes in hepatic microsomal activity and improved hepatic blood flow from closure of the ductus venosus. The variability in clearance and elimination, coupled with limited cardiovascular reserve in the neonate during the first month of life, makes opioid dosing difficult in this age group. Careful titration of 5 to 10 mcg of fentanyl or 1 to 2 mcg of sufentanil or a continuous infusion technique provides the most reliable method of achieving hemodynamic stability and an accurate dose-response. CPB, different institutional anesthetic practices, and individual patient differences all influence pharmacokinetic and pharmacodynamic disposition of the opioids in ways that are not predictable. Even certain disease states, such as tetralogy of Fallot, alter pharmacokinetic processes ( Koren et al., 1986 ).

TABLE 17-7   -- Sufentanil pharmacokinetics in pediatric cardiovascular patients[*]

Age Group

t½a (min)

t½b (min)

Clearance (mL/kg per min)

Vdss (L/kg)

1–30 days

23 ± 17

737 ± 346

6.7 ± 6.1

4.2 ± 1.0

1–24 mo

16 ± 5

214 ± 41

18.1 ± 2.7

3.1 ± 1.0

2–12 yr

20 ± 6

140 ± 30

16.9 ± 2.2

2.7 ± 0.5

12–18 yr

20 ± 6

209 ± 23

13.1 ± 0.4

2.7 ± 0.5

t½a, Slow distribution half-life; t½b, elimination half-life; Vdss, volume of distribution at steady state.



All values are mean ± SD.


An additional consideration is the success with circuit miniaturization and heparin-coated oxygenators and circuits ( Darling et al., 1998 ; Olsson et al., 2000 ; Ozawa et al., 2000 ). As the circuit prime volume is reduced, the dilutional effects of bypass become less dramatic, and plasma concentrations of anesthetic drugs should be maintained at a higher level compared with earlier reports. Heparin and biological coatings designed to minimize activation of proinflammatory mediators and endothelial cell damage are applied to oxygenators and tubing. Reducing endothelial cell leakage should stabilize the patient's plasma volume and, by reducing renal and hepatic injury, improve drug clearance ( Jensen et al., 2003 ).

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier



The management of CPB in neonates, infants, and children differs substantially from the adult patient. Pediatric patients are exposed to more severe biological extremes, including deep hypothermia (15° to 20° C), hemodilution (two- to fourfold dilution of circulating blood volume), low perfusion pressures (20 to 30 mm Hg), wide variation in pump flow rates (ranging from highs of 200 m L/kg per minute to total circulatory arrest), and wide ranging blood pH management (e.g., controlled [alpha stat, pH stat] and uncontrolled [post-TCA or post-low-flow CPB, when pH can be extremely and unpredictably low]). These physiologic stresses alter autoregulatory function. In addition to these prominent changes, subtle variations in glucose supplementation, cannula placement, presence of aortopulmonary collaterals, age, and size may impair effective perfusion during CPB.

Adult patients are rarely, if ever, exposed to these biological extremes. Temperature is rarely lowered below 25°C, hemodilution is more moderate, perfusion pressure is generally maintained at 30 to 80 mm Hg, flow rates are maintained at 50 to 65 mL/kg per minute, and pH management strategy is less influential because moderate hypothermic temperatures are commonly used, whereas deep hypothermia with or without circulatory arrest is rarely used ( Table 17-8 ). Variables, such as glucose requirements and patient size, are more consistent in adults. Venous and arterial cannulas are larger and less deforming of the atria and aorta, and their placement is more predictable. Although superficially similar, the conduct of CPB in children is considerably different from that of adults. One would expect marked physiologic differences in the response to CPB in the child in general and in neonates and young infants in particular.

TABLE 17-8   -- Differences between adult and pediatric cardiopulmonary bypass




Hypothermic temperature

Rarely below 25° to 32°C

Commonly 15° to 20°C

Total circulatory arrest

Rarely used

Used but less common

Pump priming dose



Blood volume dilution

25% to 33%

100% to 200%



Blood, albumin, FFP

Perfusion pressures

50 to 80 mm Hg

20 to 50 mm Hg

Influence of alpha stat versus pH stat management strategy

Minimal at moderate hypothermia

Marked at deep hypothermia

Measured PaCO2 differences

30 to 45 mm Hg

20 to 80 mm Hg

Glucose regulation




Rare: requires significant hepatic injury

Common: reduced hepatic glycogen stores


Frequent: generally easily controlled with insulin

Less common: rebound hypoglycemia may occur



Modified after CPB separation



Clinical Application of Cardiopulmonary Bypass in Children

The CPB circuit must replace the function of both the heart and the lungs during cardiac surgery. Because of the use of nonpulsatile flow and the need for reduced perfusion flow rates to minimize blood return to the heart, hypothermia is required. Specifically, hypothermic CPB is used to preserve organ function during cardiac surgery and to prevent end-organ ischemia through a reduction in cellular metabolism and preservation of high-energy phosphate stores. As temperature is lowered, both basal and functional cellular metabolism is reduced and the rate of adenosine triphosphate (ATP) and phosphocreatine consumption is substantially reduced ( Michenfelder and Theye, 1968 ; Swain et al., 1991 ). At deep hypothermia, cellular metabolism is so low and membrane fluidity reduced to such an extent that cellular basal metabolic needs and cellular membrane integrity can be maintained for a relatively prolonged period of time with minimal or no flow. Additional benefits include decreased production of putative mediators, a reduction in the rate of calcium influx into the cell due to the decreased fluidity of cellular membranes, and reduced rate of gene activation. These form the basis of the protective effects of deep hypothermia that allow for the implementation of low-flow deep hypothermic CPB and DHCA.

The degree of hypothermia selected is dependent on the need for reduced flow to enhance surgical repair. In general, three distinct methods of CPB are used: moderate hypothermia (25° to 32°C), deep hypothermia (15° to 20°C), or DHCA. The technique selected is based on the required surgical conditions, the patient size, the type of operation, and the potential physiologic impact on the patient.

Moderate hypothermic CPB (MHCPB) is the principal method of bypass used for older children and adolescents. In these patients, venous cannulas are less obtrusive and the heart can easily accommodate superior and inferior venae cavae (SVC and IVC, respectively) cannulation. Bicaval cannulation reduces right atrial blood return and improves the surgeon's ability to visualize intracardiac anatomy. The large cannulas used in older children are rigid and less likely to kink. Most surgeons are willing to cannulate the IVC and SVC in neonates and infants. However, in neonates and infants, this is technically more difficult and likely to induce brief periods of hemodynamic instability. In addition, the pliability of the cava and the rigidity of the cannulas may result in caval obstruction, reduced venous drainage, and elevated venous pressure in the mesenteric and cerebral circulations. When moderate hypothermia is used, perfusion flow rates must be high to meet metabolic demands of the patient. Recommendations for optimal pump flow rates for children are based on the patient's body mass and maintaining efficient organ perfusion as determined by arterial blood gases, acid-base balance, and whole body oxygen consumption during CPB ( Fox et al., 1982 ; Hickey and Andersen, 1987 ). Table 17-9 lists recommended, albeit arbitrary, normothermic flow rates for children based on body weight. At hypothermic temperatures, metabolism is reduced and therefore pump flow rates may be reduced. A physiologic basis for low-flow CPB based on measurements of cerebral metabolism is presented later in this chapter.

TABLE 17-9   -- Recommended pump flow rates for cardiopulmonary bypass in children

Patient Weight (kg)

Pump Flow Rate (mL/kg per min)















Deep hypothermic CPB is generally reserved for neonates and infants requiring complex cardiac repairs. However, certain older children with complex cardiac disease or severe aortic or mitral valve regurgitation may require deep hypothermic temperatures. For the most part, deep hypothermia is selected to allow the surgeon to operate under conditions of low-flow CPB or total circulatory arrest. Low-pump flows improve the operating conditions for the surgeon by providing a near bloodless field and generally allow the use of a single atrial cannula, facilitating better visualization of atrial anatomy and operative repairs performed through a right atriotomy.

Deep hypothermic CPB with total circulatory arrest allows the surgeon to remove the atrial and aortic cannula. With this technique, surgical repair is more precise because of the bloodless and cannula-free operative field. Arresting the circulation, even at deep hypothermic temperatures, introduces the question of how well deep hypothermia preserves organ function, with the brain being of greatest concern. Extensive clinical experience using DHCA has suggested the duration of a safe circulatory arrest period to be approximately 30 to 45 minutes, or longer ( Kirklin and Barratt-Boyes, 1986 ; Wypij et al., 2003 ).

Formal experimental data evaluating the effect of DHCA on organ function are limited; however, organ preservation during the arrest period is primarily a function of hypothermia. Hypothermia preserves organ function by maintaining cellular ATP stores despite reduced delivery, reducing excitatory neurotransmitter release and preventing calcium entry into the cell even though energy-dependent calcium pumps are depleted of ATP stores. Normothermic ischemic injury and hypothermic protection are discussed in more detail in Deep Hypothermic Circulatory Arrest: Protection from Ischemic Injury.


Pathophysiologic changes that occur during and after pediatric CPB relate to the nonendothelialized bypass circuit and oxygenator, hypothermia, the degree of hemodilution, nonpulsatile perfusion, the age of the patient, the preoperative myocardial substrate, the length of the ischemic period on the myocardium (cross-clamp time) or the entire body (low-flow CPB or total circulatory arrest), the type of anesthesia used, and the exaggerated inflammatory response present in the young child. The effects on the patient are both global and organ specific. Global hormonal and metabolic responses have been characterized as the stress response to hypothermic CPB.

General Effects of Cardiopulmonary Bypass

The release of a large number of metabolic and hormonal substances, including catecholamines, cortisol, growth hormone, cytokines, prostaglandins, leukotriene complements, glucose, insulin, beta endorphins, and other substances, characterize the stress response during hypothermic CPB ( Hindmarsh et al., 1984 ; Greeley et al., 1986b, 1988 [162] [163]). The likely causes for the elaboration of these substances include contact of blood with the nonendothelialized surface of the pump tubing and oxygenator ( Bui et al., 1991 ), nonpulsatile flow, low perfusion pressure, hemodilution, and hypothermia. Other factors that may con tribute to elevations of stress hormones include delayed renal and hepatic clearance during hypothermic CPB, myocardial injury, and exclusion of the pulmonary circulation from bypass. The lung is responsible for metabolizing and clearing many of these s tress hormones. The stress response generally peaks during rewarming from CPB. There is evidence that the hormonal component of the stress response can be blunted, but not eliminated, by increasing the depth of anesthesia (Anand et al., 1987, 1990 [12] [13]). Recent evidence suggests that the stress response in neonates and infants has less of an effect on outcome than was described in the early 1990s ( Gruber et al., 2001 ).

It is unclear at what level elevated circulating stress hormones, normally an adaptive response, become detrimental. There is little question that these substances do mediate undesirable effects such as myocardial damage (catecholamines), systemic and pulmonary hypertension (catecholamines), pulmonary endothelial damage (cytokines, complement, eicosanoids, prostanoids), and pulmonary vascular reactivity (cytokines, prostanoids). The benefits of blunting the release of stress hormones and catecholamines with fentanyl in premature infants undergoing PDA ligation have been observed ( Anand et al., 1987 ). Further, neonates with complex CHD who died in the postoperative period demonstrate much higher hormonal and metabolic responses during the intraoperative and postoperative periods than did survivors with similar cardiac defects (Anand et al., 1987, 1990 [12] [13]).

The measured levels of stress hormones and catecholamines in neonates are generally an order of magnitude greater than those measured during CPB in the adult. Although blunting the extremes of the stress hormone response seems warranted, there is additional evidence suggesting that the newborn stress response, especially the endogenous release of catecholamines, may be an adaptive metabolic response necessary for survival at birth ( Langercrantz and Slotkin, 1986 ). This suggests that complete elimination of an adaptive stress response may not be desirable, but modification of its extremes probably is desirable. To what extent acutely ill neonates with CHD are dependent on their stress response for maintaining hemodynamic stability during and after CPB is unknown.

The importance of adaptive versus maladaptive responses is even more relevant today. Earlier data from the late 1980s and early 1990s supported the concept that high-dose opioids blunt the stress hormone release and improve patient outcome (Anand et al., 1987, 1990 [12] [13]). This was viewed as a significant advance to CPB management and necessary to preserve end-organ function in neonates and infants exposed to CPB. Reappraisals of the benefits of high-dose opioids to blunt the stress response to bypass for neonates have occurred. Data from Duncan and others (2000) suggest that moderate dosing of opioids (25 to 50 mcg/kg of fentanyl) is equally effective as the formerly recommended high-dose regimen (100 mcg/kg of fentanyl) with regard to blunting the stress response to CPB ( Duncan et al., 2000 ). In another study by Gruber and others (2001) , three different anesthetic regimens were evaluated: bolus fentanyl, fentanyl infusions, and a combination of fentanyl plus midazolam. All three regimens demonstrated high levels of stress hormone release. In fact, the magnitude of epinephrine release was 8 to 15 times higher than that reported in the study of Anand and others from a decade earlier. These findings suggest that catecholamines and stress hormones may not be as important as inflammatory mediators such as cytokines and leukotrienes. Despite this marked elevation in stress hormone release, there was no evidence of increased morbidity and mortality, The fact that lower doses of opioids have similar benefits as high-dose opioids in neonates and infants undergoing CPB suggests that the evolution of the heart-lung machine with reduced prime volume and heparin- or biocompatible material-coated oxygenators and circuits, the use of modified ultrafiltration to remove mediators, and improved prebypass and postbypass management strategies are having a positive impact on the outcome of neonates and infants undergoing CPB. The role of stress hormone modification appears lessimportant today as a promotor for the activation of proinflammatory mediators.

Distinguishing maladaptive stress hormone release from adaptive release is the presence of mediators of systemic inflammation and endothelial injury. Complement activation, neutrophil activation, and release and activation of tumor necrosis factor and interleukins 1 and 6 have been well described during hypothermic CPB ( Downing and Edmonds, 1992 ; Butler et al., 1993 ). These mediators of systemic inflammation, in conjunction with ischemia-reperfusion injury and the direct effects of hypothermia, account for organ injury during and after CPB. The main target of many of these mediators is the vascular endothelium. Endothelial injury results in altered microcirculatory function, which is responsible for elevations in PVR, cerebrovascular resistance (CVR), and SVR, commonly seen after hypothermic CPB. Endothelial injury impairs release of important vasodilators such as nitric oxide (NO) and prostacyclin and enhances the effect of vasoconstrictors such as endothelin ( Aaronson et al., 2002 ). Direct endothelin antagonists such as BQ123 have been shown to reduce PVR after congenital heart surgery. Vascular endothelial damage associated with CPB is a prominent factor in post-CPB lung injury ( Schulze-Neick et al., 2002 ).

In addition to these properties, the endothelial surface, pulmonary endothelium in particular, is responsible for metabolizing vasoconstrictors such as angiotensin, catecholamines, and eicosanoids (Aaronson et al., 2002 ). Injured endothelium, by virtue of reduced production of NO and impaired metabolism of mediators of vasoconstriction, promotes vasoconstriction. Endothelial cells also play an important regulatory role in water and solute transport. Abnormalities in endothelial function promote increased capillary permeability and increased interstitial edema ( Greeley et al., 1988 ).

Evidence for microcirculatory dysfunction with nonpulsatile perfusion can be found in several studies. Ogata et al. (1960) directly observed that capillary flow in the omentum slowed and virtually ceased during nonpulsatile CPB at normothermia. At flow rates of 60 and 75 mL/kg per minute, nonpulsatile flow resulted in lower total body oxygen consumption, lower pH, and an increased base deficit than with pulsatile perfusion. Matsumato and others (1971) showed that at 37°C, nonpulsatile perfusion produced capillary sludging, dilation of the postcapillary venules, and increased edema formation in the conjunctival and cerebral microcirculation. In contrast, pulsatile CPB maintained capillary blood flow in the omentum, eliminated sludging in the conjunctival and cerebral microcirculation, and reduced jugular venous lactate levels ( Matsumato et al., 1971 ; Geha et al., 1973 ). Studies of hypothermic low-flow CPB in a canine model demonstrated that converting low-flow nonpulsatile CPB at 25 mL/kg per minute to pulsatile perfusion improved brain pH, Pco2, and Po2 ( Watanabe et al., 1989 ). Measurements of CMRO2 in neonates, infants, and children demonstrate that nonpulsatile perfusion accounts for a 9% reduction in brain metabolism ( Greeley et al., 1991) . Pulsatile perfusion may therefore improve cerebral perfusion at both normothermic and hypothermic temperatures. When low-flow CPB is used, the addition of pulsatile flow may provide improved microcirculatory perfusion and allow for better oxygen delivery to tissue at lower flow rates. Mechanistically, a lack of pulsatility alters the biomechanical forces exerted on the endothelium. This results in rather rapid changes in ion conductance, adenylate cyclase activity, and intracellular free calcium levels. These changes are similar to receptor-mediated changes in vascular tone seen with α- and β-receptor activity. If more prolonged exposure to nonpulsatile flow exists, changes in vascular tone may be augmented by release of local regulatory mediators such as endothelin.

Although improvements in microvascular flow and organ metabolism have been suggested with the addition of pulsatile flow, clinical studies have been inconclusive. Murkin and others (1987) observed a 13% improvement in CBF and cerebral oxygen consumption with pulsatile CPB in adults. In children, a small clinical trial of pulsatile CPB demonstrated no improvement in glucose, insulin, and cortisol responses during CPB ( Mori et al., 1987 ). Although trials have been limited, it appears that the benefits of pulsatile perfusion would occur in patients with limited organ reserve or those exposed to more extreme perfusion variables such as low-flow CPB or DHCA. To date, clinical assessment of pulsatile perfusion during low-flow CPB or DHCA, which are more common to the pediatric patient, has not been elucidated fully. Also, the method in which pulsatility is mechanically produced is important. To be effective, the pulse wave must duplicate the mechanics of a normal cardiac cycle. Several pulsatile systems generate wave forms that are significantly different from a physiologic heartbeat.

During high-flow hypothermic CPB, peripheral vascular resistance increases in all vascular beds throughout the body. Flows to the kidneys, gastrointestinal tract, and brain are decreased and flow is preferentially shunted toward skeletal muscle ( Lazenby et al., 1981 ). This is quite different from the intact patient who is not exposed to CPB but instead is surface cooled. These patients demonstrate an increase in peripheral vascular resistance and preferential flow to the brain, heart, and kidneys in response to a decrease in cardiac output. This difference between the hypothermic patient exposed to CPB and the surface-cooled hypothermic patient is dependent on the effects of hypothermia on cardiac output. During hypothermic CPB, when flow rates are maintained by the pump at high perfusion rates (flow rates of 3.0 L/min), blood is shunted away from the vital organs to the skeletal muscle. The vasculature of skeletal muscle serves as a large capacitance bed for excessive flow during high-flow hypothermic CPB. During low-flow hypothermic CPB, skeletal muscle vasculature constricts and total body blood flow is redistributed toward vital organs. Vital organ blood flow is maintained and provides effective oxygen delivery to maintain identical oxygen consumption at both full flow and after a 50% reduction in perfusion flow rates ( Lazenby et al., 1981 ). In patients with aortopulmonary shunts (a common finding in patients with cyanotic heart disease who are undergoing CPB), there is a greater redistribution of blood flow away from the gastrointestinal tract and kidneys ( Mavroudis et al., 1984 ). This excessive shunting may contribute to the higher incidence of end-organ injury observed in infants with large aortopulmonary shunts or aortopulmonary collaterals when exposed to prolonged periods of CPB.

pH Stat and Alpha Stat

The role of CO2 management in CPB has been studied both experimentally and clinically. Based on the effect of CO2 on arterial and intracellular pH at hypothermic temperatures, two divergent blood gas management strategies have been championed: alpha stat (temperature uncorrected) and pH stat (temperature corrected) ( Norwood et al., 1979 ; Fox et al., 1984 ; Govier et al., 1984 ; Miyamoto et al., 1986; Swain et al., 1991 ). The concept of alpha stat and pH stat blood gas strategies is based on the effect temperature has on the balance between intracellular O H- and H+ ion concentrations. The intracellular balance between O H- and H+ concentration, electrochemical neutrality, is essential to maximize enzyme efficiency. At normothermia (37°C), intracellular electrochemical neutrality occurs at pH 7.40. As the temperature of the cell is decreased, however, maintaining electrochemical neutrality requires an increase in the O H- ion concentration relative to the H+ ion concentration (alkalosis). When an alkalotic milieu is not maintained during hypothermia, many investigators believed that enzymatic function is altered and intermediary metabolites become uncharged, more lipid-soluble, and freely diffusible across cellular membranes. The net effect would be altered intracellular metabolism. Maintaining an alkaline intracellular pH was believed to enhance cellular function. Swain and colleagues (1990, 1991) [447] [448] performed NMR studies in a piglet model and used sophisticated probes to assess intracellular pH differences between alpha stat and pH stat models. They found that the intracellular pH was in fact not dramatically different between the two blood gas management strategies. This finding put the alpha stat/pH stat issue of optimal management strategy into question. The blood gas management strategy that maintains an alkaline pH is the alpha stat strategy. The alpha stat strategy is implemented by maintaining pH 7.40 measured at 37°C without correction for the effects of temperature. This strategy allows the pH of the blood and the intracellular pH to become increasingly alkalotic during cooling through a reduction in intracellular H+ ion concentration. In doing so, electrochemical neutrality is preserved. The reason uncorrected blood gas measurement is called alpha stat is because as body temperature cools, the normal blood buffers (NH3 HCO3-, etc.) become ineffective. At hypothermic temperatures, intracellular and blood proteins become the only functional buffers. Buffering is necessary to decrease the concentration of H+ ions during cooling. The alpha imidazole ring of the amino acid histidine accounts for the majority of buffering capacity at hypothermic temperature and is the reason why uncorrected blood gas management strategy is referred to asalpha stat.

Blood pH becomes increasingly alkalotic as it is cooled. To correct for this alkalotic “pH,” CO2 can be added to maintain a temperature-corrected pH of 7.40. The addition of CO2 was believed to lowerintracellular pH (pHi), resulting in an imbalance between H+ and OH- ions, that is, the loss of electrochemical neutrality. This approach of adding CO2 to correct for the effect of cooling on blood pH is termed pH stat. Although pH stat was believed to alter cellular enzymatic function by lowering pHi, and altering electrochemical neutrality, it appears that cellular enzyme function is not impaired with pH stat. However, both pH and alpha stat markedly affect microcirculatory pH and CBF.

The current controversy over blood gas management strategy in congenital heart surgery remains unresolved. Two key areas have been critically evaluated using these techniques: (1) the neurologic effects of the two strategies and (2) myocardial function. Neurologic considerations relate to the effects of CO2 on CBF, brain cooling, cerebral emboli, and the effect of pH on microcirculatory function. The pH stat strategy, by virtue of adding CO2 during cooling, increases CBF. During cooling to deep hypothermia when low-flow bypass or circulatory arrest is being used, optimal brain cooling remains pivotal for cerebral protection. With this reasoning, pH stat may seem a more effective strategy. In contrast, we have already alluded to the concern of microembolic phenomenon and the fact that air emboli may be a significant contributor to neurologic morbidity, particularly in the congenital cardiac population, where left-sided air is a common concern. Increasing brain blood flow may also increase air emboli to the brain ( Sungurtekin et al., 1999 ). In addition, if circulatory arrest is contemplated, starting the arrest period with a lower blood pH undoubtedly results in an even lower blood pH after the arrest period. The effect of a lower post-circulatory arrest pH on cerebral injury also must be considered.

In the mid 1990s, two key investigations in animals were performed evaluating alpha stat and pH stat using the percentage of cerebral metabolic recovery after circulatory arrest compared with the prearrest value as a marker for cerebral injury. In the first study by Skaryak and others (1995) , piglets were cooled using alpha stat, pH stat, or pH stat followed by conversion to alpha stat strategy before undergoing a 90-minute period of circulatory arrest. The animals cooled with pH stat had a lower cerebral metabolic recovery than the alpha stat group, suggesting a greater injury. The crossover group cooled initially with pH stat and then converted to alpha stat before circulatory arrest had the greatest recovery of CMRO2 of the three groups. These data imply that the pH stat does have a cooling benefit compared with the alpha stat, but the lower pH nullifies the benefit in animals with normal circulation. Switching over to alpha stat for a period of time before arrest optimizes pH before circulatory arrest and therefore maximizes the benefits of both strategies. The negative impact of low pH on the brain at deep hypothermia was also demonstrated by Ekroth and others (1989), when it was demonstrated that an increase in creatine kinase-BB bands occurred as the arterial pH fell below 7.1 (an uncorrected measurement using a pH stat strategy).

In a second series of experiments using a piglet model similar to that of Skaryak, Kirshbom and others (1995) surgically placed a 4-mm left subclavian-to-PA shunt on the animals before bypass and then ligated the shunt in half of the animals. The shunt was placed to simulate the effects of aortopulmonary collaterals. Collateral blood vessels are commonly present in cyanotic patients. The animals were then cooled with either a pH or an alpha stat strategy. The animals with a ligated shunt behaved similarly to those in the previous study. In contrast, the shunted animals cooled with alpha stat had significantly lower CBF, less effective brain cooling, and a significantly lower rate of metabolic recovery compared with the pH stat-cooled animals. The implication of this study was to demonstrate that pH stat strategy may have its greatest benefit in patients with cyanotic heart lesions. Cyanotic congenital cardiac defects have a tendency to develop aortopulmonary collateral blood vessels. The blood gas management strategy selected may not be equally applicable to all patients. Cyanotic patients may benefit from pH stat, whereas acyanotic lesions experimentally benefit from alpha stat. Other factors that were not evaluated by these studies include the role of intracardiac defects such as VSDs. VSDs may contribute to a greater opportunity for air to reach the left side of the circulation. With pH stat, cerebral air emboli may increase.

Hiramatsu and others (1995) evaluated the rate of pH recovery in an animal model using deep hypothermic circulatory arrest. In their study, the authors demonstrated a more rapid rate of recovery of blood pH with pH stat than with alpha stat. Although pH recovered more rapidly, the pH achieved immediately after circulatory arrest was substantially lower with pH stat. The question raised is, What is worse for the brain—the lowest pH or the rate of pH recovery?

In a clinical series designed to demonstrate a benefit for pH stat over alpha stat, Bellinger and others (2001) could not show a significant difference in neurologic outcome between alpha stat and pH stat blood gas strategies. Interestingly, a small subgroup of 16 patients with VSD had a significantly lower developmental outcome with pH stat compared with alpha stat. It was the only subgroup demonstrating a significant neurologic difference between the cooling strategies, and it favored alpha stat. The only clinical study to demonstrate a neurologic benefit to the pH stat strategy was a retrospective study of patients undergoing the Senning procedure for TGA ( Jonas et al., 1993) . These patients were operated on in the late 1970s and early 1980s. All of the patients were repaired using DHCA, and the average cooling time before initiation of circulatory was 14.5 minutes. Based on current knowledge, cooling times of less than 20 minutes have an increased risk of incomplete cerebral cooling using an alpha stat strategy ( Kern et al., 1992) . The brief cooling periods used in this study, even using a pH stat strategy, would never constitute optimal cooling and optimal cerebral protection based on current knowledge. Conclusions supporting pH stat strategy based on this analysis fail to account for the suboptimal approach to cerebral cooling.

These studies reinforce the benefits of a blood gas management strategy and one must take into account multiple factors, including bypass cooling strategy, the cardiac lesion, the presence of aortopulmonary collaterals, the presence of VSDs, the use of circulatory arrest, etc. In general, a consensus on the appropriate management strategy remains elusive. Most centers believe that cyanotic patients benefit from a pH stat strategy. Based on the data outlined earlier, alpha stat or a crossover strategy may be better for noncyanotic patients.


All organs are at risk for hypoxic-ischemic injury, but the brain is the most sensitive and therefore the limiting factor when using DHCA. In contrast to the heart, where cardioplegic arrest has improved myocardial performance, a combination of hypothermia and nonpulsatile perfusion is the only modality of cerebral protection available. To understand the protective effects of hypothermic protection, a discussion of the pathophysiology of normothermic ischemia is warranted.

At normothermic temperatures, the energy-rich compounds (ATP and phosphocreatine) are maintained through oxidative metabolism. A majority of the ATP and phosphocreatine that are produced are utilized for maintaining ion homeostasis. In fact, it is estimated that 50% to 75% of high-energy phosphate expenditure is for the maintenance of transmembrane ionic gradients ( Astrup, 1982 ; Hansen, 1985 ; Ericinska and Silver, 1989 ). Arresting the circulation at normothermia results in a rapid depletion of high-energy phosphate stores ( Norwood et al., 1979a ). After 2 minutes of complete ischemia, ATP levels fall to 10% of prearrest values ( Norstrom and Siesko, 1978 ). In association with ATP depletion, there is a release of excitatory neurotransmitters such as glutamate and aspartate ( Benveniste et al., 1984 ; Hagberg et al., 1987a, 1987b [168] [169]). Neurotransmitter release is not specific to ischemia; it can occur with other cerebral insults, such as hypoglycemia ( Benveniste et al., 1984 ; Hagberg et al., 1987a, 1987b [168] [169]). Neurotransmitter release is thought, however, to represent a stereotypic response to ischemic injury.

Neurotransmitter release adversely affects membrane ionic permeability ( Mutch and Hansen, 1984 ). ATP depletion in concert with excitatory neurotransmitter release signals a dramatic alteration in the maintenance of transmembrane ionic gradients. Electrochemical gradients for potassium, calcium, and sodium are lost, presumably due to unrestricted ion permeability across cell membranes ( Hansen and Zeuthen, 1981 ). The loss of ionic gradients does not indicate a breakdown of cell membrane integrity, and although ATP levels are reduced, there is evidence to suggest that dysfunction of the energy-dependent ionic pumps does not occur within the first few minutes of normothermic ischemia ( Mies and Paschen, 1984 ; Wieloch et al., 1984 ). However, low ATP levels may prevent the reestablishment of transmembrane ion gradients after the initial ion flux.

Calcium influx is the harbinger of permanent cellular damage. Approximately 95% of the calcium present in the extracellular space moves into the cell during the period of increased membrane permeability ( Hansen and Zeuthen, 1981 ). Calcium influx results in accelerated cellular damage through the activation of calcium-dependent enzymes (phospholipases, nucleases, and xanthine oxidase). Phospholipases C and A2 release free fatty acids (FFAs), such as arachidonic acid, from cell membrane phospholipids ( Tang and Sun, 1985 ). FFAs uncouple oxidative phosphorylation and inhibit the exchange of adenosine diphosphate (ADP) for ATP across mitochondrial membranes. Arachidonic acid is metabolized to prostaglandins through the cyclooxygenase pathway and to leukotrienes through the 5-lipoxygenase pathway during post-ischemia-reperfusion. Ischemia alters the composition of prostaglandin production ( Gaudet et al., 1980 ; Adesuyi et al., 1985 ). The potent vasoconstrictors prostaglandin GF2a (PGF2 a) and thromboxane A2 are produced in favor of vasodilators, such as prostacyclin (PGI2). Thromboxane A2 also promotes platelet aggregation, resulting in small vessel thrombosis.

Leukotrienes are undetectable in the nonischemic brain ( Gaudet et al., 1980 ; Adesuyi et al., 1985 ). However, with ischemia and reperfusion, leukotriene levels and cytokines (tumor necrosis factor2, interleukins, etc.) increase dramatically ( Moskowitz et al., 1984 ). Like prostaglandins, leukotrienes are potent cerebral vasoconstrictors. In addition, leukotrienes are mediators of “secondary ischemic damage” through increased capillary permeability and promote leukocyte entry into the ischemic tissue. The probable result is an amplification of FFA-mediated cell injury, cerebral edema, activation of coagulation cascade, platelet plugging, and accelerated vascular thrombosis.

Nucleases become active after ischemia and have been implicated in creating single-stranded breaks in DNA. A majority of nucleases are calcium dependent, and thus calcium entry is an important cofactor for nuclease activity ( Tullis and Rubin, 1982 ). In general, single-stranded DNA breaks are easily repaired ( Ward et al., 1985 ). However, extensive single-stranded regions are prone to secondary breaks in the presence of oxygen free radicals. The conversion of single-strand breaks to double-strand breaks is considered lethal to the cell ( Bryant, 1985 ; Radford, 1985 ).

The enzymatic breakdown of ADP and AMP not only wastes energy but also contributes to hypoxic cell damage ( Guitierrez, 1991 ). AMP is either dephosphorylated to adenosine or deaminated to inosine monophosphate. Adenosine crosses the cell membrane through facilitated diffusion and acts as a potent local vasodilator. It is through adenosine release that the cell attempts to improve local oxygen delivery. During severe hypoxia, however, both inosine monophosphate and intracellular adenosine can be metabolized to hypoxanthine. Hypoxanthine in the presence of the enzyme xanthine oxidase is converted to the free radicals O2 - and H2O2 -, important mediators of ischemic cerebral injury. Free radicals are catalyzed to potent oxidizing species that target proteins, unsaturated membrane lipids, and DNA. The result is extensive damage to cell membranes, nucleic acids, and enzyme systems ( Aisen, 1980 ; Freeman and Crapo, 1982 ). A flow chart describing this cascade is shown ( Fig 17-8 ).


FIGURE 17-8  Flow chart describing the cellular events occurring during normothermic arrest.



Hypothermia protects the brain from ischemic injury through preservation of high-energy phosphate stores, prevention of excitatory neurotransmitter release, restriction of membrane permeability, and prevention of calcium entry into the cell. By retarding ischemic injury, reperfusion can proceed uneventfully.

Several investigators examined the protective effect of deep hypothermia by measuring high-energy phosphate compounds with 31P nuclear magnetic resonance (31P-NMR) spectroscopy ( Norwood et al., 1979a ; Stocker et al., 1986 ; Chopp et al., 1989 ; Sutton et al., 1991 ; Jonas, 1993 ). These investigators reported that although ATP was rapidly depleted at normothermic temperatures, ATP levels were maintained at deep hypothermic temperatures (15° to 20°C) for a more prolonged period of time. At hypothermic temperatures, the rate of energy-dependent cellular enzyme systems such as Na+,K+-ATPase and Ca2+-ATPase are drastically slowed. ATP and phosphocreatine utilization is reduced, excitatory neurotransmitter release is blunted, and ion homeostasis is maintained through both energy-dependent and -independent mechanisms. At hypothermic temperatures, ischemic events do not proceed concurrently.

Norwood and others (1979a) demonstrated that 25 minutes of DHCA significantly lowers creatine phosphate levels, but ATP stores are well maintained, in isolated perfused rat brains. In contrast, larger animal studies suggest that ATP levels reach their nadir after 21 to 33 minutes of hypothermic circulatory arrest ( Norwood et al., 1979 ; Sutton et al., 1991 ; Swain et al., 1991 ; Jonas, 1993 ). After 1 hour of circulatory arrest, 31P-NMR measurements suggested a delay in ATP recovery. In sheep, after 60 minutes of arrest at 15°C, ATP levels were reduced to 36% of control. Thirty minutes of normothermic reperfusion, however, restored ATP levels to 83% of control values ( Swain et al., 1991 ). In a similar CPB model using piglets, intracellular pH did not recover from a 60-minute arrest period until after 40 minutes of normothermic reperfusion ( Jonas, 1993 ). ATP and phosphocreatine levels recovered to 90% and 98% of baseline but required 3 hours of normothermic reperfusion. These studies of whole brain cerebral metabolism in both animals and children demonstrate a significant reduction in metabolism after DHCA, which is not found after low-flow CPB. Both cellular levels of ATP and global measures of cerebral metabolism are reduced after DHCA, suggesting organ dysfunction.

Studies of ischemic brain injury in the rat by Busto and others (1987, 1989) [65] [66], using a four-vessel occlusion model, demonstrated that ischemic injury results in an increased release in both glutamate and dopamine. When temperature was lowered from 36° to 33°C, the expected rise in glutamate release did not occur and dopamine levels decreased. Interestingly, minimal levels of hypothermia (34°C) have been shown to prevent the ischemic neuronal injury on the CA-1 layer of the hippocampus compared with normothermic controls. Moderate hypothermia of 27°C was no more protective than this slight level of hypothermia (Busto et al., 1987, 1989 [65] [66]). This suggests that excitatory neurotransmitter release may act as an accelerator of neurologic injury by promoting the loss of transmembrane ionic gradients and enhancing calcium entry. More moderate levels of hypothermia may be sufficient to prevent the triggered release of excitatory neurotransmitters but do not reduce brain metabolism or alter transmembrane permeability to the same degree as more extreme hypothermic temperatures.

Deeper levels of hypothermia may provide additional protection from calcium entry through altering membrane fluidity. At deep hypothermic temperatures, cell membranes alter their permeability through changes in physical state of membrane lipids—in other words, less liquid and more semisolid ( Rich and Langer, 1982 ). This may directly affect free ion movement across cellular membranes and provide additional protection once ATP-dependent mechanisms for ion homeostasis are lost.

As discussed later in this chapter, more efficient and rigorous cooling strategies, substitution of low-flow CPB for DHCA, reperfusion periods during the arrest period, higher hematocrits, pH stat cooling in select patients, and improved rewarming strategies may minimize the risk of low-flow CPB and DHCA and reduce neuropsychological injury. It should also be noted that to date, there has been no drug study that has demonstrated improved neurologic outcome in patients or animals undergoing CPB.


CPB desensitizes cardiac β ARs. Schwinn and others (1991) examined the effects of β-agonists on adenyl cyclase activity in canine left ventricular tissue and found that maximal isoproterenol stimulation resulted in a marked increase in β-receptor-mediated adenyl cyclase activity in the pre-CPB period. After 155 minutes of CPB, reexposure to the same isoproterenol infusion resulted in a significant decrease in β AR-stimulated adenyl cyclase activity. Thirty minutes after bypass, adenyl cyclase activity was greater than on prebypass measurements. Similar responses were obtained with submaximal infusions of isoproterenol and the β2-selective drug zinterol. When β-receptor density was examined, it was found to be unchanged during CPB, suggesting that the reduction in function is related to the uncoupling of the β-receptor and adenylyl cyclase, which is the function of the Gs protein complex. It is interesting to note that 30 minutes after weaning from CPB, β-receptor number did begin to decrease, suggesting that β-receptor downregulation may play a role in the postoperative response to β-specific inotropic agents and, at least in older patients, downregulation occurs fairly quickly.

Schranz and others (1993) studied the effects of CPB on acyanotic pediatric patients and confirmed the experimental data obtained by Schwinn (1991, 1994) [417] [418]. They demonstrated that after CPB, β-agonist-induced increases in cAMP were attenuated. However, when several non-β-receptor-dependent stimulators of adenylyl cyclase were examined, adenylyl cyclase activity increased in a normal fashion despite CPB. In addition, β1- and β2-receptor density was found to be unaltered by CPB. These studies suggest a primary role for the Gs protein complex in the desensitization of β-agonist action when weaning from CPB. Neonatal myocardium does not desensitize in response to high levels of circulating catecholamines. Several animal studies have demonstrated receptor signaling is different in early neonatal life. Instead of β-agonist administration producing desensitization of responses, it promotes receptor signaling by enhancing expression and/or catalytic efficiency of adenyl cyclase (Giannuzzi et al., 1995 ). Neonatal myocytes normally have a reduced number of β-receptors. Catecholamine responsiveness is reduced in the neonate. The lack of desensitization in the neonatal population means prolonged preoperative or postoperative catecholamine administration does not reduce myocardial responsiveness to these important inotropic agents.

Myocardial Protection

Although normal neonatal hearts may be more resilient to periods of ischemia, the effects of cyanosis and CPB offset these advantages. For this reason, most pediatric cardiac surgeons believe that the basic principles of myocardial preservation practiced during adult cardiac surgery should be followed in neonatal cardiac surgery.

Research in the field of myocardial preservation demonstrated that oxygenated blood is a better cardioplegic solution than crystalloid (Buckberg, 1979, 1990 [61] [62]). Red blood cells contain large quantities of free radical scavengers and blood proteins contain histidine, an amino acid buffer that reduces intracellular acidosis during ischemia ( Jennings and Reimer, 1983 ). Histidine is particularly important because it continues to function even at profound hypothermic temperatures ( Bretschneider, 1980 ). In addition, the presence of red blood cells may increase oxygen delivery to myocardial tissue. Cold cardioplegia is generally preferred to warm cardioplegia. The rationale is similar to cerebral protection—preserving high-energy phosphate stores, preventing excitatory neurotransmitter release, restricting membrane permeability, and preventing calcium entry into the cell. Potassium is an important component of the cardioplegia solution because it maintains electrochemical quiescence. The nonbeating heart has decreased substrate utilization and allows for repletion of ATP and more homogeneous delivery of nutrients throughout the coronary system ( Danforth et al., 1960 ; Domalik-Wawrzynski et al., 1987 ). This results in better global myocardial protection.

Other factors that have been shown to be beneficial in cardioplegia solutions include leukocyte depletion, reduced calcium content, the addition of the amino acids glutamate and aspartate, and oxygen radical scavengers ( Lazar et al., 1980 ; Rozencrantz et al., 1982 ; Julia et al., 1988 ; Breda et al., 1989 ; Bolling et al., 1990 ; Chiba et al., 1993 ). Leukocytes and calcium play an important role in reperfusion injury. Glutamate and aspartate are amino acids that are intermediary metabolites of glycolysis and thereby provide substrate during reperfusion. Oxygen radical scavengers prevent membrane damage from ischemia-induced hydroxyl radicals. Blood cardioplegia has sufficient naturally occurring free radical scavengers to obviate the need for the addition of free radical scavengers to cardioplegia solution ( Julia et al., 1988 ) ( Table 17-10 ). Clermont and others (2002) indicated that systemic free radical activation may be a major contributor to myocardial oxidative stress related to CPB.

TABLE 17-10   -- Generic components of cardioplegia




Crystalloid vehicle


Electromechanical quiescence


Buffering capacity






Provides substrate during reperfusion

Blood (hematocrit 15% to 20%)

Provides increased oxygen delivery, free radical scavenging, and pH buffering

Magnesium/lidocaine and mannitol

Decreased reperfusion injury

Low ionized calcium

Decreased reperfusion injury

Leukocyte depletion

Decreased reperfusion injury



Some centers are adding low-concentration lidocaine to blood cardioplegia solutions. The rationale for this practice is to lengthen the time between cardioplegia doses, and there is some theoretical evidence of early sodium channel closure and less myocardial intracellular calcium damage during bypass. This lengthening of the inter-cardioplegia administration time points may allow a distinct surgical timing advantage in some small infants undergoing extensive aortic arch reconstructions. The rationale for this practice is shown largely in animal experimental data. Sunamori and others (1982) showed in dogs that a cardioplegia solution of lidocaine and magnesium had better myocardial preservation than cardioplegia without lidocaine. Clinical trials are needed, because a recent randomized trial in adult patients undergoing coronary revascularization has found no statistically significant improvement in myocardial preservation after the addition of lidocaine to a standard cardioplegia solution ( Rinne et al., 1998) .

Although blood cardioplegia appears beneficial, many surgical groups continue to use crystalloid cardioplegia in neonates with good success. This is particularly true when DHCA or continuous low-flow CPB at deep hypothermic temperatures is being used as part of the CPB management scheme. Deep hypothermic temperatures of 15° to 18°C are below typical cardioplegia temperatures of 22°C and provide myocardial protection.

The one exception to the use of cardioplegia is during the stage 1 repair of HLHS. Because of the risk of needle damage to the small ascending aorta, antegrade cardioplegia is avoided and cardioplegia is either not given or administered retrograde rather than prograde after aortic cross-clamping. In cases where cardioplegia is not used, the heart is protected by DHCA alone. Virtually all other patients receive cardioplegia.

One other factor in myocardial preservation comes from a study of myocardial troponin T levels in neonates and children undergoing CPB with either alpha stat or pH stat. In this study ( Nagy et al., 2003 ), there is a suggestion that a pH stat cooling regimen may provide improved myocardial protection compared with alpha stat cooling regimens.


During CPB, hypothermia is the most important factor that alters cerebral hemodynamic and metabolic parameters. Hypothermia produces a marked reduction in both CBF and brain metabolism (CMRO2) at constant pump flow rates ( Fig. 17-9 ). The coupling of flow to metabolism is an important concept in ensuring adequate oxygen delivery and limiting luxuriant perfusion to the brain. Variations in flow, metabolism, and their coupling are dramatically altered by hypothermic CPB ( Julia et al., 1988 ; Greeley et al., 1988, 1989, 1991a, 1991b [156] [157] [158] [159]).


FIGURE 17-9  Effects of hypothermia on cerebral blood flow (CBF) and cerebral metabolism. There is a positive linear relationship between hypothermia and cerebral blood flow during pediatric cardiopulmonary bypass. In contrast, there is an exponential reduction in cerebral metabolism with temperature reduction during hypothermic cardiopulmonary bypass. CMRO2, cerebral metabolic rate for oxygen.



Cerebral Blood Flow

CBF falls in a direct linear relationship with temperature (see Fig. 17-9 ). In studies where CO2, perfusion pressure, pump flow rate, and temperature were altered by the extracorporeal circulation, temperature is the most important factor influencing CBF during CPB in children (Greeley et al., 1988, 1989 [156] [157]; Julia et al., 1988 ). Moderate hypothermia and deep hypothermia have different effects on CBF and its autoregulation.

Pressure-flow autoregulation, or the ability to maintain a constant CBF despite wide ranges in mean arterial pressures, has been shown to be intact during MHCPB (26° to 30°C) in adults and children when measured using alpha stat blood gas regulation ( Govier et al., 1984 ; Murkin et al., 1987 ). In children, pressure-flow autoregulation remains intact over a range of mean arterial pressures of 15 to 80 mm Hg during MHCPB ( Greeley et al., 1988 ). During moderate hypothermia, the cerebral vasculature maintains a normal physiologic response of dilation during low perfusion pressure and constriction when perfusion pressure is high. In contrast, at deep hypothermic temperatures of 15° to 20°C, pressure-flow autoregulation is lost ( Greeley et al., 1989a ). At deep hypothermic temperatures, CVR increases with temperature reduction. CVR remains high even when pump flow rates and perfusion pressure are substantially reduced. Reduction in pump flow rates from 100 mL/kg per minute to as low as 30 mL/kg per minute does not significantly change CVR during deep hypothermic CPB. This loss of pressure-flow autoregulation is most likely due to the influence of deep hypothermic temperatures on vascular reactivity. Severe temperature reductions impair vascular relaxation ( Civalero et al., 1962 ; Tanaka et al., 1988 ; Greeley et al., 1989a ; Kern et al., 1991b ). This has been described as a cold-induced “cerebrovasoparesis” (Greeley et al., 1989a, 1989b [165] [164]).

Cerebral Metabolism

CBF decreases in a linear fashion with reduction in brain temperature, but brain metabolism (CMRO2) decreases exponentially ( Greeley et al., 1991b ; Milde, 1992 ) (see Fig. 17-9 ). A convenient expression of the effect of temperature on CMRO2 is to calculate the ratio of metabolism at a temperature gradient of 10°C, called the temperature coefficient, or Q10 ( Michenfelder and Theye, 1968 ). Cerebral oxygen consumption has been measured in a number of models during CPB (dog, monkey, and human) and has been shown to vary greatly between species and at differing ages within species (Bering, 1961 ; Michenfelder and Theye, 1968 ; Greeley et al., 1991b ; Croughwell et al., 1993) . In children and adults, Q10s of 3.65 and 2.8 have been reported, respectively ( Greeley et al., 1991 ;Croughwell et al., 1993) .

The increased metabolic suppression for younger patients may be due to more efficient cooling of the immature neurons and glial elements to hypothermia or may reflect greater brain mass as a percentage of body weight and more efficient brain cooling. Interspecies and intraspecies variability for Q10 may explain why variables other than temperature have been implicated as major contributors to cerebral protection during CPB. If adult-derived Q10 data are used, temperature-induced metabolic suppression would appear insufficient to explain clinically acceptable “safe” circulatory arrest periods.

Hypothermic protection alone may account for the majority of the protection seen during DHCA ( Greeley et al., 1991b ). Other variables, such as anesthetic agents, provide much smaller contributions to cerebral protection, once deep hypothermic temperatures (15° to 20°C) are reached ( Michenfelder and Theye, 1968 ; Michenfelder, 1988 ). At more moderate temperatures, anesthetic agents and other cerebroprotective agents, such as calcium channel blockers, barbiturates, and N-methyl-D-aspartate antagonists, may be beneficial. If deep hypothermia is the only cerebroprotective agent used, factors that enhance cerebral cooling by modifying CBF, such as the addition of CO2, may be important adjuncts to achieving uniform brain cooling and thereby improving global cerebral protection ( Michenfelder and Theye, 1968 ; Michenfelder, 1988 ; Kern et al., 1991a, 1992a [222] [223]). These considerations are discussed later in this chapter.

Cerebral Blood Flow and Metabolism Coupling

CBF decreases linearly with reduction in temperature. In contrast, cerebral metabolism decreases exponentially with reduction in temperature. The flow-metabolism ratio must increase with decreasing temperature during CPB in children. In the awake healthy child, CBF and metabolism (CMRO2) are regulated by the metabolic needs of regional areas of the brain. This has been termed cerebral flow-metabolism coupling and is an important regulatory feature of cerebral homeostasis ( Kety, 1945 ; Scheinberg and Stead, 1949 ; Stullken et al., 1977) . In humans, a mean CBF of 45 to 80 m L/100 g per minute is coupled to a CMRO2 of 3.0 to 4.0 m L/100 g per minute, for a CBF/CMRO2 ratio of 13 to 20:1 ( Michenfelder, 1988 ; Greeley et al., 1991a ). In neonates, CMRO2, CBF, and the CBF/CMRO2ratios are generally higher than those for older children and adults. This is believed to be due to increased metabolic demand for neuronal growth, myelinization, etc. ( Rosenberg et al., 1982 ). If CPB is managed using alpha stat blood gas regulation at a pump flow rate of 100 m L/kg per minute, the ratio of CBF to CMRO2 increases with decreasing temperature, so that during MHCPB, the CBF/CMRO2ratio increases to 30:1. At deep hypothermic temperatures the ratio of the CBF to CMRO2 extends to 75:1 ( Greeley et al., 1991a ). In contrast, pH stat blood gas regulation (the addition of CO2 to the gas flow mixture) results in the CBF/CMRO2 ratios of 60:1 at moderate hypothermia ( Murkin et al., 1987 ), whereas at deep hypothermic temperatures, flow-metabolism ratios using pH stat strategy are unknown. Although alpha stat regulation has been believed to maintain flow-metabolism coupling at moderate hypothermia, CBF becomes increasingly luxuriant at lower temperatures in children, even using alpha stat blood gas regulation. Luxuriant flow becomes important when low pump flows are used in conjunction with deep hypothermic CPB.


Guidelines for the safe implementation of low-flow CPB are not firmly established. Estimates for minimal acceptable pump flow rates (PFR) for children during CPB have been suggested based on metabolic measurements. Using Q10 data, one can predict CMRO2 at different temperatures. For children at 37°C, the mean CMRO2 = 1.48; at 28°C, the mean CMRO2 = 0.51 (66% reduction); at 18°C, the mean CMRO2 = 0.16 (89% reduction); and at 15°C, the mean CMRO2 = 0.11 (93% reduction) ( Greeley et al., 1991a ). By comparing the reduction in CMRO2 with proportional reductions in pump flow rates, an estimate of minimal acceptable flow rates can be predicted ( Kern et al., 1993 ). The equation describing this relationship for infants is as follows:

MPFR(T) = e · 1171(T - 37°C) · (100 mL/kg per minute)

where MPFR(T) = minimal pump flow rate at temperature T 100 mL/kg per minute = Normothermic pump flow rate e · 1171(T - 37°C) = CMRO2 at temperature T and at 37°C

Table 17-11 shows the calculated values derived from the MPFR(T) equation.

TABLE 17-11   -- Predicted minimal pump flow rates (MPFR)

Temperature (°C)

CMRO2 (mL/100 g per min)

Predicted MPFR (mL/kg per min)

























CMRO2, cerebral metabolic rate for oxygen.




Both human and animal studies suggest that these data represent a reasonable approximation of acceptable minimal pump flow rates during hypothermic CPB. Swain and others (1991) using 31P-NMR have demonstrated that flow rates of 10 mL/kg per minute at 15°C for periods of up to 2 hours in 8-week-old lambs maintain normal levels of ATP and phosphocreatine and normal brain pH. When flow is reduced to 5 mL/kg per minute at 15°C, organic phosphates become depleted and brain pH begins to fall. The effect of 5 mL/kg per minute flow rates on cerebral metabolism in these animals was indistinguishable from findings in a similar group of sheep undergoing 2 hours of circulatory arrest at 15°C. A calculated MPFR(T) of 7 mL/kg per minute is necessary to meet cerebral metabolic demands at this temperature.

In a neonatal piglet model, flow rates of between 5 and 10 mL/kg per minute at 18°C for 60 minutes resulted in only a modest reduction in metabolism (10%) after weaning from CPB compared with measurements obtained in the prebypass period ( Mault et al., 1993 ; Kern et al., 1993 ). These flow rates are just below the predicted flow rates of 11 mL/kg per minute, suggesting that these animals had borderline cerebral oxygen delivery for 18°C. Similar data in canines, monkeys, and humans have been derived and show surprisingly similar relationship between pump flow rate and metabolism at a variety of hypothermic temperatures ( Fox et al., 1982 ; Miyamoto et al., 1986 ; Watanabe et al., 1989 ; Kern et al., 1993 ).


Wernovsky and others (1992) demonstrated a higher incidence of transient postoperative seizures after 45 to 60 minutes of circulatory arrest compared with low-flow continuous perfusion. Their study results suggest that DHCA imparts a greater neurologic risk to neonates and infants compared with continuous-flow bypass. Bellinger and others (2003) reported on an 8-year neurologic assessment of 155 patients who had undergone the arterial switch procedure for TGA using either a low-flow bypass technique or a predominant DHCA strategy. In this study, full-scale IQ scores, academic achievement, memory, problem solving, visual-motor integration, and neurologic examination did not differ between the two groups. Children assigned to the circulatory arrest group performed statistically worse on motor function tests, including manual dexterity with the nondominant hand, speech apraxia, visual motor tracking, and phonologic awareness. Children who were assigned to the low-flow bypass group were associated with a more impulsive response style on continuous performance tests of vigilance and worse behavior as rated by teachers. The circulatory arrest group evaluated at 8 years out continued to demonstrate a greater motor impairment compared with the low-flow group. The low-flow group demonstrated greater behavioral problems and impulsive behavior. These findings suggest that methods to improve CPB management and the approach to DHCA could be improved based on the physiologic consequences outlined earlier. Furthermore, the approach used to implement and manage circulatory arrest in this study does not conform to, or adequately address, current techniques of brain protection and monitoring. Newer monitoring and protective strategies may significantly affect postarrest neurologic injury.

Deep Hypothermic Circulatory Arrest With Intermittent Perfusion Periods

Intermittent systemic perfusion between periods of DHCA has been suggested as an alternative to prolonged periods of DHCA. These periods of reperfusion may replete cerebral high-energy phosphate stores and preserve neurologic tissue. Data investigating this concept are limited. Reperfusion periods of 1 minute at deep hypothermic temperatures between two 30-minute circulatory arrest periods significantly improved metabolic recovery compared with a 60-minute period of circulatory arrest in neonatal piglet studies ( Mault et al., 1992 ) ( Fig. 17-10 ). This is consistent with NMR data, which suggest that high-energy phosphate compounds reach their nadir after approximately 30 minutes ( Swain et al., 1991 ). In the study by Swain and others, sheep were exposed to a 30-minute period of reperfusion following a 60-minute period of circulatory arrest. There was partial restoration of intracellular pH, ATP, and phosphocreatine levels. A second 60-minute period of circulatory arrest, however, resulted in rapid depletion of ATP and phosphocreatine levels. After a second 60-minute period of circulatory arrest, ATP, phosphocreatine, and intracellular pH values were no different than after a continuous 2-hour arrest period. These studies support the conclusion that 30 minutes of arrest at deep hypothermic temperatures does not result in full depletion of cellular ATP and phosphocreatine. Replenishing cellular metabolic stores may occur more rapidly at hypothermic temperatures, especially if cellular levels of ATP are not fully depleted before reinstituting perfusion. In contrast, after 60 minutes of total circulatory arrest, a 30-minute reperfusion period at normothermia only partially restores the brain's metabolic reserve. Reperfusion after arrest periods of 20 to 30 minutes seems to have a greater benefit than reperfusion after 60 minutes of circulatory arrest. Several additional animal studies have demon strated a benefit to intermittent perfusion ( Langley et al., 1999 ; Strauc et al., 2003 ). Clinical data, however, remain elusive. Several clinical reports have described methods for both antegrade and retrograde cerebral perfusion principally in adults undergoing surgery for aortic arch aneurysm or dissection ( Mass et al., 1997 ; Yoshii et al., 2003 ). Several studies comparing selective antegrade or retrograde cerebral perfusion in adults with ascending aortic arch disease were unable to demonstrate neurologic outcome improvement compared with circulatory arrest ( Svensson et al., 2001 ). In one study of 289 adults undergoing ascending aortic arch surgery, DHCA was compared with selective antegrade cerebral perfusion. There was no statistical difference in neurologic outcome, but the selective perfusion group did have less renal dysfunction and were extubated earlier in the postoperative period ( DiEusanio et al., 2003 ).


FIGURE 17-10  Percent cerebral metabolic rate for oxygen (CMRO2) recovery after varying lengths of circulatory arrest in a piglet model. The longer the arrest period, the less recovery there is of CMRO2. When a 60-minute period of circulatory arrest is interrupted for 1 minute and the animal is reperfused at 100 mL/kg per minute of flow, CMRO2 returns to near-normal levels.



Cerebral Cooling Strategies

Accelerated rates of cooling (>1°C/min) during CPB using alpha stat regulation are associated with a lower developmental quotient in neonates undergoing DHCA ( Bellinger et al., 1988 ). High cerebral oxygen extraction, low jugular venous bulb saturation, and a cerebral metabolic rate greater than that expected for 18°C suggest that inadequate cerebral perfusion from inefficient brain cooling is associated postoperatively with neurologic disability ( Greeley et al., 1991b ). Marked variability in cerebral cooling has been established in several clinical studies. When low jugular venous saturations were used as a marker for incomplete cerebral cooling, one third of patients demon strated significantly slower brain cooling. When cooling techniques were varied, colder blood resulting in a larger temperature gradient between the blood and tissue contributed to more rapid and complete cerebral cooling ( Kern et al., 1991a ).

Temperature gradients occur between superficial temperature measurement and deep brain structures. Studies in the rat demonstrate temperature gradients of 2° to 6°C between deep brain structures and temporalis muscle in a study using a four-vessel occlusion model of cerebral ischemia ( Busto et al., 1987 ; 1989). Significant differences in neuronal function, histopathology, free fatty acids, and excitatory neurotransmitter release were demonstrated based on regional differences in brain temperature (Busto et al., 1987, 1989 [65] [66]; Okada et al., 1988 ). Animal studies also suggest that there are regional differences in the distribution of brain blood flow during hypercarbia. Deep brain structures (thalamus, brainstem, and cerebellum) receive a significantly greater percentage of CBF than do cortical structures ( Hansen et al., 1984 ).

The ideal blood gas management strategy for children is not categorical. Just as the surgeon must decide the appropriate temperature for hypothermic bypass and whether to use moderate-flow, low-flow, or circulatory arrest, the appropriateness of a blood gas strategy depends on many modifiers. These modifiers include the degree of hypothermia, pump flow rate, use of DHCA, and cooling dynamics of the brain. Appropriate strategies can be hypothesized. During moderate hypothermia, selection of one blood gas management strategy over the other appears less critical because blood pH differences are small ( Swan, 1984 ; Bashein et al., 1990 ). During deep hypothermia with or without circulatory arrest, the addition of CO2 during active brain cooling could potentially improve the distribution of the cold perfusate to deep brain structures. Work by several investigators ( Watanabe et al., 1989 ; Jonas, 1993 ; Skaryak et al., 1995 ) suggests that pH stat management enhances the distribution of extracorporeal perfusate to the brain and may help cool the brain more thoroughly and rapidly. Although improved cooling was demonstrated in these studies, metabolic recovery after circulatory arrest was shown to be impaired, suggesting that the acid load induced by pH stat had a negative effect on enzymatic or microcirculatory function after cerebral rewarming. To retain the benefits of pH stat on cooling and to eliminate its negative effects on enzymatic function suggest the use of a combined blood gas management strategy with pH and alpha stat in succession. In another study, a group of animals underwent initial cooling with pH stat followed by alpha stat to eliminate residual acid load before the initiation of DHCA ( Skaryak et al., 1995 ). This group demonstrated improved metabolic suppression over alpha stat alone and a significant enhancement in metabolic recovery after rewarming. This suggests that initial cooling with pH stat, followed by alpha stat, may be preferable. A controlled clinical study evaluating this approach (combined pH/alpha stat cooling), however, is necessary before clinical benefits can be supported.

Other factors that may result in maldistribution of pump flow away from the cerebral circulation and contribute to inefficient cerebral cooling include anatomic variants (large aortic to pulmonary collaterals) and technical problems (aortic and venous cannula placement) ( Spach et al., 1980 ; Kern et al., 1991a ). Cyanotic patients, with known aortopulmonary collaterals, may benefit from the cerebrovasodilation of CO2 during early cooling. Once cooled, however, if DHCA or deep hypothermia with low flow is planned, conversion to an alpha stat strategy before arrest may reduce postarrest cerebral acidosis.

At some centers, a period of hyperoxia is being used before initiation of DHCA. The rationale for this is based on studies of better neurologic histology after CPB in a group of animals exposed to hyperoxia and on the fact that dissolved oxygen is used by the brain to a greater extent than is oxygen bound to hemoglobin ( Nollert et al., 1999 ).

Hematocrit and Neurologic Outcome

In a recent study, Jonas and others (2003) undertook a single-center randomized trial to assess the effects of hemodilution during CPB on neurologic and outcomes data. The patients were maintained at a hematocrit of 21.5% versus 27.8% during hypothermic CPB. The lower-hematocrit group had higher serum lactates at 1 hour after bypass and a significantly greater increase in total body water on the first postoperative day. At 1 year of age, the lower-hematocrit group had significantly worse scores on the Psychomotor Development Index. This suggests that higher hematocrits may be beneficial during hypothermic CPB ( Jonas et al., 2003 ).


The use of stress-dose preoperative corticosteroids for infant CPB has increased during the past several years. Historically, methylprednisolone was added only to the pump prime. The addition of corticosteroids to the pump prime dates back to the 1950s. Administration of corticosteroids at the time CPB is initiated theoretically reduces efficacy, because it takes 6 to 8 hours to have a significant effect on modulation of the inflammatory response. In a neonatal piglet study, Lodge and others (1999) compared three groups of animals: those that received methylprednisolone 30 mg/kg administered intravenously 8 hours and 1.5 hours before bypass, those that received methylprednisolone administered in the pump prime, and a control group that did not receive steroids. The preoperative administration of steroids resulted in a significant improvement in lung compliance, alveolar-arterial Po2 gradient, and PVR. There was a small improvement when steroids were administered to the pump prime-only group ( Lodge et al., 1999 ). Other studies have involved the role of steroids on myocardial apoptosis. Although preoperative steroid administration was no different than administration to the prime alone (Pearl et al., 2002 ), steroids reduced the incidence of postcirculatory arrest, apoptosis, and cell death. In addition, preoperative steroids have also been shown to reduce total body edema, cerebral edema, and troponin I levels and improves CMRO2 recovery after DHCA ( Langley et al., 2000 ; Schwartz et al., 2003 ). Clinical studies have demonstrated a lower ratio of proinflammatory to anti-inflammatory cytokines and extravascular fluid after the administration of dexamethasone in adults ( El Azab et al., 2002 ; Fillinger et al., 2002 ). The role of corticosteroids in pediatric cardiac surgery is still being investigated.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier



Once the aortic and venous cannulas are positioned and connected to the arterial and venous limb of the extracorporeal circuit, bypass is initiated. The technique for initiating bypass varies depending on the size of the patient and the temperature of the perfusate.

In older children and adolescents, bypass is initiated slowly. The venous cannula is unclamped and blood is siphoned from the RA into the oxygenator via gravity drainage or occasionally by vacuum assist. When the more commonly used gravity drainage is instituted, the rate at which venous blood is drained from the patient is determined by the height difference between the patient and the oxygenator inlet, and the diameter of the venous cannula and line tubing. Venous drainage can be enhanced by increasing the height difference between the oxygenator and the patient. Venous drainage can be reduced by either decreasing the height difference between the oxygenator and the patient or by partially clamping the venous line. Vacuum-assist venous drainage (VAVD) is being used more frequently in pediatric CPB. It provides a gravity-independent means for maintaining venous return and can significantly increase venous blood return during CPB procedures. If perfusion vigilance for detecting air entrainment is adequately accounted for and arterial line filters are used, then VAVD is especially useful. With the current clinical trend toward smaller venous cannulas and lower prime and circuit volumes for neonatal CPB machines, VAVD is safe and increasingly becoming a standard for neonatal and infant CPB ( Lau et al., 1999 ; Jegger et al., 2003 ).

Once venous blood begins to accumulate in the oxygenator, the arterial pump is slowly started. Its speed is gradually increased until full flow is reached. If return is diminished, line pressure is high, or mean arterial pressure excessive pump flow rates must be reduced. High line pressure and inadequate venous return are usually due to malposition or kinking of the arterial and venous cannulas, respectively.

In neonates and infants, when deep hypothermia is used, the pump prime is cold (18° to 22°C). When the cold perfusate contacts the myocardium, heart rate immediately slows and contraction is severely impaired. The contribution to total blood flow pumped by the heart rapidly diminishes. To sustain adequate systemic perfusion at or near normothermic temperatures, the arterial pump must reach full flows quickly. A major difference in the initiation of bypass in neonates and infants versus older children is the speed in which full support must be achieved. One method for initiating CPB in infants is to begin the arterial pump first; once the aortic flow is ensured, the venous cannula is unclamped and blood is siphoned out of the RA into the inlet of the oxygenator. Flowing before unclamping the venous cannula prevents the potential problem of patient exsanguination if aortic dissection or malplacement of the aortic cannula has occurred. Pump flow rates are then rapidly increased to sustain systemic perfusion. Because coronary artery disease is usually not a consideration, the myocardium should cool evenly. Exceptions to even myocardial cooling occur with coronary anomalies such as anomalous left coronary artery from the PA or pulmonary atresia with the presence of sinusoids. When a cold prime is used, caution must be exercised in using the pump to infuse volume before initiating CPB. Infusion of cold perfusate may result in bradycardia and impaired cardiac contractility before the surgeon is prepared to initiate CPB.

Once CPB begins, it is essential to observe the heart. Ineffective venous drainage can rapidly result in ventricular distention. This is especially true in infants and neonates where ventricular compliance is low and the heart may be intolerant of excessive preload augmentation. If distention occurs, pump flow must be reduced and the venous cannula repositioned. Alternatively, the heart may be vented or a pump sucker placed into the RA.


When a patient is weaned from CPB, the heart is allowed to fill by partially clamping the venous return tubing and reducing the arterial inflow until adequate blood volume is achieved. Blood volume is assessed by direct visualization of the heart and measuring cerebrovascular pressure and right atrial or left atrial filling pressures. When filling pressures are adequate, the venous cannula is clamped and the arterial inflow is stopped. The arterial cannula is left in place so that a slow infusion of residual pump blood can be used to optimize filling pressures. Myocardial function is assessed through direct cardiac visualization, central venous pressure or intracardiac monitoring, and intraoperative echocardiography. In corrected physiology, the pulse oximeter can also be used as a crude measure of cardiac output (Oshita et al., 1989 ). Low saturations or the inability of the oximeter probe to register a pulse may be a sign of very low output and high systemic resistance ( Severinghaus and Spellman, 1990 ).

After the repair of some complex congenital heart defects, the anesthesiologist and surgeon may have difficulty separating a patient from CPB. Under these circumstances, a distinction must be made between residual anatomic disease, altered loading conditions on the heart induced by the repair or the effects of CPB on systemic and PVR, myocardial function, and pulmonary compliance. Anatomic problems should be evaluated by TEE. Having the patient leave the operating room with a clinically residual defect has been associated with a high mortality rate ( Ungerleider, 1998 ). Simultaneously, the anesthesiologist should begin optimizing hemodynamics and interpreting monitored data to appropriately treat physiologic abnormalities. Echocardiography is helpful in assessing right ventricular and left ventricular function and evaluating PA pressure. A supplemental assessment of cardiac function performed in the operating room is an intraoperative “cardiac catheterization.” This is done to assess isolated pressure measurements from the various chambers of the heart; catheter pull-back measurements are used to evaluate residual pressure gradients across valves, repaired sites of stenosis and conduits; and oxygen saturation data to look for residual shunts and assess cardiac output (SVo2) ( Gold et al., 1986 ). In combination with TEE, a complete intraoperative “picture” of a structural and functional evaluation of the postoperative cardiac repair can be obtained ( Gold et al., 1986 ; Ungerleider et al., 1989a, 1989b [464] [465]; Muhiudeen et al., 1991, 1992 [317] [318]; Stevenson et al., 1993). If structural abnormalities are found, the patient can be placed back on CPB and residual defects can be repaired before leaving the operating room. Leaving the operating room with a significant residual structural defect adversely increases patient morbidity and affects survival (Goldet al., 1986; Ungerleider et al., 1989a, 1989b [464] [465]; Muhiudeen et al., 1990, 1991, 1992 [318] [319] [320]). Functional problems such as ventricular dysfunction may also be identified with echocardiography. Once diagnosed, therapy can be directed to the specific problem.

Modified Ultrafiltration

The use of modified ultrafiltration (MUF) after weaning from CPB has been advocated by a number of centers as a mechanism for removing inflammatory mediators, pulmonary vasoconstrictors, and excessive fluid at the end of CPB ( Naik et al., 1991 ; Ungerleider, 1998 ; Hiramatsu et al., 2002 ). Animal and human studies using MUF have shown a consistent improvement in lung compliance, cerebral metabolic recovery, dilutional coagulopathy, hematocrit, and myocardial function ( Skaryak et al., 1995 ; Keenan et al., 2000 ; Ootaki et al., 2002 ). Although long-term hemodynamic benefits have not been demonstrated, the acute improvement may reduce the need for more extraordinary support such as leaving the chest open or the institution of extracorporeal membrane oxygenation or ventricular assist device support in more marginal postbypass patients ( Fig. 17-11 ).


FIGURE 17-11  (A) Schematic of cardiopulmonary bypass (CPB). (B) Schematic of CPB with modified ultrafiltration (MUF) in line but clamped out of circuit. (C) Schematic of CPB with MUF in progress. After weaning from CPB, the MUF circuit is opened by removing the clamp from the ultrafiltration portion of the circuit and placing a clamp above the venous reservoir. The roller pump of the MUF circuit is turned on, and pulls approximately 5% to 10% of total cardiac output from the arterial cannula. The blood is pumped through the ultrafilter, fluid and mediators are removed in the ultrafiltrate, and blood is returned to the patient's circulation through the venous cannula. A cross-clamp is placed above the venous reservoir to prevent additional blood from entering the reservoir. If intravascular volume falls, blood can be added to the circuit at the venous reservoir and returned to the patient through the traditional portion of the bypass circuit.



Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


In this section on management strategies after weaning from CPB, we first discuss functional abnormalities common to all neonates, infants, and children after reparative or palliative operations for CHD. These include PA hypertension, right ventricular dysfunction, and left ventricular dysfunction. With this as background, approaches to specific congenital cardiac defects are noted. Because recognizing residual anatomic defects is an important part of the management of congenital heart patients who are difficult to wean from CPB, anesthesiologists must understand the repair and be cognizant of potential structural defects that may persist or become apparent after reconstructive heart surgery. The type, extent, and tolerance of residual cardiac defects are dependent on the method of repair and on whether the operation results in a complete anatomic and physiologic correction using the patient's native tissue, an anatomic and physiologic repair requiring artificial material to provide normal blood flow patterns (for children who are missing cardiac structures and require artificial material such as a conduit or a baffle), or a palliative operation necessitated by missing or diminutive ventricular chambers.

In patients receiving an anatomic and physiologic correction, residual defects are less common and occur at obvious locations. Because blood flow patterns are normal, mild to moderate anatomic defects are better tolerated in the postoperative period.

Patients who receive an anatomic and physiologic reconstruction but require prosthetic material to restore normal blood flow patterns are more likely to have anatomic problems because of the interposition of artificial material within the heart or great vessels. Residual anatomic defects are somewhat less well tolerated in this group due to the abnormal loading conditions that existed in the prerepair period and because extracardiac conduits may be positioned in a location that differs from the natural location of the great vessel it replaces. This reduces the efficiency of ventricular ejection.

In patients who have only a single ventricle and require palliative surgery, reconstruction is based first on having the single ventricle provide balanced blood flow to both the pulmonary and systemic circulation. A second procedure (or series of procedures) uses the single ventricle as a systemic ventricle. PBF is provided by channeling all of the systemic venous return directly to the PAs. PBF is maintained by a passive gradient from the systemic veins through the lungs and into the LA. This procedure known as the modified Fontan procedure is now widely applied to a variety of anatomic variants of single ventricle ( Fontan et al., 1983 ). These patients are intolerant of residual anatomic gradients, regurgitant valves, and arrhythmias. They are also less responsive to physiologic problems after weaning from CPB.


Patients with elevated PVR are also quite sensitive to blood volume changes. A patient with PA hypertension has an increased right ventricular afterload. This may result in worsening right ventricular function. Under this circumstance, right ventricular end-systolic and diastolic volumes and pressures increase, resulting in a volume- and pressure-loaded RV. Increases in right ventricular pressure decrease coronary perfusion pressure to the RV, resulting in right ventricular ischemia. In addition, the volume-underloaded LV has decreased filling due to septal shift, resulting in decreased left ventricular stroke volume. Systemic perfusion is therefore quite dependent on left atrial preload, which requires a reduction in PVR to allow effective filling. Systemic hypotension may further affect right ventricular function by reducing coronary perfusion pressure.

Ventilation Strategies

Therapy for elevated PA pressures is aimed at lowering PVR and unloading the RV. Reductions in PVR are accomplished by altering ventilation, increasing inspired oxygen concentration, optimizing lung volumes, and using alkalinization, adequate sedation, nitric oxide, milrinone, occasionally α-blockers, and catecholamines.

Drummond and others (1981) showed that reducing Paco2and increasing pH produces a consistent and reproducible reduction in PVR in infants with PA hypertension. Manipulating serum bicarbonate levels to achieve a pH between 7.5 and 7.6, while maintaining a Paco2of 40 mm Hg, has equal salutary effects on PVR ( Malik and Kidd, 1973 ; Lyrene et al., 1985 ). The increases in both arteriolar Po2 and alveolar Po2 decrease PVR. With intracardiac shunts, changes in Fio2 have little effect on the Pao2. Thus, by inference, a reduction in PVR induced by increasing the inspired oxygen concentration probably is a direct pulmonary vasodilatory effect of PAo2 rather than Pao2. Experimental work by Custer and Hales (1985) showed that increased Fio2 is a more potent vasodilator in neonatal than in adult animals. Similarly, an increase in Fio2 has a vasodilatory effect on the pulmonary vascular bed of children with VSDs in the cardiac catheterization laboratory ( Lock et al., 1982 ). Ventilatory mechanics also play a major role in reducing PVR ( Nelson, 1966 ; Zapletal et al., 1976 ; Fagan, 1977 ). Newborns and infants have a high closing volume that is near functional residual capacity; therefore, at the end of a normal breath, some airway closure occurs ( Mansell and Bryan, 1972 ). This process results in areas of lung that are perfused and unventilated. As these lung segments become increasingly hypoxemic, a secondary hypoxic vasoconstriction occurs ( Rudolph and Yuan, 1966 ). The net effect is an increase in PVR. Careful inflation of the lungs to maintain functional residual capacity is beneficial to reduce PVR. In practice, because of increased anesthesia ventilator tubing compliance and endotracheal tube leaks, this may require a relatively large set tidal volume and the addition of 5 cm H2O positive end-expiratory pressure (PEEP). After bypass, there is generally a reduction in lung compliance of approximately 20% in neonates; a higher pressure is necessary to achieve the same delivered tidal volume in most neonates.

In neonates with rapid heart rates, right ventricular output can be impaired by extending the inspiratory time beyond two cardiac cycles. With a more prolonged inspiratory time, a third consecutive heartbeat during inspiration has a marked reduction in the right ventricular stroke volume. Because PBF occurs predominantly during the expiratory phase of the respiratory cycle, the ventilatory pattern should be adjusted to allow an adequate distribution of gas throughout the lung during inspiration and a more prolonged expiratory phase to promote blood flow through the lungs.

End-expiratory pressure must be applied cautiously in patients with CHD. Low levels of PEEP (3 to 5 mm Hg) prevent narrowing of the capillary and precapillary blood vessels and reduce PVR ( Kirklin et al., 1983 ; Taeed et al., 2001 ). Higher levels of PEEP or excessive mean airway pressure results in alveolar overdistention and compression of the capillary network in the alveolar wall and interstitium. This latter circumstance causes elevated PVR and reduced PBF ( Jenkins et al., 1985 ).

The final and perhaps the least well recognized use of mechanical ventilation is to assist in unloading of the RV. During inspiration, intrathoracic pressure increases. Inspiration may therefore assist right ventricular systole by creating an increased pressure gradient from the lung to the LA. This ventilatory assist is commonly seen in patients with PA hypertension or right ventricular dysfunction. Typically, an augmentation of the arterial pressure trace during inspiration is seen. This concept is very similar to the thoracic pump mechanism of CPR ( Weisfeldt and Halperin, 1986 ). The inspiratory assist must be balanced by the potential negative effects of increased mean airway pressure on PVR, that is, RV afterload. To maximize these cardiopulmonary interactions, inspiratory time should be limited to a time of two or fewer heartbeats, and mean airway pressure should be maintained at the lowest level possible to maintain effective lung volumes and functional residual capacity (FRC).

High-frequency jet ventilation (HFJV) is an uncommonly applied alternative mode of ventilation that has been successfully used in infants with pulmonary hypertension. HFJV eliminates CO2 very efficiently and does so at a lower mean airway pressure. In patients with elevated PA pressure, HFJV reduces PVR and right ventricular afterload. The benefit of a reduced RV afterload occurs at a reduced airway pressure. For example, in postoperative Fontan patients whose cardiac output is directly linked to PVR and mean airway pressure, HFJV significantly decreases mean airway pressure, reduces PVR, and increases cardiac index ( Andrew et al., 1987 ; Dietrich et al., 1993 ).

Another mode of ventilation, high-frequency oscillatory ventilation has been advocated for newborns with persistent pulmonary hypertension. It is an effective alternative mode of ventilation, but it may be poorly tolerated in the face of secondary right ventricular dysfunction as the high mean airway pressures required with this mode contribute to a right ventricular afterload and reduced right ventricular preload. If used, preload augmentation is required.

Pharmacology Strategies

Attempts to selectively manipulate PVR through pharmacologic interventions have been less than satisfying. Drugs that have shown the greatest promise both clinically and experimentally have been the phosphodiesterase inhibitors and NO.

Phosphodiesterase inhibitors, which have been used clinically and experimentally in treating elevations in PVR, include milrinone, enoximone, and amrinone. Some investigations have shown a reduction in PVR and SVR and an increase in right ventricular contractility ( Henriksson et al., 1979 ; Colon-Otero et al., 1987 ; Vieira et al., 1991 ). Phosphodiesterase inhibitors have varying effects on cAMP and cyclic guanosine monophosphate (cGMP). cGMP is a more important mediator of pulmonary vascular smooth muscle. Amrinone, which is less commonly used today, has a more prominent effect on increasing cAMP than cGMP. It has a profound effect on SVR. Significant systemic hypotension may be seen with an amrinone loading dose. For this reason, milrinone has become the drug of choice for PA hypertension; it has a more prominent effect on cGMP, which results in pulmonary vascular relaxation. These variations in effect suggest adrenergic receptor subtypes may exist for phosphodiesterase type III inhibitors. In children, milrinone has been found to decrease SVR by 37% and PVR by 27% ( Chang et al., 1995 ). Compared with catecholamines, milrinone has not been found to increase myocardial oxygen consumption ( Chang et al., 1995 ). Milrinone also appears to have inotropic and lusitropic properties.

Nitric oxide (NO) is an endothelium-derived vasodilator of smooth muscle. The enzyme, nitric oxide synthase, converts L-arginine into NO and citrulline. NO diffuses across the endothelial cell to the adjacent smooth muscle, where it activates guanylate cyclase. The enzyme guanylate cyclase increases the production of cGMP. cGMP causes smooth muscle relaxation by preventing the release of calcium from the SR, thereby inhibiting muscle contraction. When NO diffuses into the intravascular space, it immediately binds with hemoglobin to form nitrosylhemoglobin, which is oxidized to methemoglobin. Methemoglobin is subsequently reduced to nitrates and nitrites and is excreted in the urine. The clinical importance of NO lies in the fact that it can be administered as an inhaled gas and delivered directly to the pulmonary vascular bed in contact with ventilating alveoli. The close proximity of the alveolus and the pulmonary vascular smooth muscle allows a direct effect of NO on pulmonary vascular smooth muscle. Because of the rapid binding and inactivation of NO by hemoglobin, minimal or no systemic effects occur. NO has been shown to bind 280 times faster to hemoglobin than carbon monoxide, which may explain why the systemic circulation is protected from its vasodilating properties. Clinically, reduction in PVR has been demonstrated in adult patients with mitral valve stenosis, in neonates with persistent pulmonary hypertension, and in children with reactive pulmonary hypertension after congenital heart surgery ( Jobes and Nicolson, 1988 ; Girard et al., 1991 ; Horkay et al., 1992 ; Kern et al., 1992b ; Roberts et al., 1993b ; Wessel et al., 1997 ). Experience with NO in the operating room has shown it to be beneficial, although the degree of reduction in PVR is usually not as dramatic as initially hoped. NO is in general part of a group of therapies aimed at reducing PVR ( Bender et al., 1997 ; Lillehei et al., 1999 ).

Sildenafil causes smooth muscle relaxation through the release of NO. Intravenous sildenafil has been shown to augment the pulmonary vasodilator effects of NO in infants early after cardiac surgery. However, sildenafil produced systemic hypotension and impaired oxygenation ( Stocker et al., 2003 ). Oral sildenafil is undergoing safety and efficacy trials for primary pulmonary hypertension and chronic pulmonary hypertension. It has been shown to be as efficacious as NO in patients with a reactive component to chronic pulmonary hypertension ( Michelakis et al., 2002 ). Sildenafil along with dyperidimole (which works through a cGMP mechanism) has been used to wean patients off NO.

Isoproterenol, a β1- and β2-adrenergic agonist, has mild PA vasodilating properties in the normal pulmonary circulation ( Jobes et al., 1992 ). It causes mild PVR reduction in adult patients after cardiac transplant surgery, but there are minimal data supporting the efficacy of isoproterenol in treating PA hypertension in infants and young children after weaning from CPB. The minimal responsiveness of isoproterenol in treating PA hypertension is further complicated by the undesirable side effects of isoproterenol. Isoproterenol causes tachycardia, systemic vasodilation, and an increase in myocardial oxygen consumption ( Harker, 1986 ). These effects may reduce coronary perfusion and result in myocardial ischemia. This may be especially problematic for the RV, which has to eject against a high PVR, with decreased coronary perfusion pressure and filling time.

PGE1 and PGI2 have a pulmonary vasodilating effect, but neither drug has effects that are limited to the pulmonary circulation ( Woodman and Harker, 1990 ). Although PGE1 has been used to treat pulmonary hypertensive crisis with varying degrees of success in newborns with persistent pulmonary hypertension, its beneficial effect following CPB is inconsistent. In addition, systemic vasodilation remains a common side effect, often requiring an inotropic infusion to maintain blood pressure. PGI2 infusions have been used with some success in Europe ( Saltzman et al., 1986 ). A report by Bush and others (1987) describes the successful use of PGI2 in preventing pulmonary hypertensive crisis in five patients with congenital heart defects. Prostacyclin is currently investigational in the United States. Using a piglet model, Wauthy and others (2003) found NO to be a greater pulmonary vasodilator than prostacyclin.

Similar to NO are ultrashort-acting intravenous vasodilators. Ultrashort-acting intravenous vasodilators are nonspecific potent vasodilators with a half-life measured in seconds. Infusion of these drugs into the right side of the circulation produces a potent short-lived relaxation of PA smooth muscle. Once the drug reaches the systemic circulation, it is no longer functional. Adenosine and ATP-like compounds have these properties and may have clinical applicability in the treatment of PA hypertension ( Horrow et al., 1990 ). Ng and others (2004) showed the vasodilatory properties of 50 mcg/kg per minute adenosine infusion together with 20 ppm of inhaled NO in newborns with persistent pulmonary hypertension.

Anesthesia Strategies

Reactive pulmonary vascular responses can be attenuated by increasing the depth of anesthesia (Hickey and Hansen, 1984, 1985a [183] [184]; Anand et al., 1990 ; Horrow et al., 1990 ). Opioid-based anesthetic regimens prevent sympathetic-mediated increases in PVR (Hickey and Hansen, 1984, 1985a [183] [184]). A continuous infusion of fentanyl, sufentanil, or combination of fentanyl/sufentanil plus midazolam may be particularly useful in patients who are prone to develop labile PA hypertension.

A study in newborn infants demonstrated that reactive pulmonary responses to endotracheal suction in the postoperative period can be minimized by the prior administration of fentanyl ( Hickey et al., 1985) . Because fentanyl does not block the effects of hypoxic pulmonary vasoconstriction, its effect is most likely due to the attenuated release of sympathomimetic mediators, which produce a direct vasoconstrictive effect on pulmonary arterial smooth muscle through cAMP and calcium release ( Greeley et al., 1988 ; Anand et al., 1990 ).


Primary right ventricular dysfunction is a common finding in neonates, infants, and children undergoing cardiac surgery. Hypertrophied RV in tetralogy of Fallot, surgically induced right ventricular dysfunction due to closure of a VSD through a right ventriculotomy, and the placement of a transannular patch across the right ventricular outflow tract (RVOT), causing acute pulmonary regurgitation and right ventricular volume overload, may present with right ventricular dysfunction after weaning from CPB in CHD patients ( Friedman, 1972 ; Perloff, 1982 ; Berner et al., 1983 ; Hines and Barash, 1987 ).

The treatment of right ventricular dysfunction consists of increasing coronary perfusion pressure, preload augmentation (while avoiding marked increases in right ventricular end-diastolic pressure) (Friedman, 1972 ), and inotropic support with dopamine, epinephrine, and milrinone ( Berner et al., 1983 ; Hines and Barash, 1987 ; McGovern et al., 2000 ; Hoffman et al., 2003 ). Mechanical ventilation should be adjusted to assist right ventricular function and minimize elevation in PVR.

In contrast to the LV, the low-pressure RV receives coronary blood flow during ventricular systole ( Berne and Levy, 1981 ). In patients with right ventricular dysfunction, maintaining a normal to slightly elevated systolic pressure enhances coronary perfusion to the RV and augments contractility. Infusion of drugs such as dopamine and epinephrine, as well as intermittent doses of calcium and Neo-Synephrine (phenylephrine HCl), may be helpful in augmenting right ventricular perfusion pressure. If an increase in inotropic support persists, and frequent supplemental doses of calcium or Neo-Synephrine are required after weaning from CPB, a critical evaluation for other structural and/or functional abnormalities should be aggressively pursued. McGovern (2000) reported the effects of dopamine, dobutamine, and epinephrine in a sophisticated piglet model of right ventricular dysfunction. In this model, epinephrine was the only inotropic agent to increase right ventricular contractility and reduce right ventricular afterload, suggesting that epinephrine may be the preferred inotropic agent in patients with right ventricular dysfunction ( Table 17-12 ).

TABLE 17-12   -- Effects of exogenous catecholamines on right ventricular injury








No Δ

No Δ/no Δ


No Δ


No Δ



(From McGovern JJ, Cheifertz IM, et al.: Right ventricular injury in young swine: Effects of catecholamines on right ventricular function and pulmonary vascular mechanics. Pediatr Res 48:763–769, 2000.)

↑, Increase; ↓, decrease; No Δ, no change; PVR, pulmonary vascular resistance; TVE; transpulmonary vascular efficiency defined as measure of the ease of blood flow through the lungs; PRSW, load-independent measure of ventricular contraction; PAP, pulmonary artery pressure; Rin, intrinsic resistance. (The key to this table is that only epinephrine improved the efficiency of moving blood through the lungs.)





Preload should be maintained at a normal to slightly elevated level. Because right ventricular contractility is reduced, it is important to increase preload to optimize stroke volume as seen in Starling curve (Friedman, 1972 ). Excessive volume loading, however, is not well tolerated due to ventricular noncompliance. Excessive volume loading may result in elevated end-diastolic pressure, significant tricuspid regurgitation, and impaired forward flow. In general, a central venous pressure above 10 to 14 mm Hg is poorly tolerated in neonates and infants with right ventricular dysfunction unless the RV has marked hypertrophy and poor compliance ( Rudolph, 1985 ).

If right ventricular dysfunction persists or worsens due to elevations in PA pressure or a low cardiac output state persists despite aggressive ventilatory and inotropic support, the surgical creation of a right-to-left shunt at the atrial level may significantly improve oxygen delivery. Typical patients who would benefit from this strategy include those undergoing neonatal repairs for tetralogy of Fallot and truncus arteriosus. In these patients, allowing an atrial communication to remain open with blood shunting in a right-to-left direction preserves cardiac output and oxygen delivery to the systemic circulation. Although these patients may remain cyanotic, their effective cardiac output is enhanced, systemic perfusion pressure is improved, and coronary perfusion to the RV is maintained. Over time, right ventricular pressure decreases, right-to-left shunting decreases, and systemic oxygen saturation improves. This same strategy of leaving an atrial communication to improve cardiac output has been extended to children with single-ventricle physiology through the fenestrated Fontan.

An additional strategy in neonates, infants, and children with significant postoperative right ventricular dysfunction is to leave the sternum open but covered with a Silastic membrane ( Pearl et al., 1991 ). If mild ventricular distention has occurred and lung compliance is poor, eliminating the impedance imposed by the chest wall allows the lung to inflate at a lower mean airway pressure, decreasing right ventricular afterload. In addition, by reducing intrathoracic pressure, the right ventricular end-diastolic volume can increase at a lower end-diastolic pressure.

If right ventricular dysfunction persists despite these maneuvers, consideration should be given to extracorporeal life support (extracorporeal membrane oxygenation [ECMO]). When ECMO is used for circulatory support, venoarterial cannulation is chosen. Venovenous bypass provides gas exchange but does not provide cardiac output support. Venous and arterial access may be achieved through a large central artery and vein (usually the carotid and internal jugular) or by direct chest cannulation, a more common approach used in the operating room. Recovery from severe ventricular dysfunction is predicated on the concept that the myocardium sustained a transient injury and is capable of recovery with time, that is, stunned myocardium (Dietrich et al., 1989, 1990, 1991 [103] [104] [105]; Darling et al., 2001 ). ECMO is used to decrease ventricular wall tension, increase coronary perfusion pressure, and maintain systemic perfusion with oxygenated blood; the role of ECMO in patients with myocardial injury or pulmonary hypertension is to provide adequate systemic oxygen transport and systemic perfusion, while allowing the ventricle to “rest.”

In patients without lung disease, reports of successfully using ECMO without the oxygenator has also been suggested as a way to use the ECMO system as a ventricular assist device and reduce the need for heparin in the immediate postbypass system ( Darling et al., 2001 ). Theoretically, ECMO provides adequate systemic and coronary oxygen transport while reducing myocardial oxygen demand. This favorable shift in the oxygen supply/demand ratio may allow the heart to recover from reversible myocardial failure. In a review from the ECMO registry of all patients placed on ECMO for circulatory support, patients with tetralogy of Fallot, tetralogy-like physiology, or PA hypertension had the greatest likelihood of weaning from ECMO ( Dietrich et al., 1993 ). Sixty-one percent of patients in this category were successfully weaned from mechanical support. This suggests that myocardial injury due to severe hypertrophy or PA hypertension may result in an ischemic ventricle that with extended rest, coronary reperfusion, and reduced right ventricular afterload will significantly recover function ( Dietrich et al., 1993 ). For those patients placed on ECMO for left ventricular dysfunction, successful outcome is less common. Box 17-3 outlines a management strategy for patients with PA hypertension with right ventricular dysfunction. With use of technology derived from the operating room, ECMO circuits that used heparin-coated oxygenators and tubing have allowed the initiation and maintenance of ECMO without the administration of systemic heparin for periods of 8 to 24 hours. This is particularly beneficial in patients with significant postoperative coagulopathy or in the presence of central nervous system hemorrhage after placing the patient on ECMO. Nishinaka and others (2002)demonstrated that these heparin-coated circuits allowed for 34 days of ECMO without the need for systemic anticoagulants.

BOX 17-3 

Management Strategy for Pulmonary Artery Hypertension With Right Ventricular Dysfunction

Diagnosis: Decreased oxygen delivery caused by pulmonary hypertension and decreased right ventricular cardiac output

First: Rule out residual anatomic defects or anaphylaxis.

Treatment: Decrease right ventricular afterload and improve right ventricular function.

Ventilatory Strategy



Increase alveolar and arterial oxygen.



Perform alkalinization (pH >7.5).



Decrease Paco2 (35 to 40 mm Hg).



Decrease mean airway pressure.

Pharmacologic Therapy

Optimize Coronary Perfusion.






Phenylephrine or ?vasopressin







Reduce Pulmonary Vascular Resistance/Right Ventricular Afterload.



Adequate anesthesia



Milrinone, amrinone



Nitric oxide



Sildenafil, prostacyclin, or adenosine possibly helpful

Surgical Interventions



Create atrial septal defect



Open chest

Additional Therapy



High-frequency jet or high-frequency oscillatory ventilation



Extracorporeal membrane oxygenation/right ventricular assist device


Pharmacologic Support

The contractile state of the LV may be impaired after pediatric cardiac surgery. When left ventricular ischemia is present, the causes include surgery-induced ischemia during the repair, the preoperative condition of the myocardium (myocardial hypertrophy, elevated end-diastolic pressures), and the effects of cardiopulmonary bypass with deep hypothermia and/or circulatory arrest on myocardial compliance ( Mullins, 1989 ; Hellenbrand et al., 1990 ; Dietrich et al., 1992 ). Factors further complicating left ventricular function after cardiac surgery are reduced myocardial reserve, limited recruitable stroke work, and reduced compliance characteristic of the immature neonatal heart.

Left ventricular dysfunction can be treated by optimizing preload, increasing heart rate, increasing coronary perfusion pressure, correcting ionized calcium levels, and adding inotropic support. In the neonate, a greater dependence on heart rate, a reduction in myocardial compliance, and a diminished response to calcium and catecholamines must be considered. Inotropic support usually begins with dopamine (5 to 15 mcg/kg/minute) ( Box 17-4 ).

BOX 17-4 

Management Strategy for Left Ventricular Dysfunction

Diagnosis: Decreased oxygen delivery due to left ventricular dysfunction

First: Rule out residual anatomic defect.

Treatment: Optimize cardiac output (preload, contractility, and heart rate) and reduce afterload.

Optimize preload.



Measure right and left atrial pressures.



Maintain left atrial pressure at 8 to 12 mm Hg.



Shorten inspiratory times to augment left ventricular filling.

Augment cardiac output.



Optimize heart rate (A, AV sequential, or V pacing).



Inotropic support (dopamine, epinephrine, milrinone, dobutamine)



Calcium supplementation

Reduce afterload.



Milrinone, amrinone







Reevaluate anatomic problems.



Transesophageal echocardiography



Intraoperative catheterization

Provide left ventricular mechanical support.



Extracorporeal membrane oxygenation/left ventricular assist device


Dopamine has the unique property of binding to dopaminergic receptors in the renal and mesenteric beds and improving perfusion to the gut and the kidneys. Dopamine augments cardiac contractility through two mechanisms: a direct stimulation of β1-receptors in the heart and, more important, inducing norepinephrine release from sympathetic nerve endings ( Lock et al., 1989 ). Several studies suggest that the effect of dopamine in children is age dependent. In young children after cardiac surgery, dopamine increases cardiac output, and this effect correlates more with an elevation in heart rate than augmentation of stroke volume ( Malviya et al., 1989 ; Hickey et al., 1992 ). In young adult patients, dopamine clearly increases stroke volume ( Rothman et al., 1990 ). Nonetheless, infants and neonates respond favorably to dopamine infusion with an increase in systemic blood pressure and improved peripheral and renal perfusion. At higher doses, dopamine is converted to norepinephrine and acts as a combined β/α agonist.


Calcium supplementation is also important in augmenting cardiac contractility. Although calcium has fallen into some disfavor due to concerns over reperfusion injury, calcium supplementation remains an important therapy after pediatric cardiac surgery. In particular, the underdeveloped SR in neonatal myocardium makes the neonatal heart more dependent on extracellular calcium concentrations than the adult myocardium ( Okada et al., 1988 ). Because intracellular calcium concentration plays a central role in myocardial contractility, normal or even an elevated plasma level of ionized calcium may be necessary to augment stroke volume ( Nakanishi et al., 1987 ). In addition, fluctuations in ionized calcium levels occur commonly after weaning from CPB. This effect is most often due to the relatively large transfusions of citrate and albumin-rich blood products such as whole blood, fresh frozen plasma (FFP), platelets, and cryoprecipitate necessary to promote postoperative hemostasis ( Rebeyka et al., 1990 ). Routine monitoring of ionized calcium levels and regular calcium supplementation is helpful after weaning from CPB. This is especially true in patients with diminished left ventricular function. In patients with a slow sinus or junctional rate, calcium must be administered cautiously, as marked slowing of atrioventricular conduction may occur.


Epinephrine is a potent α-, β1-, and β2-adrenergic agonist. In lower dosages of 0.03 to 0.1 mcg/kg per minute, β-mediated responses predominate. Dosages of 0.1 to 0.2 mcg/kg per minute have a mixed α/β effect; above 0.2 mcg/kg per minute, α-mediated responses predominate. Epinephrine is useful in patients with significant left ventricular dysfunction after cardiac repair who remain hypotensive or hypoperfused based on poor systemic perfusion or rising serum lactates ( Bohn et al., 1980 ). It is effective in patients who do not respond to equivalent doses of dopamine or dobutamine, and it is particularly useful in patients with mild to moderate degrees of hypotension and echocardiographic or electrocardiographic evidence of ischemia.


Milrinone, enoximone, and amrinone are nonglycoside, noncatecholamine inotropic drugs. Their mechanism of action is mediated through the inhibition of phosphodiesterase type III ( Henriksson et al., 1979 ; Colon-Otero et al., 1987 ; Vieira et al., 1991 ). Clinical studies addressing the use of phosphodiesterase III inhibitors in pediatric patients are limited. However, reports in neonates and young infants have shown a considerable benefit from the phosphodiesterase inhibitors, especially in patients whose myocardium is afterload sensitive, such as the postoperative arterial switch patient and the patient unresponsive to catecholamines ( Vieira et al., 1991 ; Chang, 1995) .

The optimal therapeutic plasma concentration of milrinone is 100 to 300 ng/mL ( Levy et al., 2002 ). The Prophylactic Intravenous Use of Milrinone After Cardiac Operation in Pediatrics (PRIMACORP) study reported that a 50 mcg/kg intravenous loading dose followed by high-dose milrinone 0.75 to 1.0 mcg/kg per minute was safe and significantly decreased the risk of low cardiac output syndrome in this surgical population group ( Hoffman et al., 2003 ). Milrinone has become an important agent to optimize left ventricular cardiac output in pediatric cardiac patients weaning from CPB. Bypass tends to increase SVR, and the combined inotropic, lusitropic, and afterload reduction of milrinone has been efficacious in treating left ventricular dysfunction after CPB. The loading dose is often administered in the operating room while the patient is on CPB.

Pharmacokinetic studies suggest that the loading dose for amrinone in children is twice the recommended adult dose (2 to 4.5 mg/kg) ( Lawless et al., 1988 ). In clinical practice, however, amrinone in loading doses of 2 mg/kg and infusion rates of 10 to 15 mcg/kg per minute are very useful in the management of low cardiac output states in neonates, infants, and children. A higher dose (3 to 4 mg/kg) has been associated with profound systemic vasodilation in the postoperative cardiac patient. Because the onset of peripheral vasodilation precedes the increase in inotropy during the loading phase, a significant reduction in afterload and hypotension may occur if the dose is administered too quickly. This effect, coupled with its prolonged elimination half-life (approximately 3 to 15 hours), necessitates extreme caution when administering a loading dose of amrinone to children with borderline low blood pressure. In the postoperative period, loading doses of amrinone and milrinone are administered over a 20- to 30-minute period. Milrinone produces less systemic hypotension and is generally better tolerated than amrinone.


Dobutamine is an effective, albeit weaker, inotropic agent in children. It has much less peripheral α-adrenergic effect than dopamine, and its β1-adrenergic effect predominates, yielding a mild peripheral vasodilation. Although reported to have less chronotropic effects than dopamine, in neonates significant tachyarrhythmias may occur. This may relate to structural similarities between dobutamine and isoproterenol ( Bohn et al., 1980 ). In children after cardiac surgery, dobutamine increases cardiac output primarily through increases in heart rate. The efficacy of dobutamine seems to be reduced in immature animals ( Rothman et al., 1990 ). This is consistent with the previous discussion of reduced β-receptors and a higher level of circulating catecholamines in newborns ( Bohn et al., 1980 ; Berner et al., 1983 ).


Synthetic 8-L-arginine vasopressin (AVP), acting at V1 receptors, is administered as a continuous infusion for refractory hypotension following CPB in adults and has been used in children, particularly those in sepsis. In one study, the dosage of AVP was adjusted for patient size and ranged from 0.0003 to 0.002 U/kg per minute. During the first hour of treatment with AVP, systolic blood pressure rose from 65 ± 14 to 87 ± 17 mm Hg (P < .0001; n = 11). Infants with refractory low blood pressure and adequate cardiac function may benefit from AVP administration after cardiac surgery ( Rosenzweig et al., 1999 ). No prospective data on the use of vasopressin are available to dictate its use in congenital heart surgery, but it is being used as an alternative agent when catecholamine-induced dysrhythmias, such as junctional ectopic tachycardia and hypotension, are present. AVP is also being recommended for the treatment of cardiac arrest. Preliminary data from out-of-hospital prospective resuscitation trials suggest that AVP is superior to epinephrine. Vasopressin may prove superior to epinephrine as a pressor agent during prolonged cardiopulmonary resuscitation in children as well ( Mann et al., 2002 ).

Assist Devices for Left Ventricle

Compared with traditional ECMO, ventricular assist devices (VADs) may be a preferred means of circulatory support for children who are unable to be weaned from CPB due to primary left ventricular dysfunction and for patients with acute cardiomyopathy or myocarditis. Reinhartz and others (2003) showed a 72% survival rate for children of less than 1.3 kg/m2 (mean, 1.09 kg/m2; range, 0.73 to 1.29 kg/m2) body surface area (BSA) placed on Thoratec LVAD (Thoratec Corp, Pleasanton, CA) for 0 to 120 days (mean, 42 days) because of cardiomyopathy or myocarditis. In this study, only one of seven patients with CHD survived.

VADs allow direct control of left atrial filling pressure and therefore left ventricular end-diastolic pressure can be minimized, allowing for a greater likelihood of left ventricular recovery. The most widely used device is the centrifugal pump, in which blood is moved by its entrainment against spinning blades and cones; an example is the Biomedicus pump. Centrifugal pumps require direct heart cannulation. A left ventricular assist device (LVAD) requires a left atrial venous cannula and an aortic cannula. Wire-reinforced cannulas are preferred, because kinking can result in marked reduction in cardiac output and/or ventricular distention. The cannulas are secured in place and connected via polypropylene tubing to the ports on the pump head. Long-term outcomes seem to be related to the underlying cardiac condition. The need for aortic cross-clamping before the patient is put on an LVAD has a negative impact on outcome ( Schindler et al., 2003 ).

When the LVAD is activated, the flow rates are increased until the patient is completely weaned from CPB and the LVAD is providing full circulatory support. Left and right atrial monitoring catheters are strongly recommended during LVAD use. Because there is no venous reservoir in the centrifugal pump system, pressure monitoring is essential to ensure adequate intravascular volume to sustain pump flow rates and prevent pumping air. Low left atrial pressures may result from hypovolemia or right ventricular failure. A right atrial pressure monitor helps differentiate volume problems from right ventricular dysfunction. If right atrial pressures are high and left ventricular filling pressures are low, significant right ventricular failure is present and conversion to a biventricular assist device system or ECMO may be indicated. Box 17-4 outlines management strategies for left ventricular dysfunction.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


The goals for congenital repair are the (1) physiologic separation of the pulmonary and systemic circulation, (2) relief of outflow obstruction, (3) preservation or restoration of ventricular mass and function, and (4) maintenance of the patient's quality of life and normalization of life expectancy. The available surgical procedures to accomplish these objectives are diverse and complex ( Table 17-13 ). Compared with cardiac operations in adult patients, congenital heart repair involves more intracardiac surgery with a greater preponderance performed through the RA, RVOT, and left ventricular outflow tract (LVOT).

TABLE 17-13   -- Common congenital heart defects and surgical approaches

Anatomic Defect


Complete Repair

Tetralogy of Fallot


VSD closure and RVOT patch

 With PA atresia



 With anomalous right coronary artery

Rastelli procedure

Patch above and below coronary artery

Hypoplastic left heart syndrome

Norwood followed by Fontan procedure


Transposition of the great arteries


Arterial switch

 Unfavorable coronary anatomy

Atrial switch (Senning)


Tricuspid atresia

Shunt followed by Fontan procedure


Pulmonary atresia with VSD

Shunt followed by Fontan procedure


With intact septum

Shunt followed by Fontan procedure


Critical aortic or pulmonary stenosis


Valvotomy or catheterization laboratory balloon valvuloplasty

Interrupted aortic arch


End-to-end anastomosis

Total anomalous pulmonary venous return


Anastomosis pulmonary veins to left atrium and ASD closure

Single ventricle/normal PA

PA band followed by Fontan procedure


With small PA

Shunt followed by Fontan procedure


Truncus arteriosus


RV-PA conduit and VSD closure

Atrioventricular canal


Repair valve clefts/patch closure of ASD/attach valves to patch

ASD, atrial septal defect; PA, pulmonary artery; RV, right ventricle; RVOT, right ventricular outflow tract; VSD, ventricular septal defect.




In general, operations performed for congenital heart defects can be divided into palliative and corrective procedures ( Arciniegas, 1985 ). The type and timing of repair depend on the age of the patient, the specific anatomic defect, and the experience of the surgeon (see Table 17-13 ). Palliation in infancy is usually performed where there are missing anatomic parts such as in pulmonary atresia (absent RV and PA), tricuspid atresia (absent RV and tricuspid valve), HLHS (atretic or severely stenosed mitral and aortic valve and hypoplastic LV), univentricular heart (absent RV or LV), or mitral atresia (absent LV). These palliative procedures can be further subdivided into those that increase PBF, those that decrease PBF, and those that increase mixing. Palliative procedures that increase PBF include shunts (modified BT), outflow patch, or enlargement of the VSD. Those that decrease PBF include PA banding and ligation of a PDA. Those that improve intracardiac mixing include atrial septostomy or septectomy (balloon, blade, or Blalock-Hanlon).

With the improvements in surgical technique coupled with the advancements in anesthetic and technological support, repair in early infancy is more effective and has been the procedure of choice for many congenital cardiac defects ( Castaneda et al., 1989 ). Neonatal repair is offered for a number of congenital heart defects (see Table 17-13 ), including total anomalous pulmonary venous return, interrupted aortic arch, and coarctation of the aorta, AS, PS, truncus arteriosus, and TGA. Atrioventricular septal defects (AVSDs) and tetralogy of Fallot are usually repaired in early infancy. In cases where pulmonary outflow obstruction is severe, tetralogy of Fallot requires neonatal repair. Repair in the neonatal period almost always requires a transannular patch, resulting in postoperative pulmonary insufficiency and a volume load on the previously pressure-loaded RV. Ventricular septal defects (VSDs) are not routinely closed in the newborn period, because many of these defects close on their own. Indications for early closure are usually related to unrelenting congestive heart failure despite medical management and impaired somatic growth. AVSDs are generally not repaired in infancy; patch closure and chordal reattachment are more readily performed in a larger heart. Occasionally, palliation is performed for some of these lesions in the newborn period with total correction later in life. Palliation is usually considered when closing the defect results in suboptimal chamber size, such as a patient with an unbalanced AVSD or with other severe congenital anomalies or when the patient is a poor candidate for CPB (birth asphyxia, intraventricular hemorrhage, etc.).

For well over a decade, the preferred approach to pediatric cardiovascular surgery has been to repair defects in infancy rather than to palliate ( Castaneda et al., 1989 ). This trend is due to the increased morbidity and mortality associated with long-term medical management and the sequelae of multiple palliative operations. Early corrective surgery, if performed well, decreases the incidence of the chronic complications of CHD such as the problems associated with ventricular pressure or volume overload, cyanosis, and pulmonary vascular obstructive disease ( Castaneda et al., 1989 ; Mahle et al., 2002 ). Complete repair, even in very low birth weight newborns weighing less than 1.5 kg, has been advocated by some ( Reddy et al., 2000) . In a series of 20 symptomatic low birth weight infants, early cardiac surgery with CPB resulted in an early infant mortality rate of 10%. At a median follow-up of 40 months, late mortality occurred in 5% and catheterization laboratory intervention was necessary in 25%. There was no evidence of intracranial hemorrhage after CPB. Although a higher surgical mortality rate was reported for low birth weight infants, prolonged medical management of these same patients generally purport a high morbidity and mortality ( Reddy et al., 2000) .

There may be a selective advantage of enhanced organ system protection during infant repair due to poorly understood factors promoting resistance to injury and enhanced recovery (i.e., enhanced organ plasticity and reserve) ( Reddy et al., 1999 ). With the continued improvement in surgical techniques, bypass technology, interventional cardiology procedures, and the early treatment of CHD, specific organ systems such as the brain, heart, and lungs sustain less damage and may be spared the detrimental effects of chronic CHD and the effects of CPB.

Another trend observed in surgery for CHD is the continued evolution of new techniques to decrease long-term morbidity and enhance survival. For example, in the 1980s, the long-term problems with right ventricular dysfunction and failure associated with the Mustard and Senning procedures for repair of TGA encouraged many surgical groups to perform the arterial switch in the neonatal period as the procedure of choice ( Castaneda et al., 1988 ). With the latter procedure, it was believed that a normal anatomic correction had better long-term results. A second example of the continuing evolution of technique is surgery for tetralogy of Fallot. Long-standing pulmonary insufficiency after right ventricular outflow repair for tetralogy of Fallot is associated with right ventricular dysfunction and failure. Preservation of the pulmonary valve at initial repair using a combined transatrial and transpulmonary approach during correction or the early insertion of a pulmonary homograft in the setting of pulmonary insufficiency is a technique that is used in the attempt to avoid the long-term problems of right ventricular dysfunction and failure ( Pacifico et al., 1987 ).

Surgery for HLHS, once considered a fatal disease, has achieved significant long-term survival after a series of creative staging procedures initially described in the mid 1980s, and it achieved worldwide application by the early to mid 1990s (Norwood et al., 1991, 1992 [340] [341]). Newer techniques evolved to minimize the morbidity of the Norwood procedure ( Mahle et al., 2000 ), including the Sano modification ( Sano et al., 2003 ). In this technique, first described by Imoto and others (2001) , a conduit is sutured from the single RV to the PA instead of creating a shunt from the innominate artery to the PA. This modification preserves systemic diastolic blood pressure, and therefore coronary blood flow, in the postoperative period. Myocardial function improves, and both hemodynamics and end-organ perfusion appear better preserved in the early postoperative period ( Mair et al., 2003 ; Pearl, 2003 ). Early pre-Glenn catheterization data have demonstrated a statistically significant finding of higher dP/dT (a measure of contractility) in the systemic ventricle after the Sano modification compared with the standard shunt approach ( Mair et al., 2003 ). Studies of survivors of the Sano modification at 1 year after repair have found excellent outcomes that are comparable to those of the traditional modified BT shunt ( Maher et al., 2003 ; Mahle et al., 2003 ; Mair et al., 2003 ; Pizarro et al., 2003, 2004 [361] [362];Sano et al., 2003 ). Also, the acute postoperative course may require less intensive management and patients may have improved mixed venous saturations. Long-term randomized outcome trials have been recommended to determine the risk of developing arrhythmias, the risk of thrombosis and infective endocarditis, and the long-term function of the single RV after a ventriculotomy incision in the systemic RV ( Pearl, 2003 ).

Another interesting therapeutic option for the neonatal HLHS has been the use of a combined interventional cardiology and cardiac surgical approach. In the catheterization laboratory, an atrial septostomy is performed and a device is placed in the PA to restrict PBF. The patient is then taken to the operating room to have an innominate-to-PA shunt placed. Alternatively, a stent is placed in the ductus arteriosus in the catheterization laboratory. The combined interventional and operative approach to hypoplastic left heart surgery allows neonatal palliation to occur without the need of DHCA or bypass in many cases ( Muller et al., 2003 ). The overall benefit is a reduced need for intensive care management and a reduced exposure to CPB. Innovations such as those outlined earlier continue to optimize the management of neonates with critical cardiac disease and reduce the long-term morbidity associated with palliative procedures and complex postoperative critical care management.

Other trends in surgical management include the broader application of surgical procedures initially designed for a specific defect. For example, the Fontan operation, which was originally devised for tricuspid atresia, has been used to repair complex univentricular hearts and HLHS ( Gildein et al., 1990 ; Kopf et al., 1992 ). Attendant with this wider application of the Fontan operation for more complex defects once considered inoperable has been the rise in morbidity and mortality. Even this trend has been reversed in patients with higher-risk defects ( Mayer et al., 1992 ).

In the early 1990s, the creation of a fenestration (or “hole”) between the RA and LA at the time of the Fontan operation allowed right-to-left shunting, maintenance of cardiac output, and avoidance of ventricular failure in the early postoperative period. At a later time, once the patient has convalesced from the acute operative events, the fenestration was closed in the catheterization laboratory with an ASD closure device or, in some centers, at the bedside with a snare placed at the time of the operation. With the advent of a three-stage approach to the repair of single-ventricle patients, including a middle-stage procedure such as a bidirectional Glenn or hemi-Fontan, which allows the patient to maintain cardiac output by leaving the inferior vena cava (IVC) blood flow to return to the single ventricle and support cardiac output, the need for fenestration at the time of the Fontan procedure has been less. With the need for fenestration reduced, extracardiac conduits rather than intracardiac baffles with a fenestration ( Kuroczynski et al., 2003 ) are used to complete the Fontan procedure. The long-term benefit of an extracardiac conduit reduces the exposure of the RA to high pressure and suture lines, of which both have been associated with an increased incidence of atrial arrhythmias ( Fishberger et al., 1997 ; Kumar et al., 2003 ). The mortality and morbidity of atrial arrhythmias are significantly higher in the single-ventricle patient ( Kumar et al., 2003 ).

Ingenuity, innovation, and continued refinement have permitted continued improvements in survival and quality of life for all patients with CHD. As the incisions in the myocardium become smaller, sutures are more precisely placed, and surgical techniques evolve, the complications of ventricular dysfunction, arrhythmias, and residual obstruction have declined and both life expectancy and life quality continue to improve.

One final difference unique to congenital heart surgery that has a major impact on anesthetic management relates to the type of cardiopulmonary support. Because of the complexity of repair in small patients, pediatric cardiac surgery involves operating at extreme biological conditions of temperature, hemodilution, and perfusion (Barratt-Boyes et al., 1971, 1980 [25] [26]). Despite a long history of use, these techniques have pronounced physiologic effects on major organ system function and are far from understood with regard to long-term outcome.


The postoperative management of the patient with a complete anatomic and physiologic correction is simplified because blood flow patterns after surgical correction are normal and, if an anatomic abnormality exists, the physiologic effect is more predictable and the location for the problem more obvious. Examples of complete anatomic repair include closure of an ASD or a VSD, repair of tetralogy of Fallot, and the arterial switch operation for TGA (see Table 17-13 ).

Atrial Septal Defects

Functional communications between the two atria are common congenital cardiac defects. ASDs are true defects, unlike a probe-patent PFO (which may be found in >20% of adults). ASDs comprise about 10% of all congenital anomalies in the pediatric population and are more common in females than in males (by 2:1). There are two main types: ostium secundum defects and ostium primum defects ( Fig. 17-12 ). Secundum defects are the more common lesions and can be subdivided further into fossa ovalis defects, sinus venosus defects, and low ASDs. Fossa ovalis defects are located posteroinferiorly (midseptally) in the atrium, at the site of the fossa ovalis. Sinus venosus defects are located superiorly in the atrial septum at the opening of the SVC into the RA. This type is often associated with one or more pulmonary veins draining into the RA (partial anomalous pulmonary venous return). A low secundum ASD may, on occasion, have no inferior border. The defect is sitting over the entrance of the inferior vena cava into the RA (see Fig. 17-12 ). Some patients with ostium secundum defects have mitral valve prolapse (10%) or, very rarely, mitral stenosis (Lutembacher's syndrome).


FIGURE 17-12  Drawing of interior of right atrium showing interatrial septum and position of three types of interatrial septal defects.  (From Reed CC, Stafford TB, editors: Cardiopulmonary bypass, 2nd ed. Houston, 1985, Texas Medical Press, p 78.)




Small septum secundum defects with PBF less than twice the systemic blood flow (   p/   s < 1.5 to 2) are well tolerated. Life expectancy is presumed to be normal, and symptoms are usually mild in childhood. In patients with a large septum secundum (   p/   s >2), the PA pressure is normal or slightly increased, and the PVR remains normal. Systemic blood flow is usually slightly decreased because of impaired delivery of blood to the LV resulting from atrial shunting. If the ASDs are untreated, structural changes in the pulmonary vascular bed may occur over several decades and may result in pulmonary hypertension and congestive heart failure.

Because of the longstanding risks of pulmonary vascular disease, arrhythmias, and paradoxical emboli (especially in pregnant women), surgical repair is usually recommended. Surgery is usually performed in children between 2 and 5 years of age but may be done earlier in infants with ostium primum defects associated with mitral regurgitation and CHF.

ASDs are surgically repaired using CPB and patch closure if the defects are large. In addition, repair of the mitral or tricuspid valve or both may be required in patients with ostium primum defects. Primum defects usually have an associated cleft in the mitral valve, which requires suture closure. Surgical mortality is less than 1% in most centers, but the risk is increased in the presence of elevated PVR or congestive heart failure. Complications include complete heart block (<1%) and supraventricular arrhythmias. Residual mitral insufficiency may remain or even worsen in patients with ostium primum defects.

Total and Partial Anomalous Pulmonary Venous Connections

Anomalous pulmonary venous connection is a rare congenital heart defect that has a vast array of abnormalities. This condition occurs if one or more of the pulmonary veins connect to the RA. The physiologic result of the abnormal connections is abnormal return of the oxygenated pulmonary venous blood to the right side of the heart.

The anomalies of pulmonary venous connection are classified based on their anatomy and physiology ( Hansen, 1985 ). If all of the pulmonary veins return to the RA, this condition is known as total anomalous pulmonary venous connection (TAPVC). If one or more, but not all, of the pulmonary veins connect to the RA, this condition is known as partial anomalous pulmonary venous connection(PAPVC). Approximately one third of patients with TAPVC have associated cardiac defects; prominent among associated defects are the heterotaxy syndromes.

The classification of TAPVC is based on the anatomic sight of the abnormal connection. TAPVC is usually classified according to the scheme proposed by Darling and others (1994) , which is based on the anatomic site of an anomalous connection: type 1 is a supracardiac defect, type 2 is a cardiac defect, type 3 is an infracardiac defect, and type 4 is a mixed defect ( Fig 17-13 ).


FIGURE 17-13  Type of total anomalous pulmonary venous connection. (A) Supracardiac type in which the pulmonary veins connect to the superior vena cava via a left vertical vein (persistent left superior vena cava). (B) Cardiac type in which the pulmonary veins drain into the coronary sinus. (C) Infracardiac type in which the pulmonary veins drain into a common pulmonary vein that passes through the diaphragm to enter the inferior vena cava.



The supracardiac type is the most common type, occurring in 41% to 49% of cases. In this type, the confluence of the pulmonary veins drains superiorly through a left superior vena cava (SVC) and, most often, enters the left innominate vein. Less frequently, there is connection of the ascending vein directly to the right SVC.

In the cardiac type of anomalous connection, the venous connection is to the coronary sinus or directly to the posterior wall of the RA. This type of TAPVC accounts for 18% to 31% of cases. Obstruction may occur at the point of entrance to the coronary sinus. A restrictive atrial communication may also occur with this type of defect.

With the infracardiac TAPVC, the distal site of connection is usually below the diaphragm, connecting with a vessel of the portal system. This type accounts for 13% to 24% of cases. In this defect, the pulmonary veins form a confluence behind the LA, and this confluence is drained through a connection that descends inferiorly, penetrates the diaphragm, and connects most frequently to the portal venous system, although connections to the ductus venosus, hepatic veins, or SVC have been reported. This type of abnormal pulmonary venous connection is usually associated with obstruction.

The mixed type of TAPVC occurs when there are two or more sites of anomalous venous connections. This type is rare, occurring in 5% to 10% of cases. Obstruction of one or more of the connections may occur.

With TAPVC, all pulmonary venous blood eventually returns to the RA. Because of the lower resistance of the pulmonary circulation, the majority of blood entering the RA is ejected out the RV into the pulmonary circulation. This situation results in right ventricular overload and may lead to symptoms and signs of congestive failure. An atrial level right-to-left shunt is obligatory to maintain cardiac output and systemic circulation. The amount of blood shunted to the LA is limited by the size of the interatrial connection and left atrial size. If there is no restriction of interatrial flow and left atrial compliance is normal, then the size of the right-to-left shunting is primarily determined by the relative compliance of the RV and LV. Dilation of the RV does occur, and PA pressure may range from being mildly elevated to systemic levels.

The presence of TAPVC with pulmonary venous obstruction results in raised pulmonary venous pressure and PA hypertension. PA hypertension is usually severe and results in decreased PBF and significant right-to-left shunting at both the atrial and ductal level. The right-to-left shunting produces severe hypoxemia in the neonatal period. The systemic hypoxemia is often serious enough to cause hemodynamic instability, resulting in hypotension, acidosis, and cardiac arrhythmias. Usually emergent surgical therapy to relieve obstruction or placement on ECMO to provide hemodynamic stability and multiorgan system resuscitation via reperfusion is necessary before operative intervention. The use of ECMO before the repair is associated with a higher mortality rate.

All patients with TAPVC require surgery. The urgency with which this surgery is performed is dictated by the anatomy, physiology, and clinical presentation, as discussed earlier. Patients with TAPVC who have pulmonary venous obstruction may be severely ill with noncompliant lungs and low cardiac output and require urgent surgical intervention. A cardiologic intervention in which the anomalous vein was opened with a stent allowed time for a patient to improve before surgical correction ( Ramakrishnan and Kothari, 2004 ). Patients who lack pulmonary venous obstruction and who have an adequate-size ASD (to allow some flow to the LA and LV) may present with pulmonary congestion and tachypnea but without severe systemic acidosis and cyanosis and can be dealt with in a less emergent, but still urgent, fashion.

The goal of surgery is to connect the pulmonary venous return to the LA and obliterate the pathway connecting the pulmonary veins to the systemic venous system. In patients with the supracardiac and infracardiac types of anomalous drainage, there is ordinarily a confluence behind the LA to which all of the pulmonary veins converge ( Delise et al., 1976 ; Eumel and Sreeram, 2004 ). This confluence drains into the systemic venous system by means of a connection that ascends toward the innominate vein (typical form of supracardiac TAPVC) or descends (through the diaphragm) toward the portal venous system (typical form of infracardiac TAPVC). For these anatomic types, surgical correction involves connection of the pulmonary venous confluence to the posterior aspect of the LA and ligation of the ascending or descending anomalous connection. The LA is almost always small in TAPVC. Embryologically, a significant portion of the LA is composed of the pulmonary venous connection derived from outgrowth from the fetal lungs. After surgical repair of TAPVC with obstruction, monitoring of both the PA and left atrial pressures is important, because these infants have a propensity to develop postoperative PA hypertension.

Ventricular Septal Defect

VSDs comprise 20% to 30% of all congenital heart defects. The interventricular septum is composed of three anatomic components: the membranous septum, the muscular septum, and the inlet septum. Anatomically, VSDs are classified according to their location. The term crista supraventricularis designates a muscular portion of the septum that separates the tricuspid and pulmonary valves and the pulmonary and aortic valves ( Fig. 17-14 ). Supracristal VSDs (type I) lie just beneath the pulmonary valve and communicate with the RVOT above the supraventricular crest. Supracristal VSDs account for 5% to 8% of all VSDs and are associated with a significant incidence of aortic valve regurgitation. Membranous or infracristal VSDs (type II) constitute the most common type of VSD (75% to 80%) and are located in the membranous portion of the septum, immediately below the septal leaflet of the tricuspid valve. An inlet VSD (type III) is usually seen in atrioventricular defects where the defect is located in the posterior region of the septum, immediately below the inlet portion of the tricuspid valve. The incidence of this defect accounts for 4% of all VSDs. Defects located in the muscular portion of the septum (type IV) are frequently multiple and are found in 10% to 15% of VSDs.


FIGURE 17-14  Classification of ventricular septal defects.  (From Cooley DA, Norman JC: Techniques in cardiac surgery. Houston, 1975, Texas Medical Press, p 119.)




A VSD is a condition of chronic volume overload in which a large increase in diastolic volume is tolerated with only minimal change in end-diastolic pressure ( Graham, 1991 ). The left ventricular end-diastolic volume is usually twice normal, whereas left ventricular end-diastolic pressure and ejection fraction are unchanged. Right ventricular dilation also occurs with VSDs, but it is usually less severe (Graham, 1982 ). The effects of a persistent VSD on left and right ventricular function depend on the length of time the ventricle has been exposed to the chronic volume-overload state. Large VSDs left unrepaired for 2 to 3 years or longer have a greater incidence of permanent ventricular dysfunction. Patients with these disorders, along with neonates and infants with poorly compensated heart failure due to a large VSD, form a subset of patients at increased risk for postoperative ventricular dysfunction.

For the most part, the overwhelming majority of children undergoing closure of a VSD do well. Problems, however, can be expected if a residual VSD remains, if significant preoperative congestive heart failure exists in infants, or if repair is associated with other cardiac anomalies such as coarctation of the aorta, interrupted aortic arch, or mitral or aortic valve anomalies ( Pacifico et al., 1990 ).

Residual septal defects are diagnosed with intraoperative TEE. Previously undiagnosed defects are usually in the muscular septum. Residual defects around the patch are usually isolated to an area poorly visualized by the surgeon or near the conduction tissue (avoided to prevent heart block). Small residual defects less than 3 to 4 mm have a high likelihood of closing on their own, especially if the child is hemodynamically stable and on minimal support ( Rychik et al., 1991 ).

Infants with a large VSD ( Fig 17-15 ) are susceptible to PA hypertensive crisis due to preexisting high PBF ( Dietrich et al., 1992 ). After closure of the VSD, the RV must pump against increased PVR. The RV can no longer recruit the LV to assist in pumping against increased PVR. It must suddenly change from pumping against a volume load to supporting a pressure load. The change in loading conditions incurred by the cardiac surgery is an example of anticipated physiologic changes in cardiac loading. Similarly, the LV must pump a smaller-volume load against SVR without the luxury of pumping any extra volume through the VSD and into the lower-pressure right-sided circulation. The LV may be at a further disadvantage if PA pressures are high, the right ventricular dilates, and the septum shifts into the LV. This significantly alters left ventricular geometry and impairs left ventricular ejection ( Meyer et al., 1972 ) ( Fig 17-16 ). Operative management is directed at reducing PVR and improving right ventricular contractility. These maneuvers improve blood return to the LV, restore left ventricular geometry, and improve cardiac output.


FIGURE 17-15  Sequential analysis of increasing pulmonary artery pressures in an experimental model. As pulmonary artery pressures are increased, the right ventricular (RV) end-diastolic pressure (EDP) increases RV hypertension (RVH), RV coronary perfusion pressure (RV CPP) decreases, and both cardiac output (CO) and mean arterial pressure (MAP) fall. As pulmonary artery hypertension increases, RV function worsens (increasing RVEDP), RV CPP narrows, and CO and MAP plummet. A bolus dose of phenylephrine (PHE) rapidly reverses RV dysfunction by improving coronary perfusion pressure to the RV. After PHE, RVEDP decreases and a statistically significant increase in CO and MAP is observed.  (With permission from Vlahakes GJ, Turley K, Hoffman JI: The physiology of failure in acute right ventricular hypertension. Circulation 63:87-95, 1980.)





FIGURE 17-16  Echocardiogram demonstrating right ventricular (RV) dilation impinging on the left ventricle (LV). This alters left ventricular filling and geometry, secondarily reducing cardiac output. AO, aorta.



Patients undergoing repair at a late age or those with poor myocardial protection during repair are similarly subject to right and left ventricular dysfunction. Elevation in PA pressure may also be a component of their disease, but reactive PVR does not commonly occur with VSD after the early infancy period. PA pressures should decrease to one-half systemic in the operating room and generally decrease to one-third systemic within 24 hours of surgery. High PA pressure is very suggestive of a residual shunt, mitral regurgitation, or, less commonly, fixed changes associated with medial hypertrophy of the pulmonary arterioles. Echocardiographic evaluation is extremely helpful in these patients in the postoperative period.

Complete heart block is an uncommon finding today due to the precise anatomic knowledge concerning the location of the conduction tissue in different types of VSDs ( Dickinson et al., 1982 ). Transient heart block, however, can occur. Stretching of the conduction tissue at the time of surgery to better visualize the defect, ion fluxes associated with cardioplegia solutions, and low calcium-containing primes are other contributing factors. Temporary ventricular pacing or, preferably, atrioventricular sequential pacing should be used in these patients. In many institutions, atrial and ventricular wires are routinely placed after all operations involving closure of a VSD.

Associated cardiac anomalies such as left-sided obstructive lesions (interrupted aortic arch and coarctation) result in a pressure-loaded, dysfunctional LV. These patients frequently require inotropic infusion to improve left ventricular contractility when being weaned from CPB.

Tetralogy of Fallot

Although the surgical approach to primary repair of tetralogy of Fallot has not changed significantly, the timing of repair has. Evidence has accumulated that a considerable portion of the myocardial, pulmonary, and central nervous system damage attributable to tetralogy of Fallot is more specifically due to prolonged exposure to abnormal blood flow patterns, hypoxemia, and cyanosis ( Borow et al., 1981 ; Castaneda et al., 1989 ; Newburger et al., 1993 ). Surgery involving multiple palliation procedures and long-term medical management does not alleviate these problems, and a marked increase in long-term morbidity and mortality has been demonstrated. For the past 15 years, there has been a greater impetus to repair rather than palliate these children in infancy ( Castaneda et al., 1989 ). Most centers prefer to operate at 4 to 8 months of age, with a few opting for repair at earlier ages. A group from The Hospital for Sick Children in Toronto ( van Dongen et al., 2003 ) reviewed their experience from 1997 through 1999. Seventy-eight consecutive patients were enrolled, and the mean age at operation was 8 months (range, 36 days to 18.5 months). The overall mortality was 1.3%, and intensive care unit stay was 2 days. Operative intervention at age younger than 3 months was associated with an increased need for inotropic support, higher postoperative fluid requirement, and higher incidence of secondary organ dysfunction.

The repair of tetralogy of Fallot includes a right ventriculotomy, patch closure of the VSD, resection of infundibular stenosis, and patch augmentation of the RVOT. The surgeon evaluates the size of the pulmonary valve and, if it is too small, the right ventricular incision is extended across the pulmonary valve annulus and onto the PA. A transannular patch creates pulmonary insufficiency, which may contribute a volume load to an already pressure-loaded RV ( Ilbawi et al., 1987 ). The right ventriculotomy may also contribute to right ventricular dysfunction during weaning from CPB and may contribute to late right ventricular dysfunction or dysrhythmias ( Atallah-Yunes et al., 1996) . These concerns are more pronounced when surgery is performed in early infancy. An alternative surgical technique is to limit or eliminate the right ventricular incision by using a transatrial or combined transatrial/transpulmonary approach to resect muscle from the RVOT and to close the VSD ( Pozzi et al., 2000 ). This approach has shown no increase in morbidity and mortality and has potential for lower long-term cardiac rhythm problems and sudden death, which is a concern in older patients who underwent repair decades earlier ( Pozzi et al., 2000 ).

Repair in the first few months of life is usually necessary when severe RVOT obstruction, a small pulmonary valve orifice, and preoperative hypercyanotic spells are present. These patients commonly have a hypertrophied, noncompliant RV and frequently require a transannular patch across the RVOT. After repair, right ventricular systolic and diastolic dysfunction is common, as well as higher right-sided filling pressures and lower PBF. Consequently, the left ventricle receives less filling, and there is reduced stroke volume and decreased cardiac output. A reflex tachycardia occurs in an attempt to augment systemic perfusion.

Leaving the foramen ovale open or creating a small ASD at the time of surgery provides a right-to-left shunt at the atrial level. This shunt bypasses the noncompliant RV, preserving left ventricular filling and systemic perfusion. The patient is somewhat hypoxemic, but this is well tolerated in the face of improved systemic oxygen delivery. An alternate approach in patients with tetralogy of Fallot and diminutive PAs has been the placement of a fenestrated VSD patch. This approach has not been associated with increased mortality and has a low incidence of excessive right-to-left shunting ( Marshal et al., 2003) .

Persistently low cardiac output despite an atrial level shunt suggests either a residual anatomic problem or severe ventricular dysfunction. Intraoperative echocardiography and assessment of the adequacy of the right ventricular outflow patch by pressure measurements should be used to evaluate the repair for a residual VSD or inadequate relief of RVOT obstruction. If anatomic considerations persist, reoperation should be considered. If primary ventricular dysfunction exists, therapy should be directed toward right ventricular or left ventricular dysfunction as previously discussed and, rarely, the use of ECMO and VADs.

Anatomic variants in coronary artery anatomy alter the surgical approach and may affect anesthetic management in patients with tetralogy of Fallot. If the right coronary artery gives rise to the left anterior descending artery, the coronary courses directly across the RVOT. A standard transannular enlargement is not feasible because the right coronary would have to be transected. Under these circumstances, the surgeon must decide to either place a conduit over the outflow tract or resect as much infundibular tissue as possible and place a patch above and below the coronary artery ( Hurwitz et al., 1980 ; Humes et al., 1987 ). In the latter case, the narrowest part of the RVOT is beneath the right coronary artery. Excessive preload may distend the RV, stretch the right coronary artery, and cause right ventricular ischemia. Although ischemic changes are uncommon in pediatric cardiac surgery, this anatomic arrangement of the coronary arteries substantially increases the likelihood of ventricular ischemia. Careful electrocardiographic observation for ischemic changes is essential when augmenting preload. Volume supplementation should be given in small increments, and right-sided filling pressures should be maintained within a narrow range to avoid right coronary stretch.

In the preoperative or the prebypass periods, tetralogy or hypercyanotic spells may occur. These episodes of acute cyanosis are due to an increase in subpulmonary infundibular spasm, a reduction in PBF, and an increase in right-to-left shunting across the VSD. Because the direction of intracardiac shunting is dependent on downstream resistance, hypercyanotic spells are due to increased resistance to blood flow across the RVOT.

Severe forms of tetralogy of Fallot present in early infancy with PS, annular stenosis, and diffuse hypoplasia of the RVOT ( Lev and Eckner, 1964 ; Rabinovitch et al., 1981 ). This subset of tetralogy patients presents with severe cyanosis at birth that is poorly responsive to medical therapy. Classic hypercyanotic spells do not occur in this group because of fixed obstruction to the RVOT. Early operative intervention is required, and complete repair in infancy generally requires a transannular patch. More typically, tetralogy patients are electively repaired during the first 4 to 8 months of life. Early repair prevents significant cardiac hypertrophy and is best if performed before the onset of significant hypercyanotic spells.

Medical management for tetralogy spells is directed at reducing right-to-left shunting across the VSD by increasing SVR through the administration of phenylephrine and by decreasing the effective narrowing of the RVOT through volume loading, sedation, and, rarely, β-blockade to decrease contractility (this is not commonly applied in the operating room setting but has been used to stabilize patients preoperatively). Increasing SVR is a critical component of treating patients with tetralogy spells ( Lang et al., 1980 ; Berner et al., 1989 ). Therapy is directed at reducing the infundibular obstruction through increasing preload and suppressing endogenous catecholamine release through increasing anesthetic depth or β-blockade. Increasing SVR rapidly reduces the right-to-left shunting across the VSD and restores oxygen saturation and oxygen delivery to acceptable levels.

The degree of ventricular shunting is dependent on the resistance to flow imposed at the level of the RVOT and the SVR. When the outflow tract resistance exceeds SVR, right-to-left shunting occurs. By increasing SVR, blood flow is preferentially in a left-to-right direction at the VSD and the hypercyanotic spell aborted. During surgery, opening the pericardium, manipulation of the RVOT, or atrial cannulation can alter preload and/or induce spasm of the infundibulum. In general, volume loading and administration of 25 to 100 mcg of phenylephrine are effective in restoring PBF and reducing right-to-left shunting at the VSD. However, occasionally severe cyanotic spells occur that are poorly responsive to medical therapy. Anesthetic management for these uncommon, yet life-threatening, events include correcting acidosis (sodium bicarbonate), increasing oxygen delivery to the tissue by increasing oxygen saturation through manipulations in SVR (phenylephrine), and increasing cardiac output through the administration of potent inotropic agents such as epinephrine via infusion or bolus to maintain cardiac output while the surgeon attempts to rapidly institute CPB. Once CPB is instituted, high flow rates should be established and maintained until the patient is fully resuscitated as judged by arterial and mixed venous blood gases and visual inspection of the heart.

Alternatively, a brief trial of placing a partial cross-clamp across the proximal aorta to reverse right-to-left shunting may be tried. This technique may increase PBF, reverse right-to-left shunting, and provide improved oxygen delivery before establishment of CPB. In patients with severe cyanotic spells, the administration of β-blocking agents may be counterproductive in that oxygen delivery may become compromised and myocardial ischemia or cardiac arrest may ensue. Hand ventilation with rapid ventilatory rates or high mean airway pressures should be discouraged during a hypercyanotic spell, because this increases intrathoracic pressure, reduces cardiac filling, and exacerbates dynamic obstruction at the RVOT.

Tetralogy of Fallot with an absent pulmonary valve is an unusual variant that occurs in 3% to 6% of tetralogy patients. The pulmonary valve is minuscule, dysplastic, and nonfunctional. Because of the absence of functional valve tissue, the RVOT remains widely patent despite moderate narrowing of the infundibulum. The associated VSD is large and malaligned. In the severe form of this disease, severe pulmonary regurgitation during gestation results in massive aneurysmal dilation of the main and branch PAs. The large conducting airways are compressed by these vessels, resulting in tracheobronchial malacia. In addition, the distal vessels do not branch in a typical segmental fashion but instead end in vascular tufts that entrap and compress the distal conducting airways. These patients present with excessive PBF, heart failure, and respiratory disease, characterized by obstruction of gas flow and increased airways resistance in the small, medium, and large airways ( Fig 17-17 ). Tetralogy of Fallot with absent pulmonary valve is associated with high morbidity and mortality.


FIGURE 17-17  Diagrammatic representation of normal relationship between pulmonary artery and left bronchus (A) and compression of left main bronchus by the aneurysmal pulmonary artery (B) in the absent pulmonary valve syndrome. Note also compression of smaller bronchi that are entwined by tufts of abnormal vessels.  (From Rabinovitch M, Grady S, David I, et al.: Compression of intrapulmonary bronchi by abnormally branching pulmonary arteries associated with absent pulmonary valves. Am J Cardiol 50:804, 1982. With permission from Excerpta Medica, Inc.)




Occasionally, palliation with a BT shunt is considered a preferred operation in tetralogy. In newborns with unrelenting hypercyanotic episodes and poor systemic perfusion who are either unresponsive or poorly responsive to medical management, a BT (or systemic-to-pulmonary) shunt can be placed. This is generally done off bypass. For newborns with tetralogy of Fallot who have hypoxic-ischemic encephalopathy or intraventricular hemorrhage, a BT shunt avoids the risks of heparinization and CPB. In general, however, exchanging tetralogy physiology for shunt physiology is not optimal. Prolonged exposure to shunt physiology risks the development of a dilated cardiomyopathy and/or pulmonary hypertension.

Atrioventricular Septal Defects

AVSDs are physiologically similar to large VSDs. AVSDs consist of an ostium primum ASD, an inlet VSD, and a cleft in the septal leaflets of the tricuspid and mitral valves (Figs. 17-18 and 17-19 [18] [19]). The management of these patients is quite similar except they are more prone to PA hypertension. This is due to the large left-to-right shunt imposed by the septal defect. AVSDs are commonly found in patients with trisomy 21. This genetic defect is associated with small peripheral airways in the lung and architecturally small PAs. Elevation in PVR may be due to an underdeveloped pulmonary vascular tree as well. Fixed PA hypertension has been described in trisomy 21 patients with AVSDs who are younger than 6 months ( Frescura et al., 1987 ).


FIGURE 17-18  Normal atrioventricular valves consisting of leaflets that originate from an annulus and attach distally to papillary muscles by chordae tendinea.  (From Lowe DA: Abnormalities of the atrioventricular valves. In Lake CL, editor: Pediatric cardiac anesthesia. Norwalk, CT, 1988, Appleton & Lange, p 300.)





FIGURE 17-19  Complete atrioventricular canal in view from right atrium. ASD, atrial septal defect; VSD, ventricular septal defect.  (From Cooley DA, Norman JC: Techniques in cardiac surgery. Houston, 1975, Texas Medical Press, p 128.)




Valvular regurgitation is not an uncommon finding after AVSD repair ( McGrath and Gonzales-Lavin, 1987 ). Mild to moderate regurgitation is generally well tolerated and is managed by decreasing systemic afterload. If concomitant ventricular dysfunction exists with AVV regurgitation, low-dose dopamine, dobutamine, or milrinone may be added. Milrinone is an extremely effective drug in this patient population. If severe AVV regurgitation exists, valve replacement may be necessary. Because the defect involves both annular and valve tissue, the artificial valve is difficult to position and there is a greater risk of perivalvular leaks. Complete heart block carries an increased risk after valve replacement in AVSDs, although it is generally uncommon ( Kirklin and Barratt-Boyes, 1986a ). This is due to a lack of normally developed annular tissue necessary to suture the mechanical valve in place and an increased likelihood of disrupting normal conduction tissue. Patients with AVSDs who do not have Down syndrome usually have greater dysplasia of valve tissue.

Transposition of the Great Arteries and Arterial Switch Operation

In TGA, the aorta arises from the RV and the PA arises from the LV. The arterial switch operation for TGA is the preferred surgical approach ( Fig. 17-20 ); it allows the LV to be the systemic ventricle. Maintaining the LV as the systemic chamber improves ventricular longevity and reduces the incidence of systemic ventricular dysfunction. The cylindrical shape, the concentric contraction, and the location of the ventricular inlet and outlet at the base of the heart of the LV allow it to function better as a pressure pump and to account for improved long-term patient survival and reduced long-term morbidity compared with alternative surgical repairs.


FIGURE 17-20  Arterial switch procedure. (A) Aorta is transected, and left and right coronary arteries and their bases are excised. (B) An equivalent segment of pulmonary arterial wall is excised, and coronary arteries are sutured to the pulmonary artery. (C) Distal pulmonary artery is brought anterior to ascending aorta, and the proximal pulmonary artery is anastomosed to the distal aorta. (D) Sites of coronary artery extraction are repaired with use of either (a) a patch of prosthetic material or (b) a segment of pericardium. Finally, proximal aorta is sutured to distal pulmonary artery.  (Modified from Castaneda AR, Norwood WI, Jonas RA, et al.: Transposition of the great arteries and intact ventricular septum: anatomical repair in the neonate. Ann Thorac Surg 38:438, 1984, with permission from the Society of Thoracic Surgeons.)




Improvements in coronary transfer technique, myocardial preservation, and postoperative management have reduced perioperative mortality rates for this operation to less than 3% in simple transposition without other associated cardiac anomalies ( Blume et al., 1999 ). In a report of 223 consecutive transposition patients, coronary artery anatomy was not found to be associated with increased mortality, although the presence of coronary abnormalities, including single right coronary, was associated with a longer duration of mechanical ventilation and need for delayed sternal closure ( Blume et al., 1999 ). These low mortality rates rival those of the atrial switch procedure and are more impressive in view of the high incidence of late complications associated with atrial switch procedures (atrial dysrhythmias [50%], baffle stenosis [8% to 10%], pulmonary venous obstruction, [8% to 10%] and right [systemic] ventricular dysfunction [10% to 20%]) ( Castaneda et al., 1989 ). The Congenital Heart Surgeons Society reviewed outcomes of 829 neonates from 24 institutions with complete TGA repaired during the years 1985 to 1989. This was the era when congenital heart centers began transitioning from the atrial switch procedure to the arterial switch procedure. In this review, the survival rate of the 516 children managed by an arterial switch operation was compared with the 285 atrial switch operation rate and found to be virtually identical (81% and 82%, respectively). In that era, the early mortality rate for the arterial switch procedure was 19%. The late mortality rate remained higher in the atrial switch group, accounting for the 15-year similarity in mortality ( Williams et al., 2003 ).

The arterial switch operation involves mobilization of the PA and dissection onto the right and left PAs (see Fig 17-20 ). The aorta and main PA are then transected. The aorta is divided above the coronary arteries and aortic valve. The LeCompte maneuver (the passage of the PA anterior to the aorta) is then performed. The coronary arteries are mobilized with 3 to 4 mm of surrounding aortic tissue and reimplanted onto the “neoaorta.” Previously harvested patches of pericardium are used to close the defects resulting from removal of the coronary arteries.

Two factors dictate the survivability of a patient after an arterial switch operation: (1) global function of the LV and (2) focal ischemia ( Wernovsky et al., 1988 ). Global ventricular function reflects how well the LV is prepared to pump against a systemic pressure load after the arterial switch procedure. In a patient who does not have VSD, the left ventricle pumps against the low resistance of the pulmonary circulation. In 2 to 4 weeks, the LV will lose its ability to pump against the systemic afterload after the arterial switch procedure ( Danforth et al., 1960 ). In these patients, the LV must be either “prepared” by producing a pressure and volume load on the LV or by anticipating the need for a VAD or ECMO after weaning from CPB. The two-stage approach to preparing the LV is accomplished intraoperatively with PA banding, which creates a left ventricular pressure of 60% of systemic or greater, and the placement of a 4-mm aorta-to-pulmonary shunt in order to pressure and volume load the LV, respectively. These patients are frequently quite ill, but left ventricular function recovers to the point of being able to pump against systemic afterload. After the LV is “conditioned” over a period of 5 to 7 days, a second operation, an arterial switch operation, is immediately performed ( Lacour-Gayet et al., 2001 ).

Inotropic support and afterload reduction are very effective in maximizing cardiac output after weaning patients from CPB after the arterial switch procedure. If inotropic support is excessive and cardiac output and systemic perfusion remain marginal, early introduction of a VAD or ECMO should be considered because myocardial function predictably worsens in the first 6 to 12 postoperative hours (Lacour-Gayet et al., 2001 ). Focal ischemia may also occur. Kinking, twisting, or excessive tension placed on the coronary arteries and compression of the coronary arteries because the main PA crosses tightly over the aorta (due to inadequate dissection of the branch PAs in combination with the LeCompte maneuver) are causes of focal ischemic changes in these patients. ST segment changes or regional wall motion abnormalities on intraoperative echocardiography should alert the anesthesiologist and surgeon to evaluate the coronary arteries and overlying structures.

Transposition of the Great Arteries: Senning or Mustard

Atrial repair of TGA is an example of a physiologic rather than a true anatomic correction ( Fig. 17-21 ). Although indications for atrial-level repairs for transposition are becoming less and less common, unusual coronary anatomy, such as coronary arteries with an intramural course, have a higher risk for coronary transfer in the arterial switch operation. Under these circumstances, atrial level repair may be considered.


FIGURE 17-21  Diagram of interior of right atrium showing the baffle that rearranges and redirects atrial blood flow following Mustard procedure. Pulmonary vein openings are visualized. CS, coronary sinus; MV, mitral valve; IVC, inferior vena cava; TV, tricuspid valve. (From Reed CC, Stafford TB, editors: Cardiopulmonary bypass, 2nd ed. Houston, 1985, Texas Medical Press, p 78.)




The Senning operation uses tissue from the right atrial wall and interatrial septum to fashion a conduit for the flow of deoxygenated (systemic venous) blood through the mitral valve and into the left (pulmonary) ventricle. The LV ejects this deoxygenated blood into the PA. The Mustard operation accomplishes the same physiologic correction as the Senning procedure, but it uses a pericardial or prosthetic baffle to redirect the flow of systemic venous blood (see Fig. 17-21 ).

Patients who underwent either the Mustard or the Senning procedure generally do well. Because the operative site is in the atrium, ventricular function is usually well preserved. In patients with low cardiac output after weaning from CPB, baffle obstruction versus primary myocardial dysfunction should be considered. Poor cardiac output and/or symptoms of SVC obstruction are highly suggestive of a baffle obstruction. The most common location for baffle obstruction is where the SVC channel crosses the atrial septum. Superior vena cava obstruction occurs in approximately 8% to 10% of patients ( Pacifico et al., 1990 ). If there is good collateral venous drainage and the obstruction is not severe, it is usually well tolerated. However, an occasional patient requires revision of the baffle.

Dysrhythmias are the most common complication after atrial correction of transposition. The incidence of dysrhythmias ranges from 20% to 60%, with the higher number being more common with the Mustard repair ( Gillette et al., 1980 ). Although bradycardia is most common, supraventricular tachyarrhythmias are also observed. In the Mustard repair, the wide excision of atrial tissue and long atrial suture line may damage the sinoatrial node or its blood supply. Temporary atrial and ventricular pacing wires should be routinely placed in these patients.

Rastelli Procedure

The Rastelli procedure (an operation for TGA with a VSD and an LVOT obstruction) combines an intracardiac baffle or tunnel from the VSD to the aorta and an extracardiac conduit to the pulmonary artery ( Fig. 17-22 ). The intracardiac baffle establishes continuity between the LV and the aorta. A large nonrestrictive VSD is required to allow uninterrupted flow between the LV and aorta. The VSD may have to be surgically enlarged during the operative procedure to prevent flow restriction. The extracardiac conduit establishes continuity between the RV and the PA.


FIGURE 17-22  (A) In the Rastelli procedure, closure of ventricular septal defect results in the left ventricle (LV) ejecting blood into the aorta (AO). (B) An external right ventricle (RV)-to-pulmonary artery (PA) conduit is placed to provide pulmonary circulation.  (From Strafford M: Transposition of the great vessels. In Lake CL, editor: Pediatric cardiac anesthesia. Norwalk, CT, 1988, Appleton & Lange, p 237.)




Postoperatively, LVOT obstruction can occur due to inadequate enlargement of the VSD or obstruction of the intracardiac tunnel, resulting in an acute pressure load on the LV. High left-sided filling pressures, low cardiac output, high inotropic requirement, and a dilated poorly contractile LV are typical findings. Intraoperative echocardiography and pressure monitoring across the baffle and VSD are helpful to localize the area of left ventricular obstruction (VSD, baffle, etc.). During post-CPB assessment in patients with low cardiac output, the anesthesiologist must maintain preload, coronary perfusion pressures, systemic afterload, and contractility. Patients with baffle obstruction have customary signs of severe left-sided obstructive physiology ( Kreselis et al., 1985 ).

The extracardiac RV-to-PA conduit can also be obstructed. Although this is more commonly seen in late follow-up, early postrepair obstruction may become apparent after chest closure. This results more commonly from extrinsic compression of the conduit by the sternum. The conduit can also drape across a coronary artery and, once the chest is closed, impinge on right coronary artery flow. Furthermore, the conduit can cause compression of the trachea or the right bronchus. Sudden changes in lung compliance, airways resistance, hypercarbia, or hypoxia suggest airway compression as a possible cause (Dekeon et al., 1992 ). Respiratory mechanics monitoring has been helpful in diagnosing this problem intraoperatively.

Heart block is an uncommon finding after Rastelli repair; however, if the VSD must be surgically enlarged, the risk of transient or permanent heart block increases. All patients undergoing VSD enlargement should have temporary atrial and ventricular pacing wires placed. It is important to ensure low pacing thresholds in the operating room, because postoperative myocardial edema usually results in increasing pacing thresholds. In neonates and infants with heart rate-dependent cardiac output, pacing wire failures resulting in a slow idioventricular rhythm are poorly tolerated and may become life threatening in the infant with marginal cardiac output in the early postoperative period.

Interrupted Aortic Arch

Interrupted aortic arch (IAA) is a congenital anomaly in which there is complete discontinuity between the ascending and descending aortas ( Fig. 17-23 ). IAA almost always occurs in association with a VSD and a patent PDA. The PDA connects the PA to the descending aorta, supplying blood flow to the lower body. Occasionally, IAA may be associated with truncus arteriosus, and it is rarely associated with other cardiac defects such as double-outlet RV. There is often a stretched PFO or ASD. The stretched PFO probably develops as a result of the in utero left-to-right shunt from the downstream aortic obstruction. IAA is classified based on the location of the interruption. Type A interruption occurs at the isthmus, which is distal to the left subclavian; type B interruption occurs between the left common carotid artery and the left subclavian and is the most common type, with an incidence between 50% and 70% in most reported series. Type C interruption is between the innominate and the left common carotid artery ( Celoria et al., 1959) . If there is no associated VSD, then there may be an aortopulmonary window (a direct connection between the ascending aorta and the main PA). Interrupted aortic arch is rare and occurs in 1% of patients with CHD ( Sandhu et al., 2002) . Type B patients may have a higher incidence of chromosomal anomalies, including DiGeorge syndrome, with thymic hypoplasia and hypoparathyroidism ( Schreiber et al., 2000 ).


FIGURE 17-23  Types of aortic arch interruption. (A) In type A, the defect is distal to the left subclavian artery. (B) Type B has the interruption between the left subclavian (LS) and left carotid (LC) arteries. (C) Type C has the discontinuity between the innominate (IA) and left carotid arteries. Ao, aorta; LPA, left pulmonary artery; MPA, main pulmonary artery; PDA, patent ductus arteriosus; RPA, right pulmonary artery.  (From Arciniegas E, editor: Pediatric cardiac surgery. Chicago, 1985, Year Book Medical, p 109, with permission.)




Unless symptoms are detected prenatally with fetal echocardiography, they begin acutely with ductal closure, when infants acutely develop severe metabolic acidosis and poor peripheral perfusion. Descending aorta blood flow decreases dramatically as the lower body is dependent on a minimal amount of collateral blood flow. There is ischemia to the liver, kidney, and intestine, and lower extremity pulses are usually absent.

Resuscitation is dependent on reestablishing ductal patency through the administration of PGE1and correction of acidosis. Occasionally, the ductus may not close, in which case the child presents with congestive cardiac failure from a left-to-right shunt through the ductus.

After stabilization and optimization of organ function, the patient is usually maintained on a low FIo2 to diminish left-to-right flow across the ductus and occasionally requires an elevated CO2 through hypoventilation or exogenous CO2 administration in order to optimize systemic blood flow. If instituted in the ICU, these management strategies should be continued through the pre-CPB period. After induction of anesthesia, placement of a lower limb and upper limb arterial catheters aids in CPB perfusion management and in the postbypass period helps to determine if a residual arch obstruction exists.

In most surgical centers a single stage repair with end-to-end anastomosis of the mobilized aorta and repair of all associated cardiac anomalies is preferred ( Sell et al., 1988 ; Schreiber et al., 2000 ). Several studies have demonstrated a significantly higher mortality rate for staged repairs ( Tlaskal et al., 1999 ; Schreiber et al., 2000 ). With adequate mobilization of the descending aorta, an interposition graft is rarely indicated.

Postsurgical complications include bleeding from dissection and mobilization of the aorta. If the aorta was stretched to complete the repair, the aorta may be under tension and obstruction of the left main bronchus can occur with evidence of air trapping on chest radiograph ( Casteneda et al., 1994 ). Left ventricular dysfunction is common after repair. Residual LVOT obstruction occurs commonly, and may contribute to high postoperative afterload. Afterload reduction with vasodilators (nitroprusside or nicardipine) or milrinone is helpful when separating from CPB but is not helpful when a fixed obstruction exists at the site of the repair. Reoperative interventions or catheterization lab procedures are commonly required to treat arch obstruction ( Sell et al., 1988 ; Castenada et al., 1994 ; Schreiber et al., 2000 ).

Ebstein's Anomaly

This is an uncommon cardiac defect of the tricuspid valve and RV. Posterior and septal leaflets of the tricuspid valve are displaced and dysplastic and the attachments to the tricuspid valve annulus are abnormal. The anterior leaflet is enlarged and often described as “sail like” in appearance. The effective tricuspid valve orifice is displaced apically into the RV. The RA is usually massively dilated and the junction between the RA and RV is enlarged. Because of the displacement of the tricuspid valve into the RV, the RV now consists of 2 parts. The inlet portion is described as “atrialized,” its wall is thin and aneurysmal. The outlet portion is referred to as the “functional” portion of the RV ( Fig. 17-24 ).


FIGURE 17-24  Anatomic features of Ebstein's anomaly. (A) Displacement of the posterior and septal leaflets into the right ventricle results in a large atrialized chamber and tricuspid valve incompetence. AV, atrioventricular. (B) External appearance. (C) Ebstein's anomaly, showing the atrialized portion of the right ventricle, with the tricuspid valve extending deep into the right ventricle and the intraventricular septum bowing into the left ventricle.



Ebstein's anomaly is a relatively rare defect accounting for less than 1% of all congenital cardiac defects. The pathophysiology depends on the extent of the tricuspid valve defect. In general, these patients have significant tricuspid valve regurgitation which reduces the volume of the blood entering the functional RV and creating a large regurgitant volume contributing to a dilated RA. If there is an associated ASD or a PFO, there is a right-to-left shunt at the atrial level. The small functional portion of the RV may have a limited stroke volume based on small size contributing to further tricuspid regurgitation. In infants with diastolic dysfunction in this small outlet portion of the RV, filling may be further impaired lessening forward flow of blood into the pulmonary circulation. An additional physiologic problem with this lesion relates to ventricular interactions. Because of the large RA and displacement of the tricuspid valve, the intraventricular septum tends to bow into the left ventricular cavity during diastole. This reduces left ventricular filling and decreases systemic cardiac output. Newborns with Ebstein's anomaly usually present with severe cyanosis. In less severe forms, they may present with supraventricular tachycardia or a murmur of tricuspid regurgitation.

Severe forms of Ebstein's anomaly presenting in neonates are usually fatal ( Lerner et al., 2003 ). Several operative interventions are described for Ebstein's anomaly. The two-ventricle repair requires a tricuspid valve annuloplasty, radical reduction atrioplasty, and closure of ASD (most groups fenestrate the patch). In addition, the atrialized portion of the RV is plicated, obliterating the aneurysmal cavity, and the large anterior leaflet is allowed to function as a monocusp valve. Further, an RVOT patch is placed to create a normal-size (7- to 8-mm) outflow tract ( Knott-Craig et al., 2002 ). Less commonly, right ventricular exclusion procedures may be performed. The RV size is reduced, the tricuspid valve is oversewn, and the coronary sinus is rerouted to the LA through an ASD ( Sano et al., 2002 ).

Anesthetic management is directed at maintaining right ventricular contractility with inotropic support and right ventricular afterload reduction. Induction of anesthesia should be cautious. In general, an intravenous induction with midazolam and fentanyl is recommended. In the sick neonate, inhalation induction may be poorly tolerated, particularly with halothane because of the high risk of atrial and ventricular dysrhythmias. Supraventricular tachycardia occurs in approximately 30% of patients and those with ventricular tachyarrhythmias, including ventricular fibrillation, may occur commonly after separation from CPB. If ventricular ectopy is present, a lidocaine infusion starting at 40mcg/kg per minute should be administered. Alternative therapy includes procainamide or amiodarone. TEE should be used to evaluate the repair and the function of both the RV and LV. Particular attention should be given to the tricuspid valve, RVOT, and intraventricular septum.


Palliative surgery is reserved for cardiac lesions in which anatomic parts are missing, such as pulmonary atresia with intact septum (absent RV and PA), tricuspid atresia (absent RV and tricuspid valve), HLHS (absent mitral valve and small LV), mitral atresia (absent LV), or univentricular heart (absent RV or LV) ( Castaneda et al., 1989 ). Underdeveloped or absent chambers and great vessels preclude complete reconstruction of a normal four-chambered heart. The intraoperative management of patients receiving palliative operations is not intuitive. These patients remain the most difficult group of patients to care for because blood flow patterns are abnormal, and the impact of preload augmentation, inotropic support, and afterload manipulations may redirect blood flow patterns as well as simply improve cardiac output and tissue oxygen delivery ( Table 17-14 ). In addition, minor changes in PA pressure, PVR, AVV function, and ventricular performance may significantly increase postoperative morbidity ( Pigott et al., 1988 ; Bridges et al., 1990 ).

TABLE 17-14   -- Single-ventricle anatomy and surgical palliation


Surgical Therapy

Tricuspid atresia

Balloon atrial septostomy➙cavopulmonary anastomosis➙Fontan

Malaligned AV canal with hypoplastic right ventricle

Pulmonary artery band➙cavopulmonary anastomosis➙Fontan

Malaligned AV canal with hypoplastic left ventricle

DKS➙cavopulmonary anastomosis➙Fontan

Pulmonary atresia with intact ventricular septum

Modified BT shunt➙cavopulmonary anastomosis➙Fontan

Pulmonary atresia with discontinuous pulmonary arteries

Unifocalization with shunt➙cavopulmonary anastomosis➙Fontan

Double-outlet right ventricle with mitral atresia

DKS/Norwood/Sano➙cavopulmonary anastomosis➙Fontan

Double-inlet left ventricle

DKS➙cavopulmonary anastomosis➙Fontan

Hypoplastic left heart syndrome

Norwood/Sano➙cavopulmonary anastomosis➙Fontan

Critical aortic stenosis with hypoplastic left ventricle

Norwood/Sano➙cavopulmonary anastomosis➙Fontan

Subset of heterotaxy syndromes

Possible DKS➙cavopulmonary anastomosis➙Fontan

AV, Atrioventricular; BT, Blalock-Taussig; DKS, Damus-Kaye-Stanzel; Fontan procedure.




The ultimate goal of palliative surgery is to separate the pulmonary and systemic circulation and in the case of a univentricular heart, to use the single ventricle as a systemic pumping chamber. The Fontan operation, initially described in 1971, and its modifications have been adapted to all forms of univentricular heart, including HLHS ( Fontan et al., 1983 ; Mayer et al., 1986 ). The modified Fontan operation provides PBF by anastomosing all of the systemic venous return (SVC and IVC) to the PAs without an interposed pumping chamber. PBF is therefore dependent on maintaining a passive gradient from the systemic veins through the pulmonary vasculature and into the LA. This arrangement requires a low PVR and low left ventricular filling pressures. Since newborns have high PVR, they are not suitable candidates for a Fontan-type procedure. The goal in the newborn period is to achieve balanced blood flow between the pulmonary and systemic circulation until the PVR falls, rendering the child suitable for a modified Fontan procedure or more commonly a cavopulmonary anastomosis.

Hypoplastic Left Heart Syndrome

A palliative procedure is performed in the newborn period to limit excessive blood flow to the pulmonary circulation and provide balanced blood flow between the systemic and pulmonary circulation. In patients with two normal-sized great arteries this may be accomplished by placing a restrictive band around the main PA. If the main PA is very small or atretic and PBF is primarily from the ductus arteriosus, PBF is maintained by placing a restrictive PA-to-aorta shunt and ligating the ductus arteriosus, thereby limiting PBF and reducing the volume load on the single ventricle. In more complex anatomy, such as HLHS, the entire aorta must be reconstructed as well ( Table 17-14 ).

In the HLHS, there is only a functional RV. The LV is severely hypoplastic, and the mitral valve is severely stenotic or atretic. In addition, the aortic valve is either atretic or severely stenotic, so much so that the PDA provides retrograde flow to the coronary circulation via the transverse and hypoplastic ascending arch and antegrade flow to the descending aorta. The single RV has to pump blood to both the pulmonary circulation (through the PA) and the systemic circulation (through the ductus arteriosus). The direction of predominant blood flow is dependent on downstream resistance, that is, the flow is determined by the PVR and the SVR. Ideally, balanced flow results in an equal portion of cardiac output going to both the systemic and pulmonary vascular beds, resulting in a 1:1 shunt ( Fig 17-25 ).


FIGURE 17-25  Hypoplastic left heart syndrome before surgery, showing the importance of maintaining a patent ductus arteriosus to maintain systemic cardiac output and balancing    p/   s. LV, left ventricle; PA, pulmonary artery; RA, right atrium; RV, right ventricle. (With permission from Rudolph A: Congenital diseases of the heart. Year Book Medical Publishers, Chicago. 1985, 2001.)




Operative repair for patients with HLHS requires a series of staged procedures culminating in the modified Fontan procedure. In infancy, the operative objective is to establish an outflow from the single RV to the systemic circulation. This is accomplished by performing an atrial septectomy to allow pulmonary venous blood to mix with systemic venous blood, transecting the main PA, creating a neoaorta using pulmonary homograft, and connecting it from the proximal pulmonary aorta to the hypoplastic ascending aorta ( Norwood et al., 1983 ; Norwood and Murphy, 1990 ) ( Fig 17-26 ). This single great vessel provides systemic and coronary blood flow. PBF is provided by a Gore-Tex shunt from the innominate artery, subclavian artery, or, rarely, the neoaorta to the main PA. The Sano modification of the stage 1 Norwood palliative procedure uses a Gore-Tex shunt from the systemic ventricle to the PA ( Sano et al., 2003 ). The first stage is a palliative procedure that eliminates the need for ductal patency. If not previously present, a large ASD is created intraoperatively, forming the RA and LA into a common chamber.


FIGURE 17-26  (A) Classic Norwood Stage 1 operation with a modified Blalock-Taussig shunt (BTS) from the innominate artery to the main pulmonary artery (PA). Also shown is reconstruction of the neoaorta and atrial septectomy. (B) Sano-modified Norwood operation, showing the right ventricle-to-PA conduit rather than a modified BT shunt. Ao, aorta; RA, right atrium.



Perioperative management of patients with HLHS is based on maintaining effective systemic cardiac output and then balancing flow between the systemic and pulmonary circulations and minimizing the volume load on the heart—maintaining a pulmonary-to-systemic ratio of 1. This can be achieved by maintaining Sao2 in the range of 70 to 80. In the preoperative period, excessive PBF is the most common problem; anesthetic management strategies are therefore directed at reducing PBF. Regardless of the anatomic lesion and the surgical technique required to achieve shunt-dependent blood flow, the physiologic goals are the same—to maintain effective systemic cardiac output, minimize volume load on the single ventricle, restrict excessive PBF, and maintain balanced flow between the pulmonary and systemic circulation ( Hansen and Hickey, 1986 ; Norwood and Murphy, 1990 ; Norwood et al., 1991) . If pulmonary resistance is low, then a greater proportion of the cardiac output is directed to the lungs and the single ventricle is required to increase its cardiac output by increasing stroke volume and heart rate ( Pigott et al., 1988 ). The newborn heart has a reduced capacity to increase stroke volume, especially when volume loaded. When flow becomes unbalanced and a large left-to-right shunt provides a disproportionate amount of blood flow to the lungs, systemic hypoperfusion and metabolic acidosis result. Once low systemic perfusion is evident, therapy must be implemented to improve cardiac output, correct metabolic acidosis, and redistribute blood flow to the systemic circulation.

Postbypass and postoperative management are based primarily on increasing total cardiac output through the use of inotropes and phosphodiesterase type III inhibitors and, secondarily, on balancing blood flow between the systemic and pulmonary circuits if PBF is excessive ( Table 17-15 ). A systemic arterial saturation of 75% to 80% with a mixed venous saturation of 50% to 60% gives a 1:1 shunt or balanced pulmonary/systemic flow. Lower mixed venous saturations are associated with poor cardiac output, and the inotrope support should be initiated or increased ( Tweddell et al., 2000 ). Mixed venous oxygen saturation monitoring measured from the internal jugular venous catheter placed in the high SVC is an important monitor for assessing cardiac output and balancing pulmonary/systemic flow.

TABLE 17-15   -- Effects of aortic and mixed venous saturation on pulmonary and systemic flows in patients with single ventricles

Aortic Saturation (%)



Clinical Considerations




Balanced shunt




Mild reduction in systemic flow




Severe reduction in systemic flow



Infinite shunt




Mild reduction in pulmonary flow




Severe reduction in pulmonary flow




Decreased cardiac ouput, increased    p




Decreased cardiac output, increased    p

   p, pulmonary blood flow;    s, systemic blood flow.




A balanced shunt can be calculated from the pulmonary/systemic flow equation:

In this physiology, PA and aortic saturations are equal. Estimating balanced shunt flow just by the aortic saturations is not accurate because low mixed venous saturations and excessive shunt flow can result in systemic saturations of 75%. A reliable way of assessing shunt physiology intraoperatively is through continuous monitoring of arterial saturation with pulse oximetry and intermittent or continuous measures of mixed venous saturations. As a surrogate measure of continuous venous oximetry, a cerebral oximetry probe can be used in these patients. The absolute number may not be 100% accurate, but as a trend monitor, it may be useful. If mixed venous saturations are not monitored, low arterial saturations may reflect low total cardiac output rather than reduced PBF, as already noted. Inotropic agents are important in improving systemic cardiac output in these patients. When comparing epinephrine, dopamine, and dobutamine, only epinephrine significantly improved systemic cardiac output and systemic oxygen delivery in a piglet model of HLHS; this is because epinephrine increased total cardiac output without increasing    p/   s. In contrast, dobutamine increased total cardiac output but increased    p/   s to greater than 2:1 and therefore systemic oxygen delivery worsened. In this model dopamine had only a modest increase in total cardiac output and did not appear to affect    p/   s, so systemic oxygen delivery was statistically unchanged ( Riordan et al., 1996 ). It has become increasingly apparent that increasing systemic cardiac output is critical in managing these patients and that inotropic support in conjunction with systemic afterload reduction is the preferred medical management strategy.

Another approach to the management of elevated SVR in the postbypass period is with the use of systemic vasodilators. Milrinone (0.5 mcg/kg per minute) in the postbypass and postoperative period and potent systemic vasodilators such as phenoxybenzamine (0.25 mg/kg) added to the bypass circuit represent two such strategies. The more commonly applied strategy initiates milrinone or other unloading agents shortly before separation from CPB and uses inotropic agents to optimize systemic cardiac output. The main advantage of this strategy is the ability to titrate increases in cardiac contractility and afterload reduction. The use of phenoxybenzamine on pump has the disadvantages that it is a potent noncompetitive antagonist of αARs that may take upward of 24 hours to regenerate, it is not titratable, and it may require low-dose norepinephrine (0.1 mcg/kg per minute) to increase mean blood pressure ( Tweddell et al., 2000 ).

Ventricular dysfunction in the stage I Norwood patient with a systemic-to-pulmonary shunt is due to shunt-dependent volume overload and reduced coronary blood flow from diversion of blood through the systemic-to-pulmonary shunt. For this reason, there has been an increasing interest in alternative strategies to the initial surgical management of HLHS. The Sano modification, which replaced the systemic-to-pulmonary shunt with a shunt from the RV to the PA, preserves coronary perfusion by preventing diastolic runoff. Several small series have demonstrated significant improvement in diastolic blood pressure, myocardial function, and inotrope use ( Maher et al., 2003 ; Mahle et al., 2003 ; Mair et al., 2003 ; Pizarro et al., 2003 ; Sano et al., 2003 ) (see Fig. 17-26 B). Bicarbonate administration is also an important adjunct to low cardiac output in shunt-dependent patients. Sodium bicarbonate corrects metabolic acidosis and provides an optimal pH for inotropic effect.

Vasodilators such as nitroprusside reduce SVR and promote systemic cardiac output. In combination with ventilatory manipulations and bicarbonate administration, vasodilators should augment systemic perfusion. Vasodilators should be used cautiously, however, if systemic pressure is low. Volume support inotropic agents should be readily available when nitroprusside is administered. Milrinone is generally a preferred first-line drug, but if blood pressure is adequately supplemented with inotropic support, sodium nitroprusside can significantly augment systemic perfusion.

Studies have demonstrated the beneficial use of inhaled CO2 as a pulmonary vasoconstrictor in patients with low PVR ( Norwood et al., 1992 ; Bradley et al., 2001 ; Keidan et al., 2003 ). This technique has proved to be most useful in the prebypass period as opposed to the post-CPB period ( Keidan et al., 2003 ). Patients almost always need to be paralyzed with a neuromuscular blocking agent when exogenous inhaled CO2 is added. Sedation alone is usually ineffective. Intraoperative control of excessive PBF in the prerepair period is essential. Surface cooling and opening of the sternum reduce CO2production and decrease PVR, respectively. The administration of 1% to 2% CO2 into the fresh gas flow may be beneficial in controlling PBF. The benefits of exogenous administration of CO2 as opposed to controlled hypoventilation are that lung volumes can be maintained while PBF is restricted.

Hypoxic gas mixtures (Fio2 = 17%) have also been advocated to control PBF. The benefits of hypoxic gas mixtures are that they can be administered without the need for heavy sedation or the use of muscle relaxants. In centers where heart transplantation has been the preferred procedure for HLHS, newborns are extubated into hypoxic environments while waiting weeks to months for a donor heart. The disadvantage of a hypoxic gas mixture is decreased oxygen delivery. Impaired oxygen delivery becomes increasingly problematic if significant pulmonary edema, lung disease, or worsening ventricular function develops. In a randomized crossover controlled study, Tabbut and others (2001) noted that in 10 paralyzed and ventilated preoperative HLHS patients, 2.7% inspired CO2was found to improve delivery of oxygen compared with 17% Fio2. A hypoxic gas mixture is rarely used after the first-stage repair for HLHS for reasons outlined earlier.

If PVR is significantly elevated, a more common occurrence during weaning from CPB, then inadequate PBF is the result and the patient becomes moderately to severely hypoxemic. Inspired Fio2of 1.0%, alkalinization, nitric oxide administration, and ensured adequate levels of anesthesia can reduce PVR and improve PBF. Persistent hypoxemia or significant hypercarbia unresponsive to ventilatory manipulations may be due to shunt dysfunction (clotting or kinking) and should be investigated by intraoperative echocardiography and visual inspection of the shunt. Urgent resumption of CPB is indicated if shunt flow remains inadequate despite these maneuvers.

Postbypass low cardiac output could be due to AVV regurgitation, neoaortic valve insufficiency, inadequate atrial septectomy, and coronary insufficiency due to kinking or poor flow through the native diminutive ascending aorta.

The end result of this first-stage procedure is the creation of a univentricular heart in which the single RV is directly connected to the systemic circulation. The newly created shunt connects either a branch of the aorta (innominate) to the PA or, in the Sano modification, a direct connection of the single ventricle to the PA. These procedures do not eliminate the need for balancing blood flow between pulmonary and systemic circulations but generally restrict PBF to a greater extent than the PDA. The palliative stage 1 procedure sets the stage for later correction with a cavopulmonary anastomosis with a bidirectional Glenn or hemi-Fontan procedure or, rarely, directly to a modified Fontan, as discussed later.

Another novel approach to the first-stage management of HLHS has been a combined cardiac catheterization and operating room approach to limit PBF and either place a stent to open the ductus or surgically place a modified BT shunt. This approach was first described by Gibbs and others (1993) . A small series of eight patients with HLHS were reported by Muller and colleagues (2003) in which a catheterization laboratory-based approach to stage I palliation was used. In this series, the initial management included ductal stenting and balloon atrial enlargement if indicated. One to 3 days later, a surgically placed PA band was used to restrict PBF. There were no deaths in the initial palliation. The second-stage procedure requires a combined stage 1 and stage 2 operative intervention; this includes construction of a neoaorta and placement of a bidirectional Glenn. In this small series, there was one death after the combined stage 1 and 2 procedure (12.5% mortality).

The overall survival of the first-stage repair for HLHS (Norwood stage I) has been reported as 68% to 77% ( Norwood and Murphy, 1990 ; Norwood et al., 1991, 1992 [340] [341]; Bove et al., 1996 ; Kern et al., 1997). A comparison of survival rates from Children's Hospital of Philadelphia demonstrates the effects on survival of newer technical advances in the management of HLHS ( Mahle et al., 2000 ) (Fig 17-27 ). In particular, early survival (first 120 days) in the era of 1984 to 1988 was 56%; from 1995 to 1998 (pre-Sano), survival increased to 71%. Late survival statistics (after 120 days) has a strong correlation with the introduction of bilateral cavopulmonary anastomosis procedures (Glenn or hemi-Fontan). The impact on late survival was demonstrated with an increase in hospital survival from 96.3% for the bilateral cavopulmonary anastomosis in 1984 to 1988 to 100% in 1995 to 1998. For the Fontan procedure, the survival rates increased from 76% to 100% during this same timeframe.


FIGURE 17-27  Graph depicting improved survival for the Norwood procedure from 1985 through 1998. The improved survival reflects the evolution of the operation. In particular, a cavopulmonary anastomosis is introduced as an interim operative procedure.



With this evolution, the hospital discharge rate after the initial palliative procedure has continued to improve, with survival rates of 90% or higher reported in some series ( Ghanayem et al., 2003 ). Home surveillance programs with strict attention to oxygen saturations with pulse oximetry and weight gain in infants following the HLHS first-stage repair are improving outcomes. Infants should achieve a minimum weight gain of 20 g during the course of 3 days, and home pulse oximetry monitoring to ensure saturations remain greater than 70% is advocated. This surveillance program has improved the survival of the interstage mortality between the Norwood operation and the subsequent cavopulmonary anastomosis ( Ghanayem et al., 2003 ).

Truncus Arteriosus

Truncus arteriosus is a congenital cardiac defect that is characterized by a single great artery arising from the base of the heart with a single semilunar valve ( van Praagh and van Praagh, 1965 ). This single artery gives rise to the systemic, pulmonary, and coronary circulations. The defect is classified by the location of the branch PAs, the presence or absence of a VSD, and the developmental characteristics of the ascending aorta and ductus arteriosus ( Fig. 17-28 ). Type I truncus has a single, short main PA arising from the truncus that divides into left and right PAs, which follow a normal course into the hilum. In type II truncus arteriosus, the branch PAs arise from separate orifices off of the truncal artery. Type III truncus arteriosus is characterized by the right and left PAs arising from opposite lateral walls of the truncal artery. Type IV truncus arteriosus or hemitruncus is characterized by a single branch PA arising from the truncal artery and a second nonconfluent branch PA arising from aortopulmonary collateral vessels or the ductus arteriosus. Type IV truncus is commonly grouped with tetralogy of Fallot or pulmonary atresia. Truncus is also classified as type A (presence of a VSD) and type B (intact ventricular septum).


FIGURE 17-28  Truncus. There are similarities between the Collett and Edwards and the Van Praagh classifications of truncus arteriosus. Type I is the same as A1. Types II and III are grouped as a single type A2 because they are not significantly distinct embryologically or therapeutically. Type A3 denotes unilateral pulmonary artery with collateral supply to the contralateral lung. Type A4 is truncus associated with interrupted aortic arch (13% of all cases of truncus arteriosus).  (From Mavroudis C, Backer CL, editors: Pediatric cardiac surgery, 3rd ed. Stamford, CT, 1999, Appleton & Lange, p 340, Chapter 19, Truncus Arteriosus.)




Patients with truncus arteriosus present with cyanosis, excessive PBF, heart failure, and truncal valve regurgitation. In the preoperative and prebypass period, management is directed at reducing PBF by increasing Paco2, maintaining Fio2 at 0.21% and, if indicated, the use of exogenous CO2 to reduce PBF. In addition, inotropic support may be necessary to augment systemic oxygen delivery by increasing systemic cardiac output (Qs) if PBF is excessive. Measures to reduce PBF may become increasingly important after opening the chest because this further reduces PVR by removing impedance of the chest wall.

Operative repair for truncus arteriosus is usually performed in the neonatal period. The operative procedure includes removing the PAs from the truncal artery, closing the VSD, and placing a valved homograft from the RV to the main PA. Regurgitation of the truncal valve is usually reduced by reducing the flow across the valve. Occasionally, however, severe truncal valve incompetence is present, necessitating truncal valve replacement. Severe regurgitation in the postbypass and postoperative periods is poorly tolerated, especially if moderate to severe myocardial dysfunction is present. Under these circumstances, truncal valve replacement is indicated. Postbypass management usually is directed at controlling PA hypertension and providing inotropic support for right and left ventricular dysfunction. Reactive PA hypertension is common in neonates with preoperative increased PBF, as is found in truncus arteriosus.


If total systemic venous return (SVC and IVC) is directed to the PAs, the arrangement is termed a modified Fontan procedure. In this operation, the IVC blood is directed to the SVC by the use of an intra-atrial tube graft or baffle ( Jonas and Castaneda, 1988 ; Mayer et al., 1992 ) or an extracardiac conduit with or without fenestration ( Stamm et al., 2002 ) ( Fig. 17-29 A, B). This surgical arrangement allows all systemic venous return to enter the pulmonary circulation. The cavopulmonary anastomosis is a “partial Fontan procedure”; that is, the SVC alone is anastomosed to the PAs and IVC flow is allowed to mix with pulmonary venous return. A classic Glenn shunt is a cavopulmonary anastomosis that connects the SVC directly to the right PA and oversews the SVC-RA junction. The left PA is left separated from the right PA and SVC, and therefore systemic venous blood flow from the SVC is directed only to the right lung. The classic Glenn is not used as a staging procedure for single-ventricle patients. The preferred forms of cavopulmonary anastomosis are the bidirectional Glenn shunt and hemi-Fontan. The bidirectional Glenn leaves the right PA and left PA in continuity ( Fig. 17-30 ), and the SVC blood flow is distributed to both the right and left PAs. The SVC is disconnected from the RA. SVC flow enters into the common atrium (physiologic LA), mixes with pulmonary venous blood, and enters the single ventricle. An alternative intermediate-stage procedure is the hemi-Fontan, which anastomoses both PAs to the SVC. The SVC is left in continuity with the RA, and a partial tube graft is sewn in place below the SVC with a dam at the base of the tube graft. This arrangement facilitates the completion of the Fontan because the surgeon needs to open the RA, remove the dam, and complete the anastomosis between the partial tube graft and the IVC.


FIGURE 17-29  (A) The modified Fontan operation using a lateral tunnel approach. The lateral tunnel is created by placing an intra-atrial Gore-Tex tube graft extending from the inferior vena cava to the superior vena cava and sewn to the posterior atrial wall. Here, the lateral tunnel is fenestrated by placing a 4-mm punch hole in the Gore-Tex portion of the graft. This allows approximately 20% of cardiac output to enter the single (systemic) ventricle without having to pass through the lungs, thereby enhancing systemic oxygen delivery. (B)Modified Fontan operation using an extracardiac conduit rather than an intra-atrial lateral tunnel. This approach minimizes exposure of the atrial tissue to high pressure. The extracardiac conduit can be fenestrated but has a lower success rate of maintaining patency due to low flow and interruption of flow caused by atrial contraction.  (With permission from O—Brien P, Boisurt JT: Current management of infants and children with single ventricle anatomy. J Pediatr Nursing 16:338-350, 2001.)





FIGURE 17-30  Bidirectional cavopulmonary anastomosis (bidirectional Glenn) demonstrating the connection of the superior vena cava (SVC) to the right pulmonary artery while maintaining continuity to both the right and left pulmonary arteries. The lower portion of the SVC is disarticulated from the right atrium, and the inferior vena cava blood flow enters the right atrium and mixes with left atrial blood through a widely patent atrial septectomy. RA, right atrium; RPA, right pulmonary artery.  (With permission from O—Brien P, Boisurt JT: Current management of infants and children with single ventricle anatomy. J Pediatr Nursing 16:338-350, 2001.)




The advantage of the cavopulmonary anastomosis operation is maintaining cardiac output even if PBF is reduced. Poor functioning of a bidirectional Glenn in the operating room is characterized by desaturation. Cardiac output is well maintained unless PBF is severely reduced, resulting in inadequate myocardial oxygen delivery. Desaturation must be carefully assessed in the operating room to ensure adequate flow across the SVC and PA anastomosis, and avoiding an elevated PVR restrictive atrial septum, or obstructed pulmonary veins ( Box 17-5 ).

BOX 17-5 

Causes of Low Cardiac Output After Fontan Completion


Systemic Venous Pressure

Common Atrial Pressure

Transpulmonary Gradient





Pulmonary artery stenosis or obstruction

Restrictive intra-atrial septum

Systemic outflow tract obstruction

N to low

Superior vena cava clot or obstruction[*]

Atrioventricular valve regurgitation or stenosis


Systemic aortic valve stenosis or regurgitation


Coarctation of the aorta


Pulmonary venous obstruction





Systemic ventricle dysfunction


Pulmonary hypertension

Note: After the cavopulmonary anastomosis, because inferior vena cava blood flow enters the systemic ventricle without first having to pass through the lungs, any anatomic obstruction through the lungs will present with low oxygen saturation but a preserved cardiac output. (↑, increases; ↓, decreases; N, normal.)



Superior vena cava–pulmonary artery pressure gradient >2 mm Hg.

Modified Fontan Procedure

The modified Fontan procedure can be completed by either a lateral tunnel procedure or an extracardiac conduit. The lateral tunnel procedure creates an intra-atrial tube graft between the SVC and IVC and is sewn so that the back wall of the atrium forms the posterior portion of the tube. Alternatively, an extracardiac conduit can be created. It connects the SVC to the IVC outside of the heart. The advantage of the lateral tunnel approach is that the posterior wall is native atrial tissue and allows the conduit to grow with the child. Also, the lateral tunnel conduit can be easily fenestrated. The extracardiac conduit is a solid tube and therefore does not grow with the patient; it also is difficult to fenestrate (see Fig. 17-29 B). Fenestration requires a direct connection to the atrium. As atrial contraction occurs, it createsperiods of low flow, and the fenestration commonly occludes early in the postoperative period. The main advantage of the extracardiac conduit is that atrial tissue is not exposed to high pressure and there are no suture lines in the atrium; atrial arrhythmias are less likely to occur in the long term. Acutely, there is also a higher incidence of effusions after the extracardiac Fontan procedure. Finally, the lack of a fenestration is generally well tolerated if a bidirectional Glenn or hemi-Fontan procedure preceded the modified Fontan procedure.

In the modified Fontan procedures, cardiac output is preload limited. Blood return to the single ventricle is dependent on maintaining a pressure gradient between the systemic veins, the pulmonary vasculature, and the single ventricle. Increased PVR, elevated PA pressures (>18 mm Hg), distortion of the PAs, obstruction of the SVC-to-PA anastomosis or intraatrial baffle, pulmonary venous obstruction, AVV regurgitation, or stenosis limits venous return to the single ventricle and decreases cardiac output ( Bridges et al., 1990 ). Clinically, these abnormalities and systolic or diastolic single-ventricle dysfunction are manifest as low cardiac output in the Fontan patient. In the operating room, pressure monitoring, including systemic venous pressure, which reflects PA pressure, and left atrial or, more accurately, common atrial pressure are helpful in distinguishing causes for poor cardiac output and assists in optimally managing these patients (see Box 17-5 ). In addition, TEE may demonstrate residual anatomic problems such as flow obstruction through the conduit or baffle, valvular regurgitation, and pulmonary venous obstruction as well as ventricular functional abnormalities. The use of TEE can help target both medical and, if indicated, additional surgical treatment.

Fontan patients with good intraoperative hemodynamics have age-appropriate blood pressure when systemic venous pressure (measured in the SVC proximal to the SVC-PA anastomosis), which reflects PA pressure, ranges from 12 to 15 mm Hg and left atrial or common atrial pressure ranges from 5 to 8 mm Hg. Acutely, physiologic systemic venous pressures may be as high as 20 to 25 mm Hg in patients with persistent elevations in PA pressures. This elevated pressure (which is reflective of the PA pressure) may be tolerated for a short period of time in the postoperative period. However, these elevated pressures should decrease with medical management. When the systemic venous pressure exceeds 15 mm Hg, the common atrial pressure is high (10 to 15 mm Hg) and the transpulmonary gradient (systemic venous pressure-common atrial pressure) is less than 10 mm Hg, then ventricular dysfunction, AVV regurgitation, or ventricular outflow obstruction (aortic valve stenosis or supervalvar stenosis) must be ruled out. When the systemic venous pressure is high, the common atrial pressure is low, and the transpulmonary gradient is also low, then elevated PVR, PA hypertension, baffle obstruction, or previously unrecognized obstruction of the branch PAs or pulmonary veins may be present. As previously mentioned, all of these problems are clinically manifest as low cardiac output.

High systemic venous and low common atrial pressures or a large gradient measured from the physiologic systemic venous catheter proximal to the anastomosis of the SVC and PAs in the Fontan patient may be a physiologic or an anatomic problem. Anatomic problems include baffle obstruction, obstructed pulmonary veins, and PA distortion. Baffle obstruction is easily remedied but must be diagnosed early before significant and irreversible consequences of abnormal hemodynamics occur. Baffle obstruction is best diagnosed by either TEE or measuring pressures in the IVC and SVC. Intraoperative echocardiography may not be able to visualize a narrowing or to appreciate minor pressure gradients within the atrial baffle or the SVC-to-PA anastomosis; this is particularly true if an extracardiac conduit is used. Mean pressure gradients as low as 3 to 4 mm Hg across the extracardiac conduit or intraatrial baffle are significant and suggest a clinically significant stenosis in the systemic venous-to-PA pathway.

Pulmonary venous obstruction can occur after Fontan operation, especially in patients with complex venous anatomy (e.g., heterotaxy syndromes) or in patients with a small LA (HLHS) ( Mayer et al., 1986). This is more common in heterotaxy syndromes where the pulmonary veins enter the atrium at an unusual location and an intra-atrial baffle is used to complete the Fontan procedure. In these cases, most surgeons are performing an extracardiac connection between the SVC and IVC ( Kumar et al., 2003 ). The use of a simple intra-atrial tube graft (lateral tunnel procedure) from the IVC to the SVC in patients with normal pulmonary veins and a single SVC minimizes the risk of pulmonary venous obstruction. If a left-sided SVC is also present, this may be directly anastomosed to the left PA and the left SVC-atrium junction oversewn. Obstructed, abnormal pulmonary veins are poorly tolerated physiologically in the Fontan procedure. In general, attempts are made to treat pulmonary venous obstruction by balloon dilation and/or placement of small stents in the cardiac catheterization laboratory. The results of this approach to treat pulmonary venous obstruction are poor in most cases but worth trying preoperatively.

PA distortion is usually due to a previous shunt procedure. This can be treated by balloon dilation or stenting in the catheterization laboratory before surgery, or the PA may need to be reconstructed in the operating room at the time of the Fontan procedure.

Physiologic problems resulting in high systemic venous and low common atrial pressures are generally due to an elevation in PVR or PA pressure. Fontan patients have a very limited ability to compensate for these changes and have diminished systemic perfusion and low cardiac output. Common treatable causes for increased PVR in the Fontan patient include hypoxia, hypercarbia, acidosis, excessive mean airway pressure or PEEP, and extrapleural compression of the lung due to pleural effusion, hemothorax, or pneumothorax. In the absence of a clearly reversible cause for increased PVR, therapy is directed toward controlling pH, Paco2, Pao2, and alkalinization. Reduced lung volumes should be treated with improving mechanical ventilation. High tidal volume ventilation with relatively short inspiratory times to achieve an arterial CO2 of 33 to 38 mm Hg, along with systemic alkalinization with sodium bicarbonate, is effective in lowering PVR. Positive pressure ventilation, which results in high mean airway pressure or use of high end-expiratory pressure, has a negative impact on PBF. In general, it is important to provide a short inspiratory phase and a prolonged expiratory phase with low mean airway pressure. PBF predominates during exhalation, so an inspiratory-to-expiratory ratio of 1:3 or longer is preferred. PEEP may be used judiciously in Fontan patients to maintain functional residual capacity, as previously discussed. Excessive PEEP, however, is poorly tolerated in Fontan physiology. Jet ventilation is an effective alternative mode of ventilation that achieves alkalization at lower mean airway pressures and significantly improves cardiac output ( Dietrich et al., 1993 ). Milrinone, epinephrine, and inhaled NO are useful therapeutic interventions. In addition, transfusion to hematocrits of 40% to 45% is useful in improving oxygen delivery.

There are three major causes for high systemic venous and high common atrial pressures in the Fontan patient: ventricular dysfunction (primarily diastolic dysfunction), AVV regurgitation, and ventricular outflow tract obstruction. The most troublesome consequence is diastolic dysfunction of the systemic ventricle. Because some preoperative Fontan candidates have a volume-loaded, hypertrophied ventricle, elevated ventricular end-diastolic pressures should be looked for after the Fontan operation ( Nishioka et al., 1981 ; Sanders et al., 1982 ). The institution of the Glenn procedure at between 4 and 8 months and improved management of stage 1 Norwood procedure have significantly reduced the volume load on the single ventricle and contributed to preservation of myocardial function and better tolerance of Fontan physiology. Inotropic agents that improve systolic function such as epinephrine also have lusitropic properties and acutely benefit diastolic function. In the long term, inotropic agents impair ventricular relaxation. Vasodilators and phosphodiesterase type III inhibitors reduce ventricular volume and are beneficial in patients with diastolic dysfunction; however, in the postoperative Fontan patient, cardiac output is dependent on adequate preload and these patients are sensitive to a reduction in filling pressure. Drugs that promote ventricular relaxation or only minimally increase contractility and unload the heart such as nitroprusside, calcium channel blockers (e.g., nicardipine), and phosphodiesterase inhibitors (e.g., milrinone) may be helpful.

AVV regurgitation may be due to either a preexisting abnormal valve or the chronic volume load on the single ventricle in the pre-Fontan period ( Nishioka et al., 1981 ; Sanders et al., 1982 ). In either case, AVV regurgitation is poorly tolerated in the postoperative Fontan patient because of the critical dependence of this physiology on ventricular filling. When valve replacement is combined with a Fontan operation, a higher-than-anticipated postoperative mortality has been observed ( Kirklin and Barratt-Boyes, 1986 ). This may be related to the gradient that is present in an artificial valve and the consequent increase in left atrial pressure incurred. Most commonly, AVVs can be repaired successfully even with a circular annuloplasty to help control regurgitation, although there are some patients in whom the AVV regurgitation improves after the cavopulmonary anastomosis and a valvuloplasty is not justified. Patients with moderate or mild preoperative AVV regurgitation usually experience improvement and do not require a valvuloplasty ( Mahle et al., 2001 ). Afterload reduction, coupled with preload augmentation and a mild increase in inotropy with a phosphodiesterase inhibitor, low-dose dopamine, or dobutamine, may be helpful in these patients.

Ventricular outflow tract obstruction results in a pressure load on a previously volume-loaded heart. This worsens ventricular systolic and diastolic function, increasing the risk of a poor outcome after Fontan. Interventional cardiac catheterization and balloon dilation of supravalvar AS are important components of preparing these patients for Fontan operation. In HLHS, aortic arch obstruction is not an uncommon finding after repair and is known to be an important factor that increases mortality after the Norwood operation. It may require close follow-up and ongoing reinterventions in the catheterization laboratory (Soongswa ng et al., 2001 ).

High common atrial pressure with low systemic venous pressure is not possible in Fontan physiology. Because blood flow through the lungs is passive (i.e., without the benefit of a pulmonary ventricle), reversing the pressure gradient would prevent filling of the single systemic ventricle and no systemic cardiac output would ensue. If readings are obtained in which the common atrial pressure is greater than the systemic venous pressure, this must be a technical monitoring problem. In Glenn physiology, because cardiac output is maintained through the IVC, it is possible to have SVC pressure below that of the common atrium and still maintain cardiac output. These patients are extremely cyanotic. The usual etiology for this rare condition is large aorta-to-pulmonary collaterals that result in reversal of flow from the collaterals through the proximal PAs and into the SVC. Treatments for this rare condition include coiling of the collaterals in the catheter laboratory or takedown of the Glenn and replacement with an aorta-to-pulmonary shunt.

Postoperatively, all patients undergoing Fontan procedures have elevated SVC pressures. Elevated pressures contribute to several complications, including pleural effusions, hepatic and renal dysfunction, ascites, and protein-losing enteropathy ( Kirklin and Barratt-Boyes, 1986 ). High systemic venous and right atrial pressures result in diminished drainage through the thoracic duct and the release of atrial natriuretic factor, which may contribute to effusions. SVC flow may also be impaired and, when coupled with low cardiac output, systemic organ perfusion is significantly reduced. The net result is diminished perfusion pressure to the abdominal viscera, hepatic and renal dysfunction, a significant accumulation of ascites, and, less commonly, a protein-losing enteropathy. A poorly functioning Fontan with high systemic venous pressure may result in severe, acute, and fatal hepatic failure due to high hepatic venous pressure and a diminution in effective hepatic blood flow (see Box 17-5 ).

Staging Operations in Single Ventricle

The Fontan procedure has been applied to an increasing number of patients, many of whom have risk factors that historically have made them poor candidates for Fontan physiology. PA distortion, increased PVR, pulmonary hypertension, AVV regurgitation, diminished ventricular performance, ventricular hypertrophy, complex cardiac anatomy (other than tricuspid atresia), and complex systemic or pulmonary venous connections increase the mortality associated with a Fontan operation from 5% to 10% to rates of 20% to 30% in some series ( Mayer et al., 1986 ; Bridges et al., 1990 ). Although the optimal management of these patients is in evolution, it is clear that continued palliation with an aorta-to-pulmonary shunt is a poor alternative, because ventricular function worsens with prolonged shunt physiology. Patients with tricuspid atresia have less ventricular dilation and hypertrophy after Fontan repair than do those who undergo a second aortopulmonary shunt procedure. Fontan candidates, due to greater distortion of the PA, experience deterioration of ventricular function and elevation of PVR ( Mietus-Snyder et al., 1987 ; Mayer et al., 1992 ).

Staging operations such as the cavopulmonary anastomosis are being advocated as an interim procedure in patients with increased risk after a Fontan procedure ( DeLeon et al., 1983 ; Mazzera et al., 1989 ). Cavopulmonary anastomosis or fenestrations are modifications that allow an atrial-level communication; that is, blood enters the systemic ventricle from the RA without passing through the lungs. The advantage of these procedures over the completed Fontan is that effective PBF is maintained while ventricular volume load is minimized and systemic oxygenation is improved over traditional shunting procedures ( Mietus-Snyder et al., 1987 ). In addition, cardiac output is not limited by high pressure or flow resistance across the pulmonary vascular bed. In the cavopulmonary anastomosis (bidirectional Glenn or hemi-Fontan), the entire IVC flow enters into the physiologic LA, mixes with pulmonary venous blood, and enters the single ventricle (see Fig. 17-30 ). Cardiac output can be augmented because all of the IVC return goes directly to the systemic ventricle and is pumped to the systemic circulation. The bidirectional Glenn and hemi-Fontan have the added advantage of allowing for a technically simple conversion to a completed Fontan ( Bridges et al., 1990 ; Mott et al., 2001 ).

The cavopulmonary anastomosis has had a major role in allowing the application of Fontan physiology to a broader array of patients who were considered poor Fontan candidates in the 1990s. This approach has a low operative mortality and facilitates adaptation to the completed Fontan physiology by limiting the damaging effects of prolonged exposure to shunt physiology ( Mott et al., 2001 ). Although the bidirectional Glenn procedure is a marked improvement over forcing young infants into Fontan physiology, it is not an optimal long-term intervention. The main reason is as children grow, the contribution of venous return through the SVC becomes appreciably less and therefore PBF decreases. Also, prolonged exposure to the physiology of the cavopulmonary anastomosis results in the development of pulmonary arteriovenous malformations (AVMs) in approximately 25% of patients. AVMs result in a progressive increase in cyanosis. The mechanism for the development of AVMs is believed to be that the liver produces inhibitors of angiogenesis that are excluded from the pulmonary circulation, so proangiogenesis factors are left unchecked ( Duncan et al., 2003) .

A fenestrated Fontan is physiologically similar to the bidirectional Glenn procedure. In this arrangement, the Fontan operation is completed using an intra-atrial lateral tunnel, and a 4-mm punch hole is placed in the tube graft connecting the IVC to the SVC. The punch hole produces a right-to-left shunt that allows approximately 20% of venous return to cross directly from the RA to the LA, thereby increasing cardiac output with minimal reductions in systemic saturation ( Bridges et al., 1990 ). This technique provides the added advantage of not requiring a second surgical procedure, because these small punch holes can be closed in the cardiac catheterization laboratory using catheter-positioned ASD closure devices, or, in many cases, the fenestration closes on its own ( Lloyd et al., 1998 ).

The main physiologic advantage of a bidirectional Glenn procedure over a fenestrated Fontan is lower IVC pressure and a larger augmentation in cardiac output. In the fenestrated Fontan, higher IVC pressure results in reduced hepatic, renal, and mesenteric perfusion, and the approximate 20% increase in cardiac output afforded by the fenestration may not be adequate.

The postoperative management of these patients is similar to that of the completed Fontan patient. Systemic venous and common atrial pressures are monitored and similarly maintained. Systemic saturation is lower because of the right-to-left atrial shunt and is generally between 80% and 90%. A lower systemic saturation is generally well tolerated because of the increase in oxygen delivery and cardiac output (right-to-left shunt). Systemic saturations in the mid to high 90 percent values generally represent closure of the fenestration.

The bidirectional Glenn is also a useful “bailout” operation. If the child does not tolerate Fontan physiology in the early postbypass period and there is no anatomically correctable problem, the patient can be converted to a bidirectional Glenn if other support options are not demeaned valuable.

Complications immediately after the Glenn procedure or fenestrated Fontan are lower than after the completed Fontan. The incidence of pleural effusions, ascites, atrial dysrhythmias, and renal, hepatic, and mesenteric perfusion problems is diminished. Very low saturation values (50% to 60%) may be seen in some of these patients. If oxygen saturations remain very low, early cardiac catheterization is recommended, because extensive venous collaterals can be coil occluded in the cardiac catheterization laboratory ( Bridges et al., 1990 ). If collaterals are not present, and there is no obstruction across the anastomosis or distortion of the pulmonary artery, then surgical options include conversion from a fenestrated Fontan to a bidirectional Glenn or placement of an aorta-to-pulmonary shunt to improve PBF.

Rhythm disturbances after the Fontan operation are common. The absence of sinus rhythm is a risk factor for Fontan operation, but evidence suggests that sinus rhythm is not an absolute requirement for successful outcome after the Fontan procedure ( Balaji et al., 1991 ). Atrial pacing can improve cardiac output and systemic blood pressure, especially when junctional rhythm is present in the early postrepair period. Atrial pacing lowers left atrial pressure and provides an atrial kick that supplements systemic stroke volume ( Alboliras et al., 1985 ). More significant rhythm disturbances such as atrial flutter and junctional ectopic tachycardia increase the risk of mortality in the early postrepair period. In a study by Balaji and others (1991) , the presence of atrial tachyarrhythmias (atrial flutter, supraventricular atrial ectopic tachycardia, and junctional ectopic tachycardia) carried a very high mortality rate in the early postoperative period. By using a total cavopulmonary connection rather than an atriopulmonary connection, atrial tachyarrhythmias were less common and were more easily controlled with antiarrhythmic therapy, overdrive pacing, or DC cardioversion ( Balaji et al., 1991 ). This finding suggests that a major contributor to postoperative arrhythmias in the Fontan patient is exposure of native atrial tissue to high pressure.

In a comparison of the hemi-Fontan with the Glenn procedure, the early postoperative incidence of sinus node dysfunction was higher in the hemi-Fontan patients. This is not surprising because the Glenn procedure is an extracardiac anastomosis ( Cohen et al., 2000 ). High pressure explains the greater likelihood of atrial arrhythmias and why medical control is so difficult. With the bidirectional Glenn procedure and extracardiac Fontan connection, native atrial tissue is not exposed to elevated pressures. Atrial tissue is primarily exposed to common atrial pressure, which is substantially lower (generally 5 to 8 mm Hg) ( Balaji et al., 1991 ; Kumar et al., 2003 ). The benefits from the extracardiac Fontan in terms of reducing cardiac dysrhythmias have led to the concept of converting lateral tunnel Fontan to an extracardiac Fontan and cryoablation surgery as an alternative to transplantation in those patients with a failing Fontan due to dysrhythmias. ( Weinstein et al., 2003 ). Results have been encouraging. The presence of AVV regurgitation may also contribute as a risk factor for increased postoperative arrhythmias, again invoking high atrial pressures as a causative factor.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


Modern pediatric cardiac anesthesia must include the principles and practice of effective anticoagulation, hemostasis, and blood conservation. Bleeding after CPB remains a significant problem in pediatric cardiac surgery ( Manno et al., 1991 ). Continuing blood loss after bypass, requiring blood component replacement, is associated with hemodynamic compromise as well as morbidity from multiple donor exposure. In pediatric patients, restoration of hemostasis has proved to be difficult, and diagnosis of the problem and treatment are marginally effective.

Neonates, infants, and children undergoing cardiac surgery with CPB have a higher rate of postoperative bleeding than do older patients ( Manno et al., 1991 ). There are a number of causative factors. There is disproportionate exposure of blood to the area of nonendothelialized extracorporeal circuit. This exposure to the nonendothelialized circuit produces a heightened inflammation-like response and activates platelets. In addition, the CPB circuit and prime volume produce a dilutional coagulopathy ( Kirklin et al., 1983 ). The inflammatory response is inversely related to patient age; the younger the patient, the more pronounced is the response ( Greeley et al., 1988 ). Inflammation is known to be intimately related to activation of the coagulation system.

Complement and platelet activation is linked to the activation of other protein systems in the blood, such as the fibrinolytic system. This generalized protein system activation plays a major role during surgery and results in impaired hemostasis and an increased bleeding tendency. The most widely used monitor of anticoagulation during CPB is the activated coagulation time (ACT). The ACT does not correlate with circulating heparin levels ( Culliford et al., 1981 ; Codispoti et al., 2001 ), particularly during hemodilution and deep hypothermia ( Martindale et al., 1996 ). Also, the type of operations performed in neonates and infants usually involves more extensive reconstruction and suture lines, creating more opportunities for surgical bleeding ( Dietrich et al., 1993 ; Martindale et al., 1996 ).

The immature coagulation system in neonates and young infants may also contribute to impaired hemostasis ( Andrew et al., 1987 ). Although procoagulant and factor levels may be reduced in the young patient with cyanotic CHD due to immature or impaired hepatosynthesis ( Colon-Otero et al., 1987 ), functional bleeding tendencies are not usually present before surgery. However, patients with cyanotic heart disease demonstrate an increased bleeding tendency after CPB for reasons that are not totally clear ( Henriksson et al., 1979 ).

CPB is a significant thrombogenic stimulus that requires anticoagulation with heparin before initiation. Heparin is usually administered empirically based on patient weight, and its effect is followed by ACT monitoring. Because heparin effect is primarily due to coupling with anti-thrombin III (AT III) and because there are age-related differences in the level of circulating procoagulants and this inhibitor, variability of heparin dosing and effect has been a concern. High heparin sensitivity is observed in the first week of life and then decreases progressively until about 3 years of age, when values approach those observed in adults ( Vieira et al., 1991 ). These findings are consistent with the variable quantities of circulating levels of procoagulants and inhibitors, especially prothrombin and AT III ( Kern et al., 1992b ). Heparin administration must also include a consideration of the quantity and composition of the priming volume for CPB, especially if FFP is added and the current trend continues of using smaller CPB circuits and prime volumes for neonates and infants. Recommendations for heparin are a dose of 300 to 400 U/kg plus an additional dose of 1 to 3 U/mL of prime and then maintaining the ACT at greater than 400 seconds. Deep hypothermia contributes to prolongation of the ACT, provides a false sense of adequate anticoagulation, and has led to an increased interest in using higher heparin doses, closer to 400 U/kg rather than the recommended 300 to 400 U/kg (Despotis et al., 1995).

Heparin is neutralized with protamine, and the dose of is calculated based on the dose of heparin given or on body weight. General requirements for protamine are 1 mg/100 U heparin administered or 5 to 7 mg/kg, and then the ACT is checked. It is unusual for a dose greater than 10 mg/kg to be required. However, the dose may have to be individualized based on the degree of observed surgical bleeding. This increased protamine requirement in young patients is indicative of the higher circulating heparin levels after CPB ( Horkay et al., 1992 ). Delayed hepatic clearance of heparin due to organ immaturity and the predominant use of deep hypothermic CPB in this age group decrease metabolism and excretion of heparin. Interpatient variability mandates some form of individual assessment to guide drug dose, to prevent excess protamine administration ( Jobes and Nicolson, 1988 ; Jobes et al., 1992 ).

As discussed, neonates and young infants with CHD have low circulating levels of procoagulants and inhibitors before surgery ( Kern et al., 1992b ). The thrombogenic and dilutional effects of CPB further contribute to hemostatic abnormalities after CPB. Formed blood elements such as leukocytes and platelets may be activated, and procoagulants diluted by CPB. DHCA causes greater fibrinolytic activity. The lower the temperature, the higher the degree of activation of the inflammatory and coagulation cascade. The causes of bleeding after CPB tend to be multifactorial. Injudicious use of blood products to separately correct individual coagulation abnormalities can further exacerbate the dilution of existing procoagulants. Severe reductions in hematocrit impair oxygen delivery and may contribute to neurologic pathology, especially after deep hypothermia.

Bleeding after CPB is not an unusual occurrence. The surgeon should first attempt to identify any obvious source of surgical bleeding at the sites of repair dissection and cannulation. Next, adequate protamine reversal of heparin is assessed by measuring the ACT. In general, standard coagulation tests show a prolongation of the bleeding time, partial thromboplastin time and prothrombin time, hypofibrinogenemia, and dilution of other procoagulants ( Table 17-16 ). The most common reason for persistent bleeding is platelet dysfunction ( Harker, 1986 ; Woodman and Harker, 1990 ; Tempe and Virmani, 2002) .

TABLE 17-16   -- Summary of coagulation tests during cardiopulmonary bypass (CPB) in infants



1 min on CPB

Cold CPB

Warm CPB


Intensive Care Unit


200 ± 59[*]

92 ± 18

94 ± 21

107 ± 24[*]

142 ± 28[*]

183 ± 33

Factor 2%

56 ± 13

30 ± 7

32 ± 8

33 ± 7[§]

48 ± 10

60 ± 33

Factor 5%

68 ± 20[*]

15 ± 4

17 ± 5

22 ± 7[*]

39 ± 11[*]

46 ± 10

Factor 7%

54 ± 11[†]

26 ± 5

27 ± 6

28 ± 5[†]

41 ± 8[†]

53 ± 16

Factor 8%

48 ± 20[‡]




32 ± 15[‡]

72 ± 45

Factor 9%

31 ± 13

20 ± 8

23 ± 5

31 ± 4

31 ± 9

40 ± 12

Factor 10%

52 ± 10

31 ± 7

31 ± 7

34 ± 8[§]

46 ± 10

47 ± 16

Platelets (k/mm3)

225 ± 54[*]

65 ± 17

45 ± 8

93 ± 28[*]

120 ± 29[*]



49 ± 22

30 ± 12

29 ± 15

32 ± 13[§]

57 ± 28

68 ± 23

Heparin (units)

0.02 ± 0.03

0.41 ± 0.08

0.42 ± 0.08

0.42 ± 0.08

0.04 ± 0.04

0.07 ± 0.06

ACT (sec)

168 ± 20




151 ± 31



P < .0001.

P < .002.

P < .005.


P < .05.


Blood transfusion practices in the context of pediatric cardiac surgery vary widely, and no one approach has received broad acceptance ( Kwiatkowski and Manno, 1999 ). In a CPB coagulation study in 494 pediatric patients undergoing cardiac surgery, the most reliable indicator of excessive bleeding and requirement for blood product transfusion after bypass was a platelet count of 108,000/μL or less (Williams et al., 1999 ). Low platelet counts in conjunction with active bleeding should be treated, first, with the administration of platelets. After platelets have been administered, if bleeding is still present, reassessment and repeat platelet infusion or the administration of cryoprecipitate in infants weighing less than 8 kg or of FFP in older children are considered. In a study of 75 pediatric patients undergoing CPB and cardiac surgery, children who weighed less than 8 kg had fewer postoperative requirements for transfusion products, if platelets were given, followed by the administration of cryoprecipitate after separation from CPB. In those children administered FFP after a platelet transfusion, greater postoperative bleeding was observed ( Miller et al., 1997 ). The excess bleeding after FFP transfusion was believed to be due to dilution of platelets and red blood cells from the higher volume of FFP compared with cryoprecipitate. Under most circumstances, meticulous surgical technique, appropriate administration of protamine, adequate patient temperature, and platelet infusion correct excessive bleeding. In neonates, excessive bleeding, as well as the escalating dilutional effects of selective component therapy on the remaining procoagulants in small patients, makes the treatment of bleeding a difficult one. The use of fresh whole blood may be warranted under these circumstances. The administration of fresh whole blood (<48 hours old) after CPB can meet all of the hematologic requirements with minimum donor exposure. The efficacy of whole blood in restoring hemostasis and reducing blood loss after CPB has been demonstrated in patients younger than 2 years who are undergoing complex surgical repairs ( Manno et al., 1991 ; Kwiatkowski and Manno, 1999 ). Most centers cannot obtain fresh whole blood due to screening requirements for blood-borne pathogens.

Many attempts have been made to reduce bleeding after CPB through pharmacologic interventions. Antifibrinolytics, ε-aminocaproic acid (EACA), and tranexamic acid ( Horrow et al., 1990 ) have been used with fair success. One double-blind study in 41 repeat-surgery pediatric patients reported 24% less blood loss after cardiac surgery when tranexamic acid 100 mg/kg loading dose, followed by 10 mg/kg per hour, was administered intravenously ( Reid et al., 1997 ). The most impressive results have been demonstrated with the use of aprotinin, a protease inhibitor ( Royston et al., 1987 ). Aprotinin has antifibrinolytic properties in low concentrations and acts as a kallikrein inhibitor at higher levels. CPB causes increased kallikrein through contact activation, promoting thrombus and fibrin generation, which promotes fibrinolysis. The inhibition of kallikrein results in an inhibition of the contact phase of coagulation, and the inhibition of fibrinolysis reduces bleeding. Reduced thrombin generation leads to diminished platelet stimulation. Better preserved platelet function has been described for patients with aprotinin ( Royston et al., 1987 ). Not surprisingly, then, aprotinin significantly reduces intraoperative and postoperative blood loss (Dietrich et al., 1989, 1990, 1991, 1992 [103] [104] [105] [106]). Aprotinin use during pediatric cardiac surgery attenuates fibrinolytic activation in a dose-dependent fashion, attenuating the hemostatic activation during CPB with less plasmin formation and, because of inhibition of contact activation, less thrombin generation, thereby reducing fibrin split product formation (Dietrich et al., 1993 ). Higher doses of aprotinin reduce thrombin-AT III complex and F1/F2 fragments, supporting the hypothesis of suppression of clotting activation with higher aprotinin doses and plasma concentrations ( Dietrich et al., 1993 ; Mossinger et al., 2003 ). Aprotinin use is recommended for reoperations and first operations in neonates where bleeding via extensive suture lines is anticipated. After a test dose, the recommended loading dose is 30,000 to 50,000 U/kg before sternotomy and then a CPB prime dose of 30,000 U/kg and continuous infusion of 7,000 to 10,000 U/kg per hour during surgery, aiming for a therapeutic plasma target range of 200 KIU/mL ( Mossinger et al., 2003 ). Higher doses are generally used in neonates and infants due to the marked hemodilution of CPB in younger infants. Aprotinin is currently not approved by the Food and Drug Administration for use in children. However, aprotinin is being administered in many centers that care for pediatric cardiac surgical patients, including its use in DHCA (Tweddell et al., 2002). Anaphylaxis on reexposure of the drug is a risk (2.7% risk; 5% risk if reexposed within 6 months; 0.95% risk if reexposed after 6 months) (Laxenaire et al., 2000). Hypersensitivity reactions with aprotinin do occur; the drug should not be used in patients at risk for thromboembolism or in those with renal insufficiency.

The treatment of postbypass coagulopathy should be dictated more by clinical bleeding than by laboratory values alone. Isolated coagulation abnormalities are often present in the uncomplicated postoperative cardiac patient (see Table 17-15 ). Usually, these coagulation abnormalities self-correct themselves during the first postoperative day and are not associated with excessive bleeding. Routine correction of these abnormalities with infusion of blood products is not warranted. Blood products should not be administered unless there is clinical evidence of bleeding or a specific defect has been identified and specific component therapy used. Routine use of blood products for volume replacement is also to be avoided; Plasmanate, lactated Ringer's, or saline solution can be satisfactorily administered at a reduced cost without the hazards associated with transfusion. Another method of minimizing homologous blood transfusions is the adoption of a preoperative autologous blood donation program. Although autologous blood donation is not commonly done in children, it has been found to avoid the need for homologous blood transfusion in 94% of situations in a group of 80 older pediatric patients undergoing corrective cardiac surgery with CPB (Masuda et al., 1995).

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


Early corrective repair in infancy with less palliation has significantly reduced the number of noncorrective closed heart operations. Noncorrective closed heart operations include PA banding, extracardiac shunts such as the modified BT shunt, and atrial septectomy. Corrective closed heart procedures include PDA ligation and repair of coarctation of the aorta. All of these procedures are performed without CPB. Venous access and intra-arterial monitoring are important in evaluating and supporting these patients, and the pulse oximeter remains an invaluable monitor during intraoperative management.


PDA ligation is performed through a left thoracotomy. The physiologic management is that of a simple shunt. Patients with a large PDA and low PVR generally present with excessive PBF and congestive heart failure. Neonates and premature infants also run the risk of having a large diastolic runoff to the PA, potentially impairing coronary perfusion. Patients range from an asymptomatic, healthy young child to the sick, ventilator-dependent premature infant on inotropic support. The former patient allows for a wide variety of anesthetic techniques and extubation in the operating room. The latter patient requires a carefully controlled anesthetic and fluid management plan. In general, a trial of medical management with indomethacin and fluid restriction is attempted in the premature infant before surgical correction. In the premature infant, transport to the operating room can be especially difficult and tedious, requiring great vigilance to avoid extubation, excessive patient cooling, and venous access disruption. For these reasons, many centers perform PDA ligation safely and cost-effectively in the neonatal intensive care unit ( Mortier et al., 1996 ).

Intraoperatively, retractors may interfere with cardiac filling and ventilatory management so that hypotension, hypoxemia, and hypercarbia may occur. Complications include inadvertent ligation of the left PA or descending aorta, recurrent laryngeal nerve damage, and excessive bleeding due to inadvertent PDA disruption. After ductal ligation in premature infants, worsening pulmonary compliance and increased ventilatory support may occur, and an increase in left ventricular afterload should be anticipated. This is problematic if left ventricular dysfunction existed preoperatively.


Coarctation accounts for about 8% of all CHD and ranks fourth or fifth among congenital cardiac anomalies. The incidence of coarctation has been reported to be 1:10,000, higher in males than in females (1.7:1). Extracardiac malformations are seen in about 7% of patients, the most frequent being Turner's syndrome (gonadal dysgenesis, webbed neck, and cubitus valgus), hypospadias, clubfoot, and ocular defects. Coarctation of the aorta is a narrowing of the descending aorta, almost always at the junction of the ductus arteriosus into the aortic arch. The posterior lateral shelf forming the localized narrowing is almost always directly across from the ductus. The term juxtaductal is used to describe the position of the coarctation.

Obstruction to left ventricular outflow results from coarctation, and this may range from severe obstruction with compromised systemic perfusion to mild upper extremity hypertension as the only manifestation. In neonates with a large posterior lateral shelf and a quickly developing obstruction, there is a sudden increase in left ventricular afterload that results in left ventricular failure. These patients usually become symptomatic in the first week of life. Left-to-right shunting occurs at the atrial level because of an enlarged and stretched foramen ovale. In the young infant with severe coarctation, systemic perfusion is dependent on right-to-left shunting across the PDA. In these circumstances, left ventricular dysfunction is very common and PGE1 is often necessary to maintain systemic perfusion. When the ductus closes gradually and aortic obstruction develops over several months, left ventricular failure is less likely to occur. In older children, the presenting symptom of coarctation may be hypertension, intermittent claudication, cerebrovascular accident, or bacterial endocarditis involving the bicuspid aortic valve (a common associated finding).

In general, a larger peripheral intravenous catheter and an indwelling arterial catheter in the right upper extremity are recommended for intraoperative and postoperative management. In patients with left ventricular dysfunction, a central venous catheter may be placed intraoperatively for pressure monitoring and inotropic support. The surgical approach is through a left thoracotomy, where the aorta is cross-clamped and the coarctation is repaired with one of three techniques: an on-lay prosthetic patch, a subclavian artery flap or resection of the coarctation and an end-to-end anastomosis. The latter is the favored approach. Intravascular volume loading with 10 to 20 mL/kg of crystalloid is administered just before removal of the clamp. The anesthetic concentration is decreased, and additional volume support is given until the blood pressure rises. Additional anesthetic concerns include spinal cord damage, rebound hypertension, and pain control. Spinal cord injury occurs in less than 0.5% of patients. Associated lesions such as VSD and aberrant left subclavian artery, as well as intraoperative fever, have been associated with an increased risk of paralysis. Other associated factors include intraoperative hypertension and reoperation for coarctation, bicuspid aortic valve, subvalvar AS, and other cardiovascular disease ( Attenhofer et al., 2002) . The use of mild hypothermia has been advocated to lower oxygen consumption during aortic cross-clamping and to reduce the incidence of paralysis. In some centers, ice application in the surgical field is used. In older children, SSEP monitoring is frequently performed (see Chapter 9 , Pediatric Anesthesia Equipment and Monitoring).

The other significant anesthetic concern is postrepair rebound hypertension. Hypertension after surgery is probably related to heightened baroreceptor reactivity and increased activation of the renin-angiotensin pathway. Perioperative hypertension is best treated with β-blockade (esmolol) or α/β-blockade (labetalol) or calcium channel blockade (nicardipine). Propranolol is useful in older patients, but it may cause severe bradycardia in infants and young children. Sodium nitroprusside does not appear to be useful except in very high doses; captopril or another angiotensin inhibitor is effective in controlling the hypertension. Most patients with coarctation are treated with β-blockers or captopril in the preoperative period.


The placement of extracardiac shunts without CPB is managed by balancing pulmonary and systemic blood flows. Consideration should be given that the newly established shunt, depending on size, should offer greater resistance to PBF rather than the PDA. In fact, the reason for placing a shunt in many cases is to replace a nonrestrictive ductus. Under these circumstances, attempts at manipulating PVR are less helpful. Central shunts are usually performed through a median sternotomy, and BT (subclavian artery-to-PA) shunts may be performed through a left or right thoracotomy. In patients in whom PBF is critically low, partial cross-clamping of the PA required for the distal anastomosis causes further reduction of PBF and desaturation. Pulse oximetry, as well as careful application of the cross-clamp to avoid PA distortion, is helpful in maintaining PBF. Rarely, severe desaturation and bradycardia occur with cross-clamping and necessitate the use of CPB to complete the procedure. Intraoperative complications include bleeding and severe systemic oxygen desaturation. During chest closure, shunt kinking or clotting can occur. In an attempt to avoid clotting of the shunt before cross-clamping, 50 to 100 U/kg heparin is administered. Pulmonary edema may result in the early postoperative period because of increased PBF from the surgically created shunt. Administering low FIO2, increasing the Paco2, and manipulating PEEP may be helpful maneuvers to decrease PBF until the pulmonary circulation can adjust. Under such circumstances, early extubation is inadvisable.


PA banding is used to restrict PBF in infants who are deemed anatomically or physiologically uncorrectable at this time and have excessive PBF from a native PA. Most patients who require PA banding have complex anatomy and include heterotaxy syndromes and lesions that are RV dominant with normal-sized PAs and LV hypoplasia. Examples of a patient who may require PA banding include those with double-outlet RV with hypoplastic LV and patients with unbalanced AVSD and hypoplastic LV. These patients are generally in heart failure with reduced systemic perfusion and excessive PBF. The surgeon places a restrictive band across the main PA to reduce flow. Band placement is very imprecise and requires careful assistance from the anesthesia team to accomplish successfully. Many approaches have been suggested; one approach is to administer the patient a 21% inspired oxygen concentration and maintain the Paco2 at 40 mm Hg to simulate the postoperative state. Direct observation of an elevation of systemic blood pressure by 10 mm Hg and a reduction in PA pressure to one-half to two-thirds systemic suggest an appropriate band size. Following the end-tidal Pco2 is helpful in ensuring that the band has had some impact on reducing PBF.


A Blalock-Hanlon atrial septectomy is an uncommon procedure for enlarging an intra-atrial connection. This procedure is done by occluding caval flow and creating an intra-atrial communication through the atrial septum. Balloon atrial septostomies (Rashkind procedure) and blade septectomies performed in the cardiac catheterization laboratory have replaced surgical intervention, except when LA size is very small or the atrial septum is thickened.


Vascular rings are anomalies of the great vessels and their branches that result in compression or obstruction of the esopha gus and trachea ( Kussman et al., 2004 ). Vascular rings may be complete or partial and are due to persistence of the normally obliterated component of the embryonic aortic arches ( Fig. 17-31 ). The most common form of a complete vascular ring is a double aortic arch. In double aortic arch, the ascending aorta arises normally, but as it leaves the pericardium, it divides into a right and left arch, which join posteriorly to form the descending aorta. This complete vascular ring entraps both the esophagus and the trachea, creating proximal airway compression and dysphagia. These patients frequently present with inspiratory and expiratory stridor, occasional wheezing, swallowing difficulties, and, in severe forms, apnea and cyanosis. Although not commonly associated with other cardiac malformations, tetralogy of Fallot and TGA may be associated malformations.


FIGURE 17-31  Anomalies of the aortic arch. (A) Double aortic arch, anterior view; note that posterior arch is larger. (B) Double aortic arch, posterior view; note that posterior arch is smaller. (C) Right aortic arch (going behind trachea and esophagus) with left ligamentum arteriosum or ductus arteriosus. (D) Aberrant right subclavian artery.(E) Schematic representation of a rare form of vascular ring comprised of a double aortic arch with an atretic right arch and a right ligamentum arteriosum. (F) Schematic representation of a vascular ring comprised of a left aortic arch, aberrant right subclavian artery, and right ligamentum arteriosum.  (Reproduced with permission from Castaneda AR, Jonas RA, Mayer JE Jr, et al.: Cardiac surgery of the neonate and infant. Philadelphia, 1994, WB Saunders.)


Various types of right aortic arches may present with a complete vascular ring due to anomalous origin of the right subclavian artery from the right arch or persistence of the ligamentum arteriosus (closed fibrous portion of the ductus arteriosus). The ring formed by the anomalous vascular origin may compress the esophagus or the trachea. Similarly, a left aortic arch with an aberrant right subclavian artery passing posterior to the esophagus and right-sided ligamentum arteriosum may cause a partial vascular ring and result in dysphagia, although symptoms are uncommon.

Another form of vascular entrapment is caused by anomalous origin of the left PA, known as a vascular sling. In this malformation, the left PA arises from the right PA and courses behind the trachea at the carina but in front of the esophagus, causing obstruction of the distal intrathoracic portion of the trachea. Anomalous innominate artery, although technically not a vascular ring, can present clinically as similar to vascular rings. An anomalous innominate artery arises more distally and leftward from the aortic arch and as a result compresses the trachea anteriorly. Localized tracheomalacia can occur, and patients frequently present with stridor and apnea. Vascular slings present with obstruction of the intrathoracic large airways and result in symptoms on exhalation that include wheezing and expiratory stridor. Anesthetic management is predicated on managing patients with a narrowed tracheal bronchial airway. These patients benefit from PEEP during mechanical or mask-bag ventilation in order to distend the tracheobronchial airway. In addition, ventilatory benefits may be attained by applying mild cricoid pressure to prevent excessive gas flow into the esophagus and stomach. This maneuver prevents gastric distension, and promotes gas flow into the narrowed airway. Intubation is usually not difficult, because the problem is extrinsic compression of the airway by a vascular structure and, once positioned, the endotracheal tube effectively stents open the airway. In patients with severe airway obstruction, bag-mask ventilation may be difficult and may worsen with administration of neuromuscular blockade. An inhalation induction with spontaneous ventilation is the preferred induction in this subgroup of patients.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


After repair of congenital cardiac defects, temporary pacing wires are generally placed to increase heart rate or provide atrioventricular synchrony in the early postbypass and postoperative periods. Temporary pacing augments cardiac output without the complications associated with pharmacologic therapy. β-Adrenergic agonists such as isoproterenol can increase heart rate; however, myocardial oxygen consumption is also increased. If the augmentation in heart rate does not effectively support oxygen delivery, myocardial ischemia may ensue. In children, increasing the heart rate to achieve age-appropriate levels or higher, or promoting atrioventricular synchrony through atrial or atrioventricular sequential pacing, may significantly improve cardiac output and oxygen delivery to tissues and is commonly used intraoperatively.

Temporary pacing may be either atrial, atrioventricular sequential, or ventricular pacing. Atrial pacing is preferred because it maintains a variable timing interval between atrial and ventricular contractions. Atrial pacing maximizes atrioventricular synchrony and augments the atrial contraction component of ventricular filling. If atrioventricular conduction is abnormal, atrioventricular sequential pacing may be used in either a synchronous (demand) mode or an asynchronous mode. In both modes, there is a fixed preset atrioventricular time interval that may or may not allow for maximal atrial contribution to cardiac output. A typical time interval used in infants is 150 milliseconds. Asynchronous pacing carries the potential for inducing ventricular tachyarrhythmias. However, this type of pacing may be the preferred method if the patient's intrinsic rhythm does not provide adequate cardiac output in a synchronized mode or competes with synchronized atrioventricular sequential pacing.

Distinguishing atrial from junctional dysrhythmias in the postoperative period is complicated by the intrinsically rapid heart rates of infants and further complicated by the use of rate-enhancing inotropic agents. In neonates with rapid heart rates (≥200 beats per minute), the P wave is typically buried in the T wave. Accurate diagnosis from a bedside monitor or 12-lead electrocardiogram is difficult. One of the most effective means of bedside electrocardiographic analysis is to use an atrial electrogram ( Waldo, 1980 ). This is obtained by attaching two atrial pacing wires to the right and left arm leads of a standard electrocardiograph machine, while maintaining normal leg lead recordings. This configuration provides maximal accentuation of the P wave.

The most benign method of treating atrial tachycardias in the postbypass period is overdrive pacing, which avoids the negative inotropic effects of antiarrhythmic drugs. Several pacing techniques have been used to interrupt atrial tachycardias, including brief bursts of pacing above the atrial rate (usually 200 to 250 beats per minute), referred to asburst suppression; pacing 10 to 15 beats per minute above the supraventricular tachycardia rate and, once a rapid atrial rate is captured, gradual reduction in the pacing rate toward normal; and introducing a premature beat, which allows a sinus beat to occur during the refractory period of the reentry loop (this approach is generally the least efficacious). If pacing is unsuccessful and the patient is hemodynamically stable, drug therapy with digoxin, procainamide, adenosine, amiodarone, or verapamil may be useful. In patients with unstable hemodynamics, synchronized cardioversion is the procedure of choice.

The placement of permanent pacemakers in children requires general anesthesia. The preferred method of placement is through a transvenous approach. Patients with slow ventricular escape rates or poorly functioning permanent or temporary wires should have an external pacing system applied before induction. With the broader availability of external pacing systems within defibrillators, this should be available for all pacemaker cases. Isoproterenol infusions have historically been made available as a method to increase the ventricular rate by 10 to 20 beats per minute using low to moderate doses.

Permanent pacing systems have become increasingly sophisticated, including rate-adaptive pacemakers, which adjust their rate to the needs of the patient. A five-position pacemaker nomenclature has been developed to provide a precise understanding of pacemaker function. The first position describes the chamber or chambers being paced (A = atrial, V = ventricular, or D = dual [atrial and ventricular]) (Table 17-17 ). The second position defines the chamber being sensed by the pacemaker and is designated A, V, D, or 0 (none). The V designation means the R waves generated by the ventricle are sensed, and the A designation implies that the P waves generated by the atrium are sensed. The third position describes the response of the pacemaker after sensing the R or P wave and the designations are I (inhibit), T (trigger), D (dual), or 0 (none). An I designation implies the pacing circuit is inhibited. The fourth position describes the programmability of the pacemaker; designations are P (rate adaptive and output are programmable), M (multiprogrammable), R (rate adaptive), and 0 (none). Because almost all pacemakers are multiprogrammable, this designation is usually omitted. The fifth position designates arrhythmia control. The nomenclature includes P (pacing), S (shock), D (dual pacing and shocking), and 0 (none). This designation is for automatic implantable cardiac defibrillators (AICDs), which are not commonly used in pediatric patients. AICDs are implanted in children with prolonged QT interval or a history of ventricular tachycardia.

TABLE 17-17   -- Pacemaker nomenclature

Pacers: Five-Letter Code

First letter: chamber paced

A = atrium, V = ventricle, D = dual (A + V)

Second letter: chamber sensed

A = atrium, V = ventricle, D = dual (A + V), O = none

Third letter: response after sensing

I = pacing inhibited, T = pacing triggered, D = dual (I + T), O = none

Fourth letter: programmability

P = rate and output, M = multiprogrammable, C = communicating


R = rate adaptive, O = none

Fifth letter: arrhythmia control

P = pacing, S = shock, D = dual (P + S), O = none


AVI = atrial paced, ventricle sensed; pacing inhibited if beat sensed.

VVIR = demand ventricular pacing with physiologic response to exercise.

DDD = both atrium and ventricle paced, both sensed; pacing triggered in each chamber if beat not sensed.




Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


Advances in interventional cardiac catheterization techniques are significantly changing the operative and nonoperative approaches to the patient with CHD. Intracardiac echocardiography, technological advancement in interventional devices (stents, closure devices, valves), and three-dimensional cardiac magnetic resonance imaging are greatly increasing the scope and complexity of nonoperative interventional cardiac catheterization and diagnostic procedures ( Fogel, 2005 ; Thanopoulos et al., 2003 ; Zanchetta et al., 2003 ). As a result, less cardiac surgery and CPB are required for the safe closure of small secundum ASDs, VSDs, and PDAs. Stenotic aortic and pulmonic valves and recurrent aortic coarctations can be dilated in the catheterization laboratory, avoiding surgical intervention as well (Lock et al., 1989 ; Mullins, 1989 ; Hellenbrand et al., 1990 ; Hickey et al., 1992 ). These techniques shorten hospital stays and are particularly beneficial to patients with recurrent coarctation and muscular or apical VSDs, who are at a higher risk for surgical intervention. Many other patients with complex cardiac defects are poor operative risks. Innovative interventional procedures improve vascular anatomy, reduce pressure loads on ventricles, and decrease the repetitive operative risk for these patients. For example, in tetralogy of Fallot with hypoplastic PAs, balloon angioplasty and vascular stenting procedures create favorable PA anatomy and reduce proximal PA pressure and right ventricular end-diastolic pressure.

The anesthetic management of interventional procedures in the catheterization laboratory must include the same level of preparation that would apply in caring for these patients in the operating room. The patients have the same complex cardiac physiology and, in some cases, greater physiologic complexity and less cardiovascular reserve because of their poorer operative risks. Interventional catheterization procedures can impose acute pressure load on the heart during balloon inflation. Large catheters placed across mitral or tricuspid valves create acute valvular regurgitation or, in the case of a small valve orifice, transient valvular stenosis. When catheters are placed across shunts, severe reductions in PBF and marked cyanosis may occur ( Malviya et al., 1989 ; Hickey et al., 1992 ). The anesthetic plan must consider the specific cardiology objectives of the procedure and the impact of anesthetic management in facilitating or hindering the interventional procedure. In general, there are three distinct periods involved in an interventional catheterization: the data acquisition period, the interventional period, and the postprocedural evaluation period.

During the data acquisition period, the cardiologist per-forms a hemodynamic catheterization to evaluate the need for and extent of the planned intervention. Catheterization data are obtained under baseline or normal physiologic conditions (i.e., room air, a physiologic Paco2, and either controlled or spontaneous ventilation). Because of the complexity of the interventional procedures and the need to maintain normal Paco2 and immobility, general anesthesia with controlled ventilation is increasingly required for children in the catheterization laboratory. Increased Fio2 or changes in Paco2 may affect physiologic data. A secured airway allows the anesthesiologist to concentrate on hemodynamic issues. Positive pressure ventilation also reduces the risk of air embolism. During spontaneous ventilation, a large reduction in intrathoracic pressure may entrain air into vascular sheaths and result in moderate to large pulmonary or systemic air emboli. Precise device placement is also facilitated with controlled ventilation by reducing the respiratory shifting of cardiac structures.

Substantial blood loss and changes in ventricular function occur commonly during the intervention. Blood volume replacement and inotropic support may be necessary during or immediately after the interventional procedure. In the postprocedural period, the success and the physiologic impact of the intervention are evaluated. Blood pressure, mixed venous oxygen saturation, ventricular end-diastolic pressure, and cardiac output, when available, are used to assess the impact of the intervention. Postintervention recovery is generally done in an intensive care unit environment for monitoring and/or respiratory or cardiovascular support.

Because of the hemodynamic variability of many of these patients and the changing anesthetic requirements, continuous intravenous infusion with ketamine/midazolam, or propofol with fentanyl, alfentanil, or remifentanil is used. Many children with congenital cardiac disease may safely undergo induction of anesthesia with sevoflurane and then, after an intravenous cannula is secured, a maintenance infusion of propofol (50 to 125 mcg/kg per minute) along with remifentanil can be started. Potent inhaled anesthetics are generally not used as the primary anesthetic because of their negative inotropic effects; they are reserved for adjunctive anesthesia.

A brief description of some of the interventional procedures and the associated anesthetic implications are presented. The success and safety of these procedures, together with the miniaturization of intracardiac stents and closure devices, expand the therapeutic options for pediatric interventional cardiologists.

A fenestrated Fontan may have a closure device deployed if needed or a prematurely closed fenestration reopened and stented if necessary (Chatrath et al., 2003). Many of these procedures require TEE or intracardiac echocardiography ( Zanchetta et al., 2003 ) to guide accurate stent or occluder deployment. The provision of anesthesia in the cardiac catheterization laboratory is challenging: the room is dark, the cardiac catheterization laboratory is generally poorly designed for anesthesia equipment, and communication can sometimes be less than ideal.


During the past 10 years, several different closure devices have reached the market, and most have been removed. Only the Amplatzer (AGA Medical Corporation, Golden Valley, MN) expanding heat-treated nickel-titanium double-wire disc is available in the United States ( Fig. 17-32 ). The double disc flattens out once deployed with each side of the disc on opposite sides of the atrial septum. The original device used in the late 1980s and early 1990s was a collapsed, double-umbrella or clamshell device. This device had six arms within a meshed patch. The clamshell device was taken off the market because follow-up evaluation demonstrated cracks or breaks in some of the arms.


FIGURE 17-32  Amplatzer (AGA Medical Corporation, Golden Valley, MN) heat-treated nickel-titanium wires. (A) Open Amplatzer device. (B) Deployed Amplatzer device with the two flattened titanium-nickel discs that lie on each side of the atrial or ventricular wall after deployment for repair of an atrial or ventricular septal defect.



The ASD closure device is deployed by the cardiologist through a large introducer sheath that is placed in the femoral vein. A catheter containing the device is advanced from the femoral vein to the RA and placed across the ASD into the left atrial chamber. With the use of biplane fluoroscopy and TEE, the catheter is positioned in the LA away from the mitral valve ( Lock et al., 1989 ). The sheath is pulled back to open the distal side of the device into the LA. The sheath and device are then pulled back so the distal side makes contact with the left atrial side of the septum. Fluoroscopy and TEE are used to confirm that the device is appropriately positioned and does not interfere with mitral valve motion. Once adequately seated, the sheath is pulled farther back to expose the proximal side of the device, which is then opened to engage on the right side of the atrial septum. When proper positioning is certain, the device is released ( Lock et al., 1989 ). In a report of 122 children undergoing transcatheter ASD closures using the clamshell device, there was a 9% incidence of procedural complications resulting in hemodynamic complications requiring treatment ( Hickey et al., 1992 ). Device embolization to either the right or left side of the circulation does occur and requires further interventions to retrieve them. One series reported an embolization incidence as high as 7% ( Thanopoulos et al., 2003 ). There is an institutional learning curve with these newer devices, and most centers with experience can lower the incidence of device embolization to about 2% to 3% ( Godart et al., 2003 ) ( Table 17-18 ).

TABLE 17-18   -- Atrial septal defect (ASD) closure device


No. (%)


Device embolization

6 (5)

Two retrieved in operating room

Air embolization

4 (3.3)

Pressors, atropine, volume, cardiopulmonary resuscitation seizure

Acute mitral regurgitation (MR), tricuspid regurgitation (TR)

2 (1.7)

Pressors reposition device

Chronic TR

1 (0.8)

Device removed in operating room; ASD closed

Brachial plexus injury

3 (2.5)


Dye reaction


Pressors and ventilation


2 (1.7)

Apnea/airway obstruction




Most VSDs that are electively closed in the catheterization laboratory are midmuscular or apical VSDs, which are either difficult to close in the operating room or would require a left ventriculotomy. Left ventriculotomies are associated with a high incidence of late left ventricular dysfunction and have fallen into disfavor as a surgical option. The transcatheter approach requires a blade atrial septostomy and a retrograde catheter placed through the femoral artery and advanced to the LA. This catheter is pulled across the atrial septum into the RA and is used to guide an SVC catheter (placed through the internal jugular vein) across the created ASD into the LA, across the mitral valve, and into the LV. The VSD is approached from the left ventricular side. The left side of the intraventricular septum is smooth, whereas the right side is trabeculated. By placing the device on the left side, the VSD opening can be identified. The large sheath containing the closure device interferes with closure of the mitral valve, resulting in acute mitral regurgitation or, in cases where the VSD is large or the mitral annulus small, acute severe mitral valve obstruction. In this latter case, systemic outflow is decreased and a period of severe hypotension may be experienced. Judicious use of vasoconstrictors to maintain coronary perfusion during the catheter placement, followed by volume and inotropic resuscitation after the VSD device is secured, may be required.


One of the most important areas of interventional catheterization has been the dilation and stenting of hypoplastic or stenotic branch PAs. In patients with tetralogy of Fallot with hypoplastic PAs, pulmonary atresia, or single ventricle with surgically induced peripheral stenosis at the site of a previous shunt, the use of balloon angioplasty and stenting procedures creates favorable PA anatomy and reduces the risk of subsequent surgical repairs ( Fig. 17-33 ). Balloon angioplasty is accomplished by tearing the vascular intima and media, allowing the vessel to remodel and heal with a larger diameter. The balloon is placed across the stenotic lesion so the middle of the balloon is at the stenosis. The balloon is inflated until the waist of the balloon is eliminated. Ideally, the most stenotic lesions are dilated first tominimize the impact on PBF and cardiac output. When the balloon is inflated, PBF is reduced, right ventricular afterload is increased, and cardiac output falls. In patients with an associated VSD, right-to-left shunting and desaturation occur with balloon inflation. Occasionally, balloon catheters must be placed across aortopulmonary shunts, significantly reducing PBF. The procedure is successful in approximately 60% of patients. Complications include hypotension (40%), PA rupture (3%), unilateral reperfusion pulmonary edema (4%), aneurysmal dilation of the dilated pulmonary vessel (8%), death (1.5%), and transient postprocedural right ventricular dysfunction. Anesthetic support minimizes hemodynamic compromise by anticipating changes in blood flow patterns, treating transient hypotension, and providing airway support to minimize the risks of PA disruption and acute unilateral pulmonary edema ( Rothman et al., 1990 ).


FIGURE 17-33  (A) Severe bilateral branch pulmonary artery stenosis at the distal end of a conduit in a patient with pulmonary atresia and ventricular septal defect. Stents were placed in the right and left pulmonary arteries. (B) Follow-up angiogram in the same projection and magnification showed marked improvement of both right and left pulmonary artery stenosis.



Balloon valvotomies are common interventional procedures. Acute inflation of the balloon results in a transient absence of forward flow and marked elevation in pressure in the chamber proximal to the valve. For aortic and pulmonary valvotomy, there may be acute ischemic injury to the ventricle with a drop in cardiac output. Measurement of the gradient across the valve may be artificially lower because of a reduction in flow across the valvular orifice. Successful balloon dilation of a stenotic valve also requires tearing of fused valve leaflets. Occasionally, the valve leaflets tear at nonfused regions, or the valves are damaged during the procedure, causing regurgitation. A mild degree of valvular regurgitation is well tolerated. Severe valvular regurgitation is generally poorly tolerated, because there is a rapid change in loading conditions from stenosis to severe valvular insufficiency. Patients with significant ventricular dysfunction before the intervention are particularly at risk for postprocedural hemodynamic instability.


Radiofrequency ablation is a nonsurgical approach to eliminat ing atrial or ventricular reentrant tachyarrhythmias. The technique requires arrhythmia pathway mapping and precision ablation of the aberrant pathway using a radiofrequency ablation catheter. During the ablation, unexpected patient movement may result in catheter dislodgment and damage to normal conducting tissue, so general anesthesia is usually required in younger children. Anesthetic agents and techniques should be chosen to increase circulating catecholamines and avoid suppression of arrhythmogenesis so as to promote identification of the aberrant pathway. Several authors ( Sharpe et al., 1994 ; Erb et al., 2002b ) described the electrophysiologic effects of volatile anesthetics (enflurane, isoflurane, and halothane), propofol, and opioid-benzodiazepine-based anesthetics in patients undergoing radiofrequency ablation procedures. Volatile anesthetics prolonged the refractory periods of both normal atrioventricular conduction and the accessory pathway. Volatile anesthetic gases interfered with data interpretation and rendered postablative studies less reliable in judging the success of the procedure. The use of volatile anesthetics should be discouraged during electrophysiologic procedures. Conversely, patients with reentry tachycardias requiring anesthesia for noncardiac surgery may benefit from volatile anesthetic agents as they may reduce the incidence of supraventricular tachycardia during the operative procedure. Opioid-benzodiazepine anesthesia had no demonstrable effect on electrophysiologic measurements and therefore provided electrophysiologic data unencumbered by anesthetic effect.

A combination of propofol and opioids as a continuous intravenous anesthetic, along with nitrous oxide and neuromuscular blockade, does not affect the electrophysiology data. In addition, this anesthetic combination has been shown to decrease the incidence of postoperative nausea and vomiting compared with an inhalational anesthetic technique (Erb et al., 2002a, 2002b [117] [118]).

Rapid atrial pacing and, occasionally, an isoproterenol infusion are required during the mapping procedure. Severe postprocedural cardiomyopathy has been described, but it is very unusual. An underlying cardiomyopathy from frequent episodes of supraventricular tachycardia and myocardial oxygen imbalance caused by prolonged periods of rapid atrial pacing and isoproterenol infusions are the presumed causative factors.

Strict attention to arm support and padding of all pressure points are essential. Tension on the brachial plexus must be avoided, especially if arms are secured next to head at less than 90 degrees flexion/extension. Also, pressure on the radial nerve at the elbow can occur, especially for the longer radiofrequency arrhythmia ablation procedures. A peripheral arterial catheter is helpful during these lengthy procedures for continuous monitoring of arterial blood pressure and blood gases.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


Refinements in diagnostic and operative techniques, as well as advances in medical and anesthetic care, have resulted in marked improvement in survival of patients with CHD. Consequently, some of these patients may be subjected to elective or emergency noncardiac surgery either before or after correction of the cardiac anomaly. The anesthetic management of patients with uncorrected heart defects allows for little margin of error. It is essential that the anesthesiologist be knowledgeable not only of the basics of pediatric anesthesia but also of the pathophysiology of the cardiac lesions ( Salem and Griffin, 1981 ).

The preoperative evaluation and preparation of patients with CHD for noncardiac surgery are not any different from the evaluation and preparation of those children undergoing repair of CHD (as described earlier). Because of the high incidence of associated congenital anomalies in patients with CHD, these anomalies must be ascertained, and the anesthetic plan must be altered appropriately ( Moore, 1981 ;Salem and Griffin, 1981 ; Schwartz, 1985 ).

If the child has undergone one or more cardiac procedures in the past, it is essential to know whether the surgery was palliative or corrective, as well as if any complications or residual cardiac disease exists. For example, after repair of coarctation, some patients develop aortic stenosis or aortic insufficiency from the gradual thickening or everting of the bicuspid aortic valve. In some patients with repaired ASDs, 8% to 37% have angiographic evidence of mitral valve prolapse ( Betrin et al., 1975 ). In patients who underwent a repair of tetralogy of Fallot, unifocal, multifocal, and repetitive premature ventricular contractions at rest or after exercise have been associated with ventricular tachycardia, ventricular fibrillation, and sudden death ( Quattlebaum et al., 1976 ; Gillette et al., 1977 ). These patients with arrhythmias usually have residual hemodynamic abnormalities such as infundibular stenosis that are arrhythmogenic. Sinoatrial node dysfunction has been seen in patients after the Mustard procedure and ASD repair.

In children with uncomplicated CHD who undergo noncardiac procedures, surgery may be scheduled as an outpatient procedure. For patients with complex CHD, hospital admission should be considered the day before surgery, so that appropriate consultation and preparation can be made. In all patients, consultation with a pediatric cardiologist and access to recent echocardiography and catheterization data should be an integral part of the preoperative management. The anesthetic management decisions regarding induction technique, airway management, and maintenance of anesthesia are based on the patient's functional status, the pathophysiology of the underlying defect, the proposed operative procedure, and the anticipated hemodynamic response to the anesthetic agent. The cardiac grid ( Table 17-19) may be used to construct an anesthetic plan.

TABLE 17-19   -- Cardiac grid for common congenital heart lesions (desired hemodynamic changes)





Heart Rate





VSD (left-to-right)



VSD (right-to-left)






N ↑




Coarctation of the aorta








N ↑




N ↑



N ↑

N ↑



N ↑

N ↑

AR, aortic regurgitation; AS, aortic stenosis; ASD, atrial septal defect; IHSS, idiopathic hypertrophic subaortic stenosis; MR, mitral regurgitation; MS, mitral stenosis; PDA, patent ductus arteriosus; PS, pulmonary stenosis; PVR, pulmonary vascular resistance; SVR, systemic vascular resistance; VSD, ventricular septal defect.




Full resuscitative capabilities, including a defibrillator, must be available in the anesthetizing locations ( Schwartz and Jobes, 1982 ). Resuscitation drugs must be readily accessible before anesthetic induction. These drugs should include epinephrine, sodium bicarbonate, atropine, lidocaine, and phenylephrine. The drugs may be drawn up and diluted to appropriate concentrations if the patient's clinical condition warrants it.

The only patients in whom antibiotic prophylaxis is not recommended are those with an isolated unrepaired secundum ASD, with a secundum ASD repaired with a patch, and with a previously ligated PDA.

Except in superficial and peripheral surgical procedures of short duration, controlled ventilation is desirable in these patients. Hypoventilation is poorly tolerated in CHD patients with limited cardiac reserve. In patients with tetralogy of Fallot, excessive mean airway pressure reduces cardiac filling and leads to increased right-to-left shunting across a VSD and cyanosis.

Preinduction monitoring of the child with cardiac disease for noncardiac surgery includes precordial stethoscope, electrocardiography, noninvasive blood pressure measurements, pulse oximetry, and temperature probe. Additional monitoring depends on the child's underlying cardiac status, disease state, and extent and duration of the anticipated surgery ( Moore, 1981 ; Schwartz, 1985 ).

Postoperative respiratory support is dependent on the type and duration of surgery and/or the presence of significant cardiac or pulmonary dysfunction. Weaning from mechanical ventilatory support can gradually take place in the intensive care unit. For children in whom tracheal extubation is planned for at the end of the operative procedure, the criteria for extubation are similar to those for any child. The cardiovascular responses to reversal of neuromuscular blockade have been studied in healthy children and in children with cyanotic and acyanotic CHD ( Salem et al., 1970 ; Wong et al., 1974 ). Despite slight changes in heart rate during reversal, cardiac output remains essentially unchanged during reversal ( Sale et al., 1977) . Antagonism of neuromuscular blockade need not be withheld for fear of arrhythmias. When the procedure is complete, the patient is transferred to the recovery room (or an intensive care environment if indicated) while oxygen is being administered. Close observation and monitoring of these children in the early postoperative period are essential.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier



Cardiomyopathies are defined as “diseases of the myocardium associated with cardiac dysfunction” ( Richardson et al., 1996) . They are classified based on newer pathophysiologic criteria. The five classes include dilated, hypertrophic, restrictive, arrhythmogenic RV, and unclassified. This classification is helpful to the anesthesiologist in optimizing an anesthetic plan for patients with cardiomyopathy. Cardiac surgery per se is not often performed in these children; however, they do present at end stage when cardiac transplantation or bridging to transplant is often the only alternative. Other patient encounters include placement of AICDs, central venous catheter placement, cardiac catheterization laboratory, muscle biopsies, dental rehabilitation, endoscopic diagnostic procedures, or childbirth.

Cardiomyopathies may be caused by ischemic heart disease (uncommon in children), toxin exposure, valvular heart disease, hypertension, inflammatory diseases (e.g., lupus), viral infections, metabolic diseases, and inherited disorders (e.g., glycogen storage disease type II, Pompe's disease). These etiologies usually fall into one of the pathophysiologic groups.

Dilated Cardiomyopathy

Systolic function is depressed usually with large ventricular volumes, elevated left ventricular end-diastolic pressures, high pulmonary venous pressures, and the presence of mitral or tricuspid valve regurgitation. The electrocardiogram may show bundle branch block, there may be ST-T wave abnormalities, and it is prudent to measure electrolytes before anesthesia, because these patients are on diuretics and possibly digoxin. Anesthetic management should focus on maintaining adequate preload, but over hydration causes pulmonary edema. Heart rate should be maintained at preinduction levels, because stroke volume is limited. The response to inotropes may be blunted, because stroke volume may increase minimally while heart rate increases excessively, resulting in impaired myocardial oxygen balance and ischemia. Mild afterload reduction is generally a preferred approach to enhance forward flow. Milrinone is a good inotrope and vasodilator for augmenting cardiac output in this type of cardiomyopathy. Afterload should be maintained at a normal-to-low SVR. A high afterload is not tolerated by a hypocontractile dilated cardiomyopathy. Etomidate 0.3 mg/kg is a hemodynamically stable induction agent; very low-dose maintenance inhalational agent may be used, together with fentanyl 2 to 5 mcg/kg and low-dose midazolam to prevent excessive SVR elevation. Monitoring may require a central venous catheter and an intra-arterial catheter to measure beat-to-beat blood pressure ( Schechter et al., 1995 ).

Hypertrophic Cardiomyopathy

The pathophysiology associated with this diagnosis is noted for preservation of systolic function but poor diastolic function. Ventricular concentric or asymmetric hypertrophy usually develops, and over time, this decreases the ventricular cavity size. Frequently, the hypertrophied myocardium compresses endocardial blood vessels ( Mohiddin et al., 2002) . LVOT obstruction may be present. A reduction in preload increases dynamic obstruction and worsens cardiac output. An excessive decrease in SVR reduces coronary perfusion pressure and could result in ischemia. This sets up a downward spiral of diminished coronary blood flow, increased myocardial ischemia, and the risk of intraoperative ventricular fibrillation and sudden cardiac arrest. The anesthetic management of patients with this pathophysiology should be aimed at maintaining normal to slightly elevated SVR and preload at normal to moderately elevated levels. Heart rate is usually kept at a low to normal level to optimize diastolic filling time and enhance stroke volume. Inotropes are seldom needed and may decrease cardiac output in very severe hypertrophic cardiomyopathy, because systolic cavity obliteration may occur and diastolic filling time is reduced due to increased heart rate. Spontaneous ventilation is optimal for these patients because it preserves cardiorespiratory interactions. In the infantile form of glycogen storage disease type II or Pompe's disease, in which patients develop a severe form of hypertrophic cardiomyopathy, an anesthetic technique using spontaneous ventilation under ketamine and fentanyl titration with local or regional anesthesia has been found to be an alternative to inhalational general anesthesia (I ng et al., 2004 ). Propofol as the sole anesthetic agent in glycogen storage disease type II may not offer the safest anesthetic hemodynamic profile in the severe form of this cardiomyopathy (I ng et al., 2004 ), because it significantly decreases preload and SVR ( Williams et al., 1999 ).

Restrictive Cardiomyopathy

This is a rare form of cardiomyopathy. It usually results in endomyocardial fibrosis with a severe decrease in ventricular compliance and poor ventricular filling. In the early stage of the disease, ejection fraction is maintained; as restriction progresses with time, left ventricular end-diastolic pressure increases, stroke volume significantly decreases, and cardiac output decreases. One of the distinguishing features of restrictive cardiomyopathy is the relentless progression of an elevated left ventricular end-diastolic pressure and resultant increase in PVR. By the time many children with a restrictive cardiomyopathy present with exercise intolerance, PVR is already greater than 10 to 15 Woods units; this prohibits them from being a recipient of a heart transplant ( Weller et al., 2002 ).

The most common etiologies for restrictive cardiomyopathy in children are endomyocardial fibrosis and Loffler's endocarditis. Some believe them to be part of the same spectrum, but in Loffler's disease, intraventricular cavity thrombus formation is common, and eosinophilic infiltration into other organs causes small vessel arteritis. Endomyocardial fibrosis is characterized by fibrosis of the ventricular apex and mitral valve. LVOT obstruction is rare (Davies, 1960). The diagnosis is often made by echocardiography, showing small ventricles, massively dilated atria, and, rarely, pericardial effusion ( Weller et al., 2002 ). The jugular venous pressure is often raised and can show a paradoxical increase in height during spontaneous inspiration. This is due to increased venous return into a noncompliant ventricle.


Acute Pericarditis

Acute pericarditis rarely causes restriction. It can occur in children at any age, and quite commonly the etiology is purulent pericarditis. Staphylococcus causes about 40% of cases; collagen vascular disease, 30%; virus, 20%; and radiation therapy and, rarely, neoplastic disease, 10% ( Roodpeyma and Sadeghian, 2000 ; Kohli et al., 2001 ). A pericardiectomy is almost never required with acute pericarditis, but the pericardial cavity may need to be drained by a subxiphoid catheter that is usually inserted with transthoracic echocardiographic guidance in a sterile fashion under sedation. Propofol infusion with local anesthetic infiltration or occasionally ketamine is commonly used. Cardiac tamponade may be seen with associated acute carditis, but this is rare ( Roodpeyama and Sadeghian, 2000 ).

Constrictive Pericarditis

Constrictive pericarditis is a chronic condition more commonly seen in the developing world. Tuberculosis is the number 1 offending pathogen. In more-developed nations, neoplastic disease, inflammatory conditions, and radiation therapy-induced constrictive pericarditis are more commonly seen ( Kohli et al., 2001 ). Constrictive pericarditis is clinically suspected when there is a raised jugular venous pressure, hepatomegaly, pedal edema, pleural effusions, and massive ascites. Surgical therapy necessitates a pericardiectomy through a midline sternotomy, and although the heart may be cannulated for emergent CPB, all surgical attempts are made to complete the procedure without CPB because dissection of the adherent pericardium may cause myocardial stunning. Postoperatively, low cardiac output syndrome is commonly observed. Theoretically, the pericardium over the LV is removed first. It is argued that if the right side were released first, an increased ventricular volume may increase PBF and increase volume return to a small noncompliant LV, and this may cause pulmonary edema. This is rarely seen clinically, because as a practical matter, both ventricles are often released simultaneously as the surgical dissection steadily progresses.

Pericardial Effusion and Cardiac Tamponade

Postoperative tamponade may be caused by a rapidly increasing collection of fluid into the pericardial space if intact or directly into either the mediastinum or the pleural space if the peri cardium is left open after cardiac surgery. A combination of fresh blood and formed clot is present and often requires urgent/emergent surgical reexploration for evacuation. The signs and symptoms are clinical and should be acted on before significant hemodynamic compromise occurs. If time permits, TEE should be obtained to confirm the diagnosis. The clinical presentation includes tachycardia, hypotension, central venous pressure elevation, worsening pulmonary compliance, and oxygen desaturation with poor peripheral perfusion. Early diagnosis and prompt therapy can ensure good outcome even in very small infants ( Wirrell et al., 1993) . In older children, pulsus paradoxus and distended neck veins are seen.

Anesthetic management of pericardial effusion depends on whether there is tamponade. In the presence of tamponade, a pericardial catheter should be placed using local anesthesia. If tamponade is not present based on echocardiographic evaluation, the use of a slow titration of anesthetic agents such as propofol is well tolerated. Patients should be volume loaded to optimize stroke volume before the administration of anesthetic drugs, and resuscitative agents should be available for use at the bedside.

A common source of unexpected cardiac tamponade in neonates, infants, and children is related to vascular or heart perforation from central lines. Line location has been looked at as a cause for cardiac perforation, with a higher risk identified for neonates. In neonates, general recommendations for the optimal placement of the distal tip of peripherally inserted central catheters, as well as percutaneously placed internal jugular or subclavian venous catheters, should lie at the SVC-RA junction (Nadroo et al., 2001). This corresponds to a level above T-2 as documented by chest radiography ( Bargy et al., 1986 ).

In the operating room, central venous catheters are placed without radiologic confirmation. On arrival in the intensive care unit, a chest radiograph is obtained and the catheter placement is adjusted if indicated. Despite these recommendations, numerous reports exist of iatrogenic cardiac tamponade, often suspected as delayed migration or excessive movement of the catheter tip due to abduction of the arm or head flexion/extension changes. Catheter mobility of 3 to 5 cm has been described ( Brandt et al., 1970 ) for subclavian and internal jugular central catheters, respectively. For peripherally placed central catheters, especially when crossing a joint area, greater degrees of movement have been documented, and it is recommended to limit limb movement ( Henzel et al., 1971) . This may explain the continued occurrence of cardiac tamponade despite appropriate catheter positioning. Although the recommendation for optimal placement of the catheter tip is at the SVC-RA juncture, it may actually be 1 cm within the RA. The rationale for this location is that the pericardium extends 1 to 2 cm proximal from the SVC RA juncture. The narrow SVC can be perforated by the tip of the catheter, leading to infusion of fluid into the pericardial space. By placing the tip within the wider opening of the RA, the tip swings freely and does not engage the wall of the RA, similar to a clapper in a bell.

Infants with catheter migration and gradual pericardial infusion of fluid usually present with sudden cardiac arrest; very few are successfully resuscitated with pericardiocentesis ( Aiken et al., 1992 ; Nadroo, 2001). Retrospectively, it is sometimes noted that the patient may have been gradually deteriorating hemodynamically, with signs of dyspnea, tachycardia, or nausea misinterpreted for 12 to 24 hours before the cardiac arrest ( Arbitman and Kart, 1979 ). For any patient with a central venous catheter in situ and a diagnosis of pericardial effusion, cardiac echocardiography with a small injection of agitated saline should be performed to confirm that the catheter tip is not the source of the pericardial infusion. If the catheter tip is seen in the pericardial space and that central catheter is used, any further injection of fluid worsens the effusion and contributes to worsening of the cardiac tamponade.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


The anesthetic management of children undergoing cardiac surgery requires a sound understanding of the basic principles of cardiovascular anesthesia, pediatric anesthesia, and intensive care and a working knowledge of the pathophysiology of congenital cardiac disease. Anesthesiologists must also know how to treat right ventricular dysfunction, left ventricular dysfunction, and PA hypertension and be aware of the potential residual anatomic problems associated with cardiac surgical interventions. This requisite knowledge, coupled with an open mind toward the ever-expanding strategies and techniques being developed worldwide for optimizing care of complex CHD patients, is essential. Through constant evolution and hard work, the overall life expectancy and life quality of the child with CHD continue to improve and we hope approach those of children born with a normal cardiovascular system.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


Aaronson et al., 2002. Aaronson PI, Robertson TP, Ward JP: Endothelium-derived mediators and hypoxic pulmonary vasoconstriction.  Respir Physiol Neurobiol  2002; 132:107-120.

Abdul-Khaliq et al., 2002. Abdul-Khaliq H, Troitzsch D, Schubert S, et al: Cerebral oxygen monitoring during neonatal cardiopulmonary bypass and deep hypothermic circulatory arrest.  Thorac Cardiovasc Surg  2002; 50:77-81.

Adesuyi et al., 1985. Adesuyi S, Cockrell C, Gamache D, et al: Lipoxygenase metabolism of arachidonic acid in brain.  J Neurochem  1985; 45:770.

Aiken et al., 1992. Aiken G, Porteous L, Tracy M, Richardson V: Cardiac tamponade from a fine Silastic central venous catheter in a premature infant.  J Paediatr Child Health  1992; 28:325.

Aisen, 1980. Aisen P: The transferrins. Iron in biochemistry and medicine, II,  New York, Academic Press, 1980.

Alboliras et al., 1985. Alboliras E, Porter C, Danielson G, et al: Results of the modified Fontan operation for congenital heart lesions in patients without preoperative sinus rhythm.  J Am Coll Cardiol  1985; 6:228-235.

Allen et al., 2003. Allen SW, Gauvreau K, Bloom BT, et al: Evidence-based referral results in significantly reduced mortality after congenital heart surgery.  Pediatrics  2003; 112:24-28.

Alvis and Reves, 1985. Alvis JM, Reves JG, et al: Computer-assisted continuous infusions of fentanyl during cardiac surgery: Comparison with a manual method.  Anesthesiology  1985; 63:41-50.

American Academy of Pediatrics, Section on Cardiology and Cardiac Surgery, 2002. American Academy of Pediatrics, Section on Cardiology and Cardiac Surgery : Guidelines for pediatric cardiovascular centers.  Pediatrics  2002; 109:544-549.

Anand et al., 1990. Anand KJS, Hansen DD, Hickey PR: Hormonal–metabolic stress response in neonates undergoing cardiac surgery.  Anesthesiology  1990; 73:661.

Anand and Hickey, 1992. Anand KJS, Hickey PR: Halothane-morphine compared with high-dose sufentanil for anesthesia and postoperative analgesia in neonatal cardiac surgery.  N Engl J Med  1992; 326:1-9.

Anand et al., 1987. Anand KJS, Sippell WG, Aynsley-Green A: Randomized trial of fentanyl anesthesia in preterm neonates undergoing surgery: Effects on the stress response.  Lancet  1987; 1:243.

Anderson, 1989. Anderson PAW: Maturation and cardiac contractility.  Cardiol Clin  1989; 7:209-225.

Andrew et al., 1987. Andrew MB, Paes B, Milner R, et al: Development of the human coagulation system in the full-term infant.  Blood  1987; 70:165-172.

Arbitman and Kart, 1979. Arbitman M, Kart H: Hydromediastinum after aberrant central venous catheter placement.  Crit Care Med  1979; 7:27-29.

Arciniegas, 1985. Arciniegas E: Pediatric cardiac surgery,  Chicago, Year Book Medical Publishers, Inc., 1985.

Ashwal, 1997. Ashwal S: Brain death in the newborn.  Curr Perspect Clin Perinatol  1997; 24:859-882.

Astrup, 1982. Astrup J: Energy-requiring cell functions in the ischemic brain.  J Neurosurg  1982; 56:482.

Atallah-Yunes and Kavey, 1996. Atallah-Yunes NH, Kavey RE, et al: Postoperative assessment of a modified surgical approach to repair of tetralogy of Fallot: Long term follow up.  Circulation  1996; 94(Suppl II):II-22-II-26.

Attenhofer Jost et al., 2002. Attenhofer Jost CH, Schaff HV, Connolly HM, et al: Spectrum of reoperations after repair of aortic coarctation: Importance of individualized approach because of coexistent cardiovascular disease.  Mayo Clin Proc  2002; 77:646-653.

Ausems et al., 1986. Ausems ME, Hug CC, Stanski DR, et al: Plasma concentrations of alfentanil required to supplement nitrous oxide anesthesia for general surgery.  Anesthesiology  1986; 75:362-370.

Austin et al., 1997. Austin 3rd EH, Edmonds Jr HL, Auden SM, et al: Benefit of neurophysiologic monitoring for pediatric cardiac surgery.  J Thorac Cardiovasc Surg  1997; 114:707-715.

Balaji et al., 1991. Balaji S, Gewillig M, Bull C, et al: Arrythmias after the Fontan procedure: Comparison of total cavopulmonary connection and atriopulmonary connection.  Circulation  1991; 84(Suppl III):III-162-III-167.

Bargy et al., 1986. Bargy F, Barbet P, Houette G: The pericardium of the newborn infant. Anatomic and radioanatomic study with the view toward better positioning central catheters in the superior vena cava.  Bull Assoc Anat  1986; 70:47.

Barratt-Boyes et al., 1971. Barratt-Boyes BG, Simpson M, Neutze JM: Intracardiac surgery in neonates and infants using deep hypothermia with surface cooling and limited cardiopulmonary bypass.  Circulation  1971; 53:25-30.

Barratt-Boyes et al., 1980. Barratt-Boyes BG, Simpson MM, Neutze JM: Intracardiac surgery in neonates and infants using deep hypothermia.  Circulation  1980; 62(Suppl III):III-73-III-80.

Barton et al., 2004. Barton P, Kalil AC, Nadel S, et al: Safety, pharmacokinetics and pharmacodynamics of drotrecogin alfa (activated) in children with severe sepsis.  Pediatrics  2004; 113(1 Pt 1):7-17.

Bashein et al., 1990. Bashein GBD, Townes BD, Nessly BS, et al: A randomized study of carbon dioxide management during hypothermic cardiopulmonary bypass.  Anesthesiology  1990; 72:7-15.

Baylen et al., 1977. Baylen B, Meyer RA, Korfhagen J, et al: Left ventricular performance in the critically ill premature infant with patent ductus arteriosus and pulmonary disease.  Circulation  1977; 55:182-188.

Becker and Caruso, 1981. Becker AE, Caruso G: Congenital heart disease—a morphologist's view on myocardial dysfunction,  Edinburgh, Churchill Livingstone, 1981.

Bellinger et al., 1988. Bellinger DC, Wernovsky G, Rappaport L, et al: Rapid cooling of infants on cardiopulmonary bypass adversely affects later cognitive function.  Circulation  1988; 78:A358.

Bellinger et al., 1999. Bellinger DC, Wypij D, Kuban KC, et al: Developmental and neurological status of children at 4 years of age after heart surgery with hypothermic circulatory arrest or low-flow cardiopulmonary bypass.  Circulation  1999; 100:526-532.

Bellinger et al., 2001. Bellinger DC, Wypij D, du Plessis AJ, et al: Developmental and neurologic effects of alpha-stat versus pH-stat strategies for deep hypothermic cardiopulmonary bypass in infants.  J Thorac Cardiovasc Surg  2001; 121:374.

Bellinger et al., 2003. Bellinger DC, Bernstein JH, Kirkwood MW, et al: Visual-spatial skills in children after open-heart surgery.  J Dev Behav Pediatr  2003; 24:169-179.

Bellinger et al., 2003. Bellinger DC, Wypij D, duPlessis AJ, et al: Neurodevelopmental status at eight years in children with dextro-transposition of the great arteries: The Boston circulatory arrest trial.  J Thorac Cardiovasc Surg  2003; 26:1385-1396.

Bender et al., 1997. Bender KA, Alexander JA, Enos JM, et al: Effects of inhaled nitric oxide in patients with hypoxemia and pulmonary hypertension after cardiac surgery.  Am J Crit Care  1997; 6:127-131.

Bengur et al., 1998. Bengur AR, Li JS, Herlong JR, et al: Intraoperative transesophageal echocardiography in congenital heart disease.  Semin Thorac Cardiovasc Surg  1998; 10:255-264.

Benveniste et al., 1984. Benveniste H, Drejer J, Schousboe A, et al: Elevations of the extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis.  J Neurochem  1984; 43:1369-1374.

Bering, 1961. Bering EAJ: Effect of body temperature change on cerebral oxygen consumption during hypothermia.  Am J Physiol  1961; 200:417.

Berman, 1985. Berman W: The hemodynamics of shunts in congenital heart disease,  New York, Raven Press, 1985.

Berne and Levy, 1981. Berne RM, Levy MN: Coronary circulation and cardiac metabolism.   In: Berne RM, Levy MN, ed. Cardiovascular physiology,  St. Louis: The CV Mosby Co.; 1981.

Berner et al., 1989. Berner M, Oberhansli I, Rouge J, et al: Chronotropic and inotropic supports are both required to increase cardiac output early after corrective operations for tetralogy of Fallot.  J Thorac Cardiovasc Surg  1989; 97:297-302.

Berner et al., 1983. Berner MJC, Rouge JC, Friedli B: The hemodynamic effect of phentolamine and dobutamine after open-heart operations in children: Influence of underlying heart defect.  Ann Thorac Surg  1983; 35:643-650.

Bers, 2001. Bers D: Excitation-contraction coupling and cardiac contractile force,  2nd ed.. Dordrecht, the Netherlands: Kluwer; 2001:125-130.

Betrin et al., 1975. Betrin A, Wyle E, Felderhof C, et al: Prolapse of the posterior leaflet of the mitral valve associated with secundum atrial septal defect.  Am J Cardiol  1975; 35:363.

Bettex et al., 2003. Bettex DA, Schmidlin D, Bernath MA, et al: Intraoperative transesophageal echocardiography in pediatric congenital cardiac surgery: A two-center observational study.  Anesth Analg  2003; 97:1275-1282.

Bishop et al., 1986. Bishop CCR, Powell S, Rutt D, et al: Transcranial Doppler measurement of middle cerebral artery blood flow velocity: A validation study.  Stroke  1986; 17:913-915.

Blume et al., 1999. Blume ED, Altmann K, Mayer JE, et al: Evolution of risk factors influencing early mortality of the arterial switch operation.  J Am Coll Cardiol  1999; 33:1702-1709.

Bohn et al., 1980. Bohn DJ, Poirer CS, Edmonds JF: Efficacy of dopamine, dobutamine, and epinephrine during emergence from cardiopulmonary bypass in children.  Crit Care Med  1980; 8:367-372.

Bolling et al., 1990. Bolling SF, Bies LE, Bove E, et al: Augmenting intracellular adenosine improves myocardial recovery.  J Thorac Cardiovasc Surg  1990; 99:469-474.

Borow et al., 1981. Borow KM, Keane JF, Castaneda A, et al: Systemic ventricular function in patients with tetralogy of Fallot, ventricular septal defect and transposition of the great arteries repaired during infancy.  Circulation  1981; 64:878.

Bove and Lloyd, 1996. Bove EL, Lloyd TR: Staged reconstruction for hypoplastic left heart syndrome.  Ann Surg  1996; 224:387.

Bradley et al., 2001. Bradley SM, Simsic JM, Atz AM: Hemodynamic effects of inspired carbon dioxide after the Norwood procedure.  Ann Thorac Surg  2001; 72:2088-2093.

Brandt et al., 1970. Brandt RL, Foley WJ, Fink GH, et al: Mechanism of perforation of the heart with production of hydropericardium by a venous catheter and its prevention.  Am J Surg  1970; 119:311-316.

Breda et al., 1989. Breda MA, Drinkwater DC, Laks H, et al: Prevention of reperfusion injury in the neonatal heart with leukocyte-depleted blood.  J Thorac Cardiovasc Surg  1989; 97:654-658.

Bretschneider, 1980. Bretschneider HJ: Myocardial protection.  Thorac Cardiovasc Surg  1980; 28:295-299.

Bridges et al., 1990. Bridges ND, Jonas RA, Mayer J, et al: Bidirectional cavopulmonary anastomosis as interim palliation for high risk Fontan candidates: Early results.  Circulation  1990; 82(Suppl IV):IV-170-IV-176.

Brodde et al., 1992. Brodde O, Broede A, Daul A, et al: Receptor systems in the non-failing human heart.  Basic Res Cardiol  1992; 87(Suppl):1-14.

Brown et al., 2003. Brown JW, Ruzmetov M, Vijay P, et al: Surgery for aortic stenosis in children: A 40-year experience.  Ann Thorac Surg  2003; 76:1398-1411.

Bryant, 1985. Bryant PE: Enzymatic restriction of mammalian cell DNA: Evidence for double-strand breaks as potentially lethal lesions.  Int J Radic Res  1985; 48:55.

Buckberg, 1979. Buckberg GD: A proposed solution to the cardioplegia controversy.  J Thorac Cardiovasc Surg  1979; 77:803-806.

Buckberg, 1990. Buckberg GD: Oxygenated cardioplegia: Blood is a many splendored thing.  Ann Thorac Surg  1990; 50:175-177.

Bui et al., 1991. Bui KC, Hammerman C, Hirshl RB, et al: Plasma prostanoids in neonates with pulmonary hypertension treated with conventional therapy and with extracorporeal membrane oxygenation.  J Thorac Cardiovasc Surg  1991; 101:973-983.

Bush et al., 1987. Bush A, Busset C, Knight W, et al: Modification of pulmonary hypertension secondary to congenital heart disease.  Am Rev Respir Dis  1987; 136:767.

Busto et al., 1987. Busto R, Dietrich WD, Globus MY, et al: Small differences in the intra-ischemic brain temperature critically determine the extent of ischemic neuronal injury.  J Cereb Blood Flow Metab  1987; 7:729-736.

Busto et al., 1989. Busto R, Mordecai Y-T, Dietrich D, et al: Effects of mild hypothermia on brain ischemia.  Stroke  1989; 20:904-910.

Butler et al., 1993. Butler J, Rocker GM, Westaby S: Inflammatory responses to cardiopulmonary bypass.  Ann Thorac Surg  1993; 55:552-559.

Castaneda et al., 1989. Castaneda AR, Mayer JEJ, Jonas RA, et al: The neonate with critical congenital heart disease: Repair—a surgical challenge.  J Thorac Cardiovasc Surg  1989; 98:869-875.

Castaneda et al., 1988. Castaneda AR, Trusler GA, Paul MA, et al: The early results of treatment of simple transposition in the current era.  J Thorac Cardiovasc Surg  1988; 95:4-22.

Casteneda et al., 1994. Casteneda AR, Jonas RA, Mayer JE, et al: Chapter 22.  Cardiac surgery of the neonate and infant,  Philadelphia, PA: WB Saunders; 1994:353-362.

Celoria and Patton, 1959. Celoria GC, Patton RB: Congenital absence of the aortic arch.  Am Heart J  1959; 58:407.

Chang et al., 1995. Chang AC, Atz AM, Wernovsky G, et al: Milrinone: Systemic and pulmonary hemodynamic effects in neonates after cardiac surgery.  Crit Care Med  1995; 23:1907.

Chiba et al., 1993. Chiba Y, Muraoka R, Ihaya A, et al: The effects of leukocyte depletion on the prevention of reperfusion injury during cardiopulmonary bypass.  J Thorac Cardiovasc Surg  1993; 1:350-356.

Chizzonite and Zak, 1981. Chizzonite R, Zak R: Calcium-induced cell death: Susceptibility of cardiac myocytes is age-dependent.  Science  1981; 213:508-511.

Chopp et al., 1989. Chopp M, Knight R, Tidwell CD, et al: The metabolic effects of moderate hypothermia on global cerebral ischemia and recirculation in the cat: Comparison to monothermia and hypothermia.  J Cereb Blood Flow Metab  1989; 9:141-148.

Choudry and Brenn, 2002. Choudry DK, Brenn BR: Bispectral index monitoring: A comparison between normal and children with quadriplegic cerebral palsy.  Anesth Analg  2002; 95:1582-1585.

Civalero et al., 1962. Civalero L, Moreno J, Senning A: Temperature conditions and oxygen consumption during deep hypothermia.  Acta Chir Scand  1962; 123:179-188.

Clermont et al., 2002. Clermont G, Vergely C, Jazayeri S, et al: Systemic free radical activation is a major event involved in myocardial oxidative stress related to cardiopulmonary bypass.  Anesthesiology  2002; 96:80-87.

Codispoti et al., 2001. Codispoti M, Ludlam CA, Simpson D, Mankad PS: Individualized heparin and protamine management in infants and children undergoing cardiac operations.  Ann Thorac Surg  2001; 71:922.

Cohen et al., 2000. Cohen MI, Bridges ND, Gaynor JW, et al: Modifications to the cavopulmonary anastomosis do not eliminate early sinus node dysfunction.  J Thorac Cardiovasc Surg  2000; 120:891-900.

Colon-Otero et al., 1987. Colon-Otero G, Gilchrist GS, Holcomb GR, et al: Preoperative evaluation of hemostasis in patients with heart disease.  Mayo Clin Proc  1987; 62:379-385.

Cook, 2001. Cook DJ: Con: Low-flow cardiopulmonary bypass is not the preferred technique for patients undergoing cardiac surgical procedures.  J Cardiothorac Vasc Anesth  2001; 15:652-654.

Croughwell et al., 1993a. Croughwell N, Smith L, Quill T, et al: The effect of temperature on cerebral metabolism and blood flow in adults during cardiopulmonary bypass.  J Thorac Cardiovasc Surg  1993; 103:549-554.

Croughwell et al., 1993b. Croughwell N, Newman MF, Blumenthal JA, et al: Jugular venous saturation and cerebral arterial venous oxygen difference predict cognitive dysfunction after cardiac surgery.  Circulation  1993; 88(Suppl II):II-289.

Culliford et al., 1981. Culliford AT, Gitel SN, Starr N, et al: Lack of correlation between activated clotting time and plasma heparin during cardiopulmonary bypass.  Ann Surg  1981; 193:105.

Custer and Hales, 1985. Custer JR, Hales CA: Influence of alveolar oxygen on pulmonary vasoconstriction in newborn lambs vs sheep.  Am Rev Respir Dis  1985; 132:326.

Cyran et al., 1991. Cyran SE, Myers JL, Gleason MM, et al: Application of intraoperative transesophageal echocardiography in infants and small children.  J Cardiovasc Surg (Torino)  1991; 32:318-321.

Danforth et al., 1960. Danforth WH, Naegle S, Bing R: Effect of ischemia and reoxygenation on glycolytic reactions and adenosine triphosphate in heart muscle.  Circ Res  1960; 8:965-969.

Darling et al., 1994. Darling EM, Shearer IR, Nanry K, et al: Modified ultrafiltration in pediatric cardiopulmonary bypass.  J Extra Corpor Technol  1994; 26:205-209.

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

Darling et al., 2001. Darling EM, Kaemmer D, Lawson DS, et al: Use of ECMO without the oxygenator to provide ventricular support after Norwood stage I procedures.  Ann Thorac Surg  2001; 71:735-736.

Davidson et al., 2001. Davidson AJ, McCann M, Devavaram P, et al: The difference in the bispectral index between infants and children during emergence from anesthesia after circumcision surgery.  Anesth Analg  2001; 93:326-330.

Davis et al., 1987. Davis PJ, Cook DR, Stiller RL, et al: Pharmacodynamics and pharmacokinetics of high-dose sufentanil in infants and children undergoing cardiac surgery.  Anesth Analg  1987; 66:203-210.

Davis et al., 1999. Davis PJ, Wilson AS, Siewers RD, et al: The effects of cardiopulmonary bypass on remifentanil kinetics in children undergoing atrial septal defect repair.  Anesth Analg  1999; 89:904.

Deal et al., 1970. Deal CW, Warden JC, Monk I: Effect of hypothermia on lung compliance.  Thorax  1970; 25:105-109.

Decoodt et al., 1992. Decoodt P, Kacenelenbogen R, Bar JP, et al: Clinical usefulness of biplane transesophageal echocardiography.  Echocardiogr J Cardiovasc Ultrasound Allied Techn  1992; 9:257-264.

Dekeon et al., 1992. Dekeon M, Meldrum M, Meliones J, et al: Pulmonary function testing demonstrates airway obstruction secondary to cardiovascular abnormalities.  Pediatr Crit Care Colloquium  Jan 30, 1992;

DeLeon et al., 1983. DeLeon S, Idriss F, Ilbawi M, et al: The role of the Glenn shunt in patients undergoing the Fontan operation.  J Thorac Cardiovasc Surg  1983; 85:669-677.

Delise et al., 1976. Delise G, Ando M, Calder AL, et al: Total anomalous pulmonary veins connection: Report of 93 autopsied cases with emphasis on diagnostic and surgical considerations.  Am Heart J  1976; 91:99-122.

DiEusanio et al., 2003. DiEusanio M, Wesselink RM, Morshuis WJ, et al: Hypothermic circulatory arrest and antegrade selective cerebral perfusion during ascending aorta-hemiarch replacement: A retrospective comparison.  J Thorac Cardiovasc Surg  2003; 125:849-854.

Dickinson et al., 1982. Dickinson D, Wilkinson J, Smith A, et al: Variations in the morphology of the ventricular septal defect and disposition of the atrioventricular conduction tissues in tetralogy of Fallot.  Thorac Cardiovasc Surg  1982; 30:243-247.

Dietrich et al., 1992. Dietrich W, Barankay A, Hahnel C, et al: High-dose aprotinin in cardiac surgery: Three years' experience in 1,784 patients.  J Cardiothorac Vasc Anesth  1992; 6:324-327.

Dietrich et al., 1989. Dietrich W, Henze R, Barankay A, et al: High-dose aprotinin application reduces homologous blood requirement in cardiac surgery.  J Cardiothorac Anesth  1989; 3:79-84.

Dietrich et al., 1993. Dietrich W, Mossinger H, Spannagl M, et al: Hemostatic activation during cardiopulmonary bypass with different aprotinin dosages in pediatric patients having cardiac operations.  J Thorac Cardiovasc Surg  1993; 105:712-720.

Dietrich et al., 1990. Dietrich W, Spannagl M, Jochum M, et al: Influence of high-dose aprotinin treatment on blood loss and coagulation patterns in patients undergoing myocardial revascularization.  Anesthesiology  1990; 73:1119-1126.

Dietrich et al., 1991. Dietrich W, Spannagl M, Schramm W, et al: The influence of preoperative anticoagulation on heparin response during cardiopulmonary bypass.  J Thorac Cardiovasc Surg  1991; 102:505-514.

Domalik-Wawrzynski et al., 1987. Domalik-Wawrzynski LJ, Powell Jr WR, Guerrero L, et al: Effect of changes in ventricular relaxation on early diastolic coronary blood flow in canine hearts.  Circ Res  1987; 65:1751-1756.

Downing and Edmonds, 1992. Downing SW, Edmonds LH: Release of vasoactive substances during cardiopulmonary bypass.  Ann Thorac Surg  1992; 54:1236-1243.

Drummond et al., 1981. Drummond WH, Gregory GA, Heyman M, et al: The independent effects of hyperventilation, tolazoline, and dopamine in infants with persistent pulmonary hypertension.  J Pediatr  1981; 98:603-608.

Drummond and Lock, 1984. Drummond WH, Lock JE: Neonatal pulmonary vasodilator drugs.  Dev Pharmacol Ther  1984; 7:1.

Dubost et al., 1989. Dubost C, Deloche A, Carpentier A, et al: Endomyocardial fibrosis.  Chirurgie  1989; 115:689-703.

Duncan and Desai, 2003. Duncan BW, Desai S: Pulmonary arteriovenous malformations after cavopulmonary anastomosis.  Ann Thorac Surg  2003; 76:1759-1766.

Duncan et al., 2000. Duncan HP, Cloote A, Weir PM, et al: Reducing stress response in the prebypass phase of open heart surgery in infants and young children: A comparison of different fentanyl doses.  Br J Anesth  2000; 84:556-564.

El Azab et al., 2002. El Azab SR, Rosseel PM, deLange JJ, et al: Dexamethasone decreases the pro- to anti-inflammatory cytokine ratio during cardiac surgery.  Br J Anaesth  2002; 88:496-501.

Emmel and Sreeram, 2004. Emmel M, Sreeram N: Total anomalous pulmonary vein connection: Diagnosis, management and outcome.  Curr Treat Options Cardiovasc Med  2004; 5:423-439.

Engle and Perloff, 1983. Engle MA, Perloff JK: Congenital heart disease after surgery: Benefits, residua, and sequelae,  New York, Yorke Medical Books, 1983.

Erb et al., 2002a. Erb TO, Hall JM, Ing RJ, et al: Postoperative nausea and vomiting in children and adolescents undergoing radiofrequency catheter ablation: a randomized comparison of propofol and isoflurane-based anesthetics.  Anesth Analg  2002; 95:1577.

Erb et al., 2002b. Erb TO, Kanter RJ, Hall JM, et al: Comparison of electrophysiologic effects of propofol and isoflurane-based anesthetics in children undergoing radiofrequency catheter ablation for supraventricular tachycardia.  Anesthesiology  2002; 96:1386.

Ericinska and Silver, 1989. Ericinska M, Silver IA: ATP and brain function.  J Cereb Blood Flow Metab  1989; 9:2.

Fagan, 1977. Fagan DG: Shape changes in V-P loops for children's lungs related to growth.  Thorax  1977; 32:198.

Fillinger et al., 2002. Fillinger MP, Rassias AJ, Guyre PM, et al: Glucocorticoid effects on the inflammatory and clinical responses to cardiac surgery.  J Cardiothorac Vasc Anesth  2002; 16:163-169.

Fishberger et al., 1997. Fishberger SB, Wernovsky G, Gentles TL, et al: Factors that influence the development of atrial flutter after the Fontan operation.  J Thorac Cardiovasc Surg  1997; 113:80-86.

Fishberger et al., 2003. Fishberger SB, Wernovsky G, Gentles TL, et al: Factors that influence the development of atrial flutter after the Fontan operation.  J Thorac Cardiovasc Surg  2003; 11:70-74.

Fogel, May 5, 2005. Fogel MA: Is routine cardiac catheterization necessary in the management of patients with single ventricles across staged Fontan reconstruction? No.  Pediatr Cardiol  May 5, 2005;

Fontan et al., 1983. Fontan F, Deville C, Quagebeur J, et al: Repair of tricuspid atresia in 100 patients.  J Thorac Cardiovasc Surg  1983; 85:647.

Fox et al., 1982. Fox LS, Blackstone EH, Kirklin JW, et al: Relationship of whole body oxygen consumption to perfusion flow rate during hypothermic cardiopulmonary bypass.  J Thorac Cardiovasc Surg  1982; 83:239.

Fox et al., 1984a. Fox L, Blackstone E, Kirklin J, et al: Relationship of brain blood flow and oxygen consumption to perfusion flow rate during profound hypothermic cardiopulmonary bypass an experimental study.  J Thorac Cardiovasc Surg  1984; 87:658-664.

Fox et al., 1984b. Fox L, Blackstone E, Kirklin J, et al: Cerebral monitoring of somatosensory evoked potentials during profoundly hypothermic circulatory arrest.  Circulation  1984;

Fox et al., 1984c. Fox L, Blackstone E, Kirklin J, et al: Relationship of whole body oxygen consumption to perfusion flow rate during hypothermic cardiopulmonary bypass.  J Thorac Cardiovasc Surg  1984; 83:239-248.

Franciosi and Blanc, 1968. Franciosi RA, Blanc WA: Myocardial infarction in infants and children. I. A necropsy study in congenital heart disease.  J Pediatr  1968; 73:309.

Freeman and Crapo, 1982. Freeman BA, Crapo JD: Free radicals and tissue injury.  Lab Invest  1982; 47:412.

Frescura et al., 1987. Frescura C, Thiene G, Franceschini E, et al: Pulmonary vascular disease in infants with complete atrioventricular septal defect.  Int J Cardiol  1987; 15:91.

Friedman, 1972. Friedman W: The intrinsic properties of the developing heart.  Prog Cardiovasc Dis  1972; 15:87-111.

Friesen and Lichtor, 1983. Friesen RH, Lichtor JL: Cardiovascular effects of inhalational induction with isoflurane in infants.  Anesth Analg  1983; 62:411-417.

Friesen et al., 2003. Friesen RH, Veit AS, Archibald DJ, et al: A comparison of remifentanil and fentanyl for fast track paediatric cardiac anesthesia.  Paediatr Anaesth  2003; 13:122-125.

Fyler, 1980. Fyler D: Report of the New England Regional Infant Cardiac Program.  Pediatrics  1980; 65(Suppl):376.

Gaudet et al., 1980. Gaudet RJ, Alam I, Levine L: Accumulation of cyclooxygenase products of arachidonic acid metabolism in gerbil brain during reperfusion after bilateral common carotid artery occlusion.  J Neurochem  1980; 35:653-658.

Gaynor et al., 2003. Gaynor JW, Gerdes M, Zackai EH, et al: Apolipoprotein E genotype and neurodevelopment sequelae of infant cardiac surgery.  J Thorac Cardiovasc Surg  2003; 26:1736-1745.

Geha et al., 1973. Geha AS, Salaymeh MT, Abe T, et al: Pressure-dependent cerebral ischemia during cardiopulmonary bypass.  Neurology  1973; 23:521-529.

Ghanayem et al., 2003. Ghanayem NS, Hoffman GM, Mussatto KA, et al: Home surveillance program prevents interstage mortality after the Norwood procedure.  J Thorac Cardiovasc Surg  2003; 126:1257.

Giannuzzi et al., 1995. Giannuzzi CE, Seidler FJ, Slotkin TA: Beta-adrenoceptor control of cardiac adenylyl cyclase during development: agonist pretreatment in the neonate uniquely causes heterologous sensitization, not desensitization.  Brain Res  1995; 694:271.

Gibbs et al., 1993. Gibbs JL, Wren C, Watterson KG, et al: Stenting of the arterial ductus combined with banding of the pulmonary arteries and atrial septostomy: A new approach to palliation for the hypoplastic left heart syndrome.  Br Heart J  1993; 69:551-555.

Gildein et al., 1990. Gildein HP, Ahmadi A, Fontan F, et al: Special problems in Fontan-type operations for complex cardiac lesions.  Int J Cardiol  1990; 29:21-28.

Gillette et al., 1980. Gillette PC, Kugler JD, Garson AJ: Mechanisms of cardiac arrhythmias after the Mustard operation for transposition of the great arteries.  Am J Cardiol  1980; 45:1225-1229.

Gillette et al., 1977. Gillette PC, Yeoman M, Mullins C, et al: Sudden death after repair of tetralogy of Fallot. Electrocardiographic and electrophysiologic abnormalities.  Circulation  1977; 56:566.

Girard et al., 1991. Girard C, Lehot J, Clerc J, et al: Inhaled nitric oxide (NO) in pulmonary hypertension following mitral valve replacement.  Anesthesiology  1991; 75:A984.

Glass et al., 1997. Glass PS, Bloom M, Kearse L, et al: Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers.  Anesthesiology  1997; 86:836-847.

Godart et al., 2003. Godart F, Rey C, Devos P, et al: Transcatheter occlusion of moderate to large patent arterial ducts, having a diameter above 2.5 mm, with the Amplatzer duct occluder. Comparisons Rashkind buttoned devices and coils in 116 consecutive patients.  Cardiol Young  2003; 13:413.

Gold et al., 1986. Gold JP, Jonas RA, Lang P, et al: Transthoracic intracardiac monitoring lines in pediatric surgical patients: A ten year experience.  Ann Thorac Surg  1986; 42:185.

Goldberg et al., 2000. Goldberg CS, Schwartz EM, Brunberg JA, et al: Neurodevelopment outcome of patients after the Fontan operation: A comparison between children with hypoplastic left heart syndrome and other functional single ventricle lesions.  J Pediatr  2000; 137:646-652.

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

Graham, 1982. Graham T: Ventricular performance in adults after operation for congenital heart disease.  Am J Cardiol  1982; 50:612-620.

Graham, 1991. Graham TP: Ventricular performance in congenital heart disease.  Circulation  1991; 84:2259-2274.

Greeley et al., 1991a. Greeley WJ, Bracey VA, Ungerleider RM, et al: Recovery of cerebral metabolism and mitochondrial oxidation state is delayed after hypothermic circulatory arrest.  Circulation  1991; 84(Suppl III):III-400-III-415.

Greeley et al., 1986a. Greeley WJ, Bushman GA, Davis DP, et al: Comparative effects of halothane and ketamine on systemic arterial oxygen saturation in children with cyanotic heart disease.  Anesthesiology  1986; 65:666-668.

Greeley et al., 1988. Greeley WJ, Bushman G, Kong D, et al: Effects of cardiopulmonary bypass on eicosanoid metabolism during pediatric cardiovascular surgery.  J Thorac Cardiovasc Surg  1988; 95:842-849.

Greeley and de Bruijn, 1988. Greeley WJ, de Bruijn NP: Changes in sufentanil pharmacokinetics within the neonatal period.  Anesth Analg  1988; 67:86-90.

Greeley et al., 1987a. Greeley WJ, de Bruijn NP, Davis DP: Sufentanil pharmacokinetics in pediatric cardiovascular patients.  Anesth Analg  1987; 66:1067-1072.

Greeley et al., 1993. Greeley WJ, Kern FH, Meliones J, et al: Monitoring the brain during cardiac surgery in children [editorial; comment].  Can J Anaesth  1993; 40:291-297.

Greeley et al., 1991b. Greeley WJ, Kern FH, Ungerleider RM, et al: The effect of hypothermic cardiopulmonary bypass and total circulatory arrest on cerebral metabolism in neonates, infants, and children.  J Thorac Cardiovasc Surg  1991; 101:783-794.

Greeley et al., 1987b. Greeley WJ, Leslie JB, Reves JG: Prostaglandins and the cardiovascular system: A review and update.  J Cardiothorac Anesth  1987; 1:331-349.

Greeley et al., 1986b. Greeley WJ, Leslie JB, Reves JG, et al: Eicosanoids (prostaglandins) and the cardiovascular system.  J Cardiol Surg  1986; 1:357-378.

Greeley and Ungerleider, 1991. Greeley WJ, Ungerleider RM: Assessing the effect of cardiopulmonary bypass on the brain [editorial; comment].  Ann Thorac Surg  1991; 52:417-419.

Greeley et al., 1989b. Greeley WJ, Ungerleider RM, Kern FH, et al: Effects of cardiopulmonary bypass on cerebral blood flow in neonates, infants, and children.  Circulation  1989; 80:1209-1215.

Greeley et al., 1989a. Greeley WJ, Ungerleider RM, Smith L, et al: The effects of deep hypothermic cardiopulmonary bypass and total circulatory arrest on cerebral blood flow in infants and children.  J Thorac Cardiovasc Surg  1989; 97:737-745.

Gruber et al., 2001. Gruber EM, Laussen PC, Casta A, et al: Stress response in infants undergoing cardiac surgery: A randomized study of fentanyl bolus, fentanyl infusion, and fentanyl-midazolam infusion.  Anesth Analg  2001; 92:882-890.

Guitierrez, 1991. Guitierrez G: Cellular energy metabolism during hypoxia.  Crit Care Med  1991; 19:619-626.

Hagberg et al., 1987a. Hagberg J, Andersson P, Kjellmer I, et al: Extracellular overflow of glutamate, aspartate, GABA, and taurine in the cortex and basal ganglia of fetal lambs during hypoxia-ischemia.  Neurosci Lett  1987; 78:311-317.

Hagberg et al., 1987b. Hagberg J, Andersson P, Lacarewicz J, et al: Extracellular adenosine, inosine, hypoxanthine, and xanthine in relation to tissue nucleotides and purines in rat striatum during transient ischemia.  J Neurochem  1987; 49:227.

Hagler et al., 1988. Hagler DJ, Tajik AJ, Seward JB, et al: Intraoperative two-dimensional Doppler echocardiography. A preliminary study for congenital heart disease.  J Thorac Cardiovasc Surg  1988; 95:516-522.

Hannan et al., 1998. Hannan EL, Racz M, Kavey RE, et al: Pediatric cardiac surgery: The effect of hospital and surgeon volume on in-hospital mortality.  Pediatrics  1998; 101:963-969.

Hansen, 1985. Hansen AJ: Effect of anoxia on ion distribution in the brain.  Physiol Rev  1985; 65:101.

Hansen and Zeuthen, 1981. Hansen AJ, Zeuthen T: Extracellular ion concentrations during spreading depression and ischemia in the rat brain cortex.  Acta Physiol Scand  1981; 113:437.

Hansen and Hickey, 1986. Hansen D, Hickey P: Anesthesia for hypoplastic left heart syndrome: Use of high-dose fentanyl in 30 neonates.  Anesth Analg  1986; 65:127-132.

Hansen et al., 1984. Hansen N, Brubakk A, Bratlid D, et al: The effects of variations in PaCO2 on brain blood flow and cardiac output in the newborn piglet.  Pediatr Res  1984; 18:1132-1136.

Harker, 1986. Harker LA: Bleeding after cardiopulmonary bypass.  N Engl J Med  1986; 314:446-448.

Hausdorf et al., 1992. Hausdorf G, Friedel N, Berjis F, et al: Enoximone in newborns with refractory postoperative low output status.  Eur J Cardiothorac Surg  1992; 6:311-317.

Hawkins et al., 1989. Hawkins J, Van Hare GF, Schmidt KG: Effects of increasing afterload on left ventricular output in fetal lambs.  Circ Res  1989; 65:127-134.

Hellenbrand et al., 1990. Hellenbrand WE, Fahey JT, McGowan F, et al: Transesophageal echocardiographic guidance for transcatheter closure of atrial septal defects.  Am J Cardiol  1990; 66:207-213.

Henriksson et al., 1979. Henriksson P, Varendh G, Lundstrom NR: Haemostatic defects in cyanotic congenital heart disease.  Br Heart J  1979; 41:23-27.

Henzel and DeWeese, 1971. Henzel JH, DeWeese MS: Morbid and mortal complications associated with prolonged central venous cannulation. Awareness, recognition, and prevention.  Am J Surg  1971; 121:600-605.

Hickey and Andersen, 1987. Hickey P, Andersen N: Deep hypothermic circulatory arrest: A review of pathophysiology and clinical experience as a basis for anesthetic management.  J Cardiothorac Anesth  1987; 1:137-155.

Hickey and Hansen, 1984. Hickey P, Hansen D: Fentanyl and sufentanil-oxygen-pancuronium anesthesia for cardiac surgery in infants.  Anesth Analg  1984; 63:117-124.

Hickey et al., 1985a. Hickey PR, Hansen DD, Wessel DL, et al: Pulmonary and systemic hemodynamic responses to fentanyl in infants.  Anesth Analg  1985; 64:483-490.

Hickey et al., 1985b. Hickey PR, Hansen DD, Wessel DL, et al: Blunting of stress responses in the pulmonary circulation of infants by fentanyl.  Anesth Analg  1985; 64:1137-1143.

Hickey and Wessel, 1987. Hickey PR, Wessel DL: Anesthesia for treatment of congenital heart disease,  Miami, Grune & Stratton, 1987.

Hickey et al., 1992. Hickey PR, Wessel DL, Streitz SL, et al: Transcatheter closure of atrial septal defects: Hemodynamic complications and anesthetic management.  Anesth Analg  1992; 74:44-50.

Hillier et al., 1991. Hillier SC, Burrows FA, Bissonnette B, et al: Cerebral hemodynamics in neonates and infants undergoing cardiopulmonary bypass and profound hypothermic circulatory arrest: Assessment by transcranial Doppler sonography.  Anesth Analg  1991; 72:723-728.

Hindmarsh et al., 1984. Hindmarsh KW, Sankaran K, Watson V: Plasma beta-endorphin concentrations in neonates associated with acute stress.  Dev Pharmacol Ther  1984; 7:198.

Hines and Barash, 1987. Hines R, Barash PG: Right ventricular failure,  New York, Grune & Stratton, 1987.

Hiramatsu et al., 1995. Hiramatsu T, Miura T, Forbess JM, et al: pH strategies and cerebral energetics before and after circulatory arrest.  J Thorac Cardiovasc Surg  1995; 109:948-957.

Hiramatsu et al., 2002. Hiramatsu T, Imai Y, Kurosawa H, et al: Effects of dilutional and modified ultrafiltration in plasma endothelin-1 and pulmonary vascular resistance after the Fontan procedure.  Ann Thorac Surg  2002; 73:861-865.

Hoerter, 1976. Hoerter J: Changes in the sensitivity to hypoxia and glucose deprivation in the isolated perfused rabbit heart during perinatal development.  Pflugers Arch  1976; 303:1-6.

Hardman and Limbird, 2001. In: Hardman J, Limbird L, ed. Goodman and Gilman's pharmacological basis of therapeutics,  New York: McGraw-Hill; 2001:136-146.

Hoffman et al., 2003. Hoffman TM, Wernovsky G, Atz AM, et al: Efficacy and safety of milrinone in preventing low cardiac output syndrome in infants and children after corrective surgery for congenital heart disease.  Circulation  2003; 107:996-1002.

Horkay et al., 1992. Horkay F, Martin P, Rajah SM, et al: Response to heparinization in adults and children undergoing cardiac operations.  Ann Thorac Surg  1992; 53:822-826.

Horrow et al., 1990. Horrow JC, Hlavacek J, Strong MD, et al: Prophylactic tranexamic acid decreases bleeding after cardiac operations.  J Thorac Cardiovasc Surg  1990; 99:70-74.

Hug et al., 1994. Hug Jr CC, Burm AG, de Lange S: Alfentanil pharmacokinetics in cardiac surgical patients.  Anesth Analg  1994; 78:231.

Humes et al., 1987. Humes RA, Driscoll DJ, Danielson G, et al: Tetralogy of Fallot with anomalous origin of the left anterior descending coronary artery.  J Thorac Cardiovasc Surg  1987; 94:784-789.

Humpherys and Cummings, 1984. Humpherys JE, Cummings P: Atrial and ventricular tropomyosin and troponin-I in the developing bovine and human heart.  J Mol Cell Cardiol  1984; 16:643.

Hurwitz et al., 1980. Hurwitz RA, Smith W, King H, et al: Tetralogy of Fallot with abnormal coronary artery: 1967 to 1977.  J Thorac Cardiovasc Surg  1980; 80:129-134.

Ilbawi et al., 1987. Ilbawi MN, Idriss FS, DeLeon SY, et al: Factors that exaggerate the deleterious effects of pulmonary insufficiency on the right ventricle after tetralogy repair: Surgical implications.  J Thorac Cardiovasc Surg  1987; 93:36-41.

Ilke et al., 2003. Ilke L, Hale K, Fashaw L, et al: Developmental outcome of patients with hypoplastic left heart syndrome treated with heart transplant.  J Pediatr  2003; 142:20-25.

Imoto et al., 2001. Imoto Y, Kado H, Shiokawa Y, et al: Experience with the Norwood procedure without circulatory arrest.  J Thorac Cardiovasc Surg  2001; 122:879-882.

Jarmakani et al., 1971. Jarmakani JM, Graham TP, Canent R: The effect of corrective surgery on heart volume and mass in children with ventricular septal defect.  Am J Cardiol  1971; 27:254-258.

Jarmakani et al., 1972. Jarmakani JM, Graham TP, Canent R: Left ventricular contractile state in children with successfully corrected ventricular septal defect.  Circulation  1972;I-102-I-110.

Jarmakani et al., 1978. Jarmakani JM, Nagatomo T, Nakazawa M, et al: Effect of hypoxia on myocardial high-energy phosphates in the neonatal mammalian heart.  Am J Physiol  1978; 235:H475-H481.

Jegger et al., 2003. Jegger D, Tevaearai HT, Mueller XM, et al: Limitations using the vacuum-assist venous drainage technique during cardiopulmonary bypass procedures.  J Extra Corpor Technol  2003; 35:207-211.

Jenkins et al., 1985. Jenkins J, Lynn A, Edmonds J, et al: Effects of mechanical ventilation on cardiopulmonary function in children after open-heart surgery.  Crit Care Med  1985; 13:77-80.

Jennings and Reimer, 1983. Jennings R, Reimer K: Factors involved in salvaging ischemic myocardium: Effect of reperfusion of arterial blood.  Circulation  1983; 68(Suppl I):I-25-I-29.

Jensen et al., 2003. Jensen E, Andreasson S, Bengtsson A, et al: Influence of two different perfusion systems on inflammatory response in pediatric heart surgery.  Ann Thorac Surg  2003; 75:919-925.

Jobes et al., 1992. Jobes DR, Shaffer GW, Aitken GL: Heparin/protamine dosing guided by in vitro testing reduces blood loss and transfusion in cardiac surgery.  Anesthesiology  1992; 77:A137.

Jobes and Nicolson, 1988. Jobes JR, Nicolson SC: Monitoring of arterial hemoglobin oxygen saturation using a tongue sensor.  Anesth Analg  1988; 67:186-193.

Jonas, 1993. Jonas RA: Experimental studies of hypothermic circulatory arrest and low flow bypass.  Cardiol Young  1993; 3:299-307.

Jonas and Castaneda, 1988. Jonas RA, Castaneda AR: Modified Fontan procedure: Atrial baffle and systemic venous to pulmonary artery anastomotic techniques.  J Cardiol Surg  1988; 3:91-96.

Jonas et al., 2003. Jonas RA, Wypij D, Roth SJ, et al: The influence of hemodilution on outcome after hypothermic cardiopulmonary bypass: Results of a randomized trial in infants.  J Thorac Cardiovasc Surg  2003; 26:1765-1774.

Julia et al., 1990. Julia P, Kofsky E, Buckberg G, et al: Studies of myocardial protection in the immature heart. I. Enhanced tolerance of immature vs adult myocardium to global ischemia with reference to metabolic differences.  J Thorac Cardiovasc Surg  1990; 100:879-888.

Julia et al., 1991. Julia P, Kofsky E, Buckberg G, et al: Studies of myocardial protection in the immature heart. III. Models of ischemic and hypoxic/ischemic injury in the immature puppy heart.  J Thorac Cardiovasc Surg  1991; 101:14-19.

Julia et al., 1988. Julia P, Partington M, Buckberg G, et al: Superiority of blood cardioplegia in limiting reperfusion damage: Importance of endogenous oxygen free-radical scavengers in red blood cells.  Surg Forum  1988; 39:221.

Keenan et al., 2000. Keenan HT, Thiagarajan R, Stephens KE, et al: Pulmonary function after modified venovenous ultrafiltration in infants: A prospective, randomized trial.  J Thorac Cardiovasc Surg  2000; 119:501-505.

Keidan et al., 2003. Keidan I, Mishaly D, Berkenstadt H, et al: Combining low inspired oxygen and carbon dioxide during mechanical ventilation for the Norwood procedure.  Paediatr Anaesth  2003; 13:58-62.

Kern et al., 1991a. Kern F, Jonas R, Mayer J, et al: Conventional temperature monitoring is a poor correlate of efficient brain cooling.  Anesthesiology  1991; 75(Suppl):A57.

Kern et al., 1992a. Kern F, Morana N, Sears J, et al: Coagulation defects in neonates during CPB.  Ann Thorac Surg  1992; 54:541-546.

Kern et al., 1991b. Kern F, Ungerleider R, Quill T, et al: Cerebral blood flow response to changes in PaCO2 during hypothermic cardiopulmonary bypass in children.  J Thorac Cardiovasc Surg  1991; 101:618-622.

Kern et al., 1993. Kern F, Ungerleider R, Reves J, et al: The effect of altering pump flow rate on cerebral blood flow and cerebral metabolism in neonates, infants and children.  Ann Thorac Surg  1993; 56:1366-1372.

Kern et al., 1992b. Kern F, Jonas R, Mayer J, et al: Temperature monitoring during infant CPB: Does it predict efficient brain cooling?.  Ann Thorac Surg  1992; 54:749-754.

Kern et al., 1991c. Kern F, Ungerleider R, Jacobs J, et al: Computerized continuous infusion of intravenous anesthetic drugs during pediatric cardiac surgery.  Anesth Analg  1991; 72:487-492.

Kety, 1945. Kety SS: The determination of cerebral blood flow in man by the use of nitrous oxide in low concentrations.  Am J Physiol  1945; 143:53.

Kirklin et al., 1983. Kirklin J, Westaby S, Blackstone EH, et al: Complement and the damaging effects of cardiopulmonary bypass.  J Thorac Cardiovasc Surg  1983; 86:845-852.

Kirklin and Barratt-Boyes, 1986b. Kirklin JW, Barratt-Boyes BG: Hypothermia, circulatory arrest, and cardiopulmonary bypass,  New York, John Wiley & Sons., 1986.

Kirklin and Barratt-Boyes, 1986a. Kirklin J, Barratt-Boyes B: Cardiac surgery,  New York, Churchill Livingstone, 1986.

Kirshbom et al., 1995. Kirshbom PM, Skaryak LA, DiBernardo LR, et al: Effects of aortopulmonary collaterals on cerebral cooling and cerebral metabolic recovery after circulatory arrest.  Circulation  1995; 92(Suppl II):II-490-II-494.

Kloth and Baum, 2002. Kloth RL, Baum VC: Very early extubation in children after cardiac surgery.  Crit Care Med  2002; 30:787-791.

Knott-Craig et al., 2002. Knott-Craig CJ, Overholt ED, Ward KE, et al: Repair of Ebstein's anomaly in the symptomatic neonate: An evolution of technique with 7-year follow-up.  Ann Thorac Surg  2002; 73:1789-1792.

Kohli et al., 2001. Kohli A, Rao KS, Nachiappan M, et al: Pericardiectomy for constrictive pericarditis: Is prolonged inotropic support a bad omen?.  J Indian Med Assoc  2001; 99:499-501.

Kopf et al., 1992. Kopf GS, Kleinman CS, Hijazi ZM, et al: Fenestrated Fontan operation with delayed transcatheter closure of atrial septal defect—improved results in high-risk patients.  J Thorac Cardiovasc Surg  1992; 103:1039-1048.

Koren et al., 1986. Koren G, Goresky G, Crean P, et al: Unexpected alterations in fentanyl pharmacokinetics in children undergoing cardiac surgery: Age related or disease related?.  Dec Pharmacol Ther  1986; 9:183.

Kreselis et al., 1985. Kreselis D, Rocchini A, Rosenthal A, et al: Hemodynamic determinants of exercise induced ST segment depression in children with valvular aortic stenosis.  Am J Cardiol  1985; 55:1133-1136.

Kumar et al., 2003. Kumar SP, Rubinstein CS, Simsic JM, et al: Lateral tunnel versus extracardiac conduit Fontan procedure: A concurrent comparison.  Ann Thorac Surg  2003; 76:1389-1396.

Kuroczynski et al., 2003. Kuroczynski W, Kampmann C, Choi Y, et al: The Fontan-operation: From intra- to extracardiac procedure.  J Thorac Cardiovasc Surg  2003; 11:70-74.

Kussman et al., 2004. Kussman BD, Geva T, McGowan FX: Cardiovascular causes of airway compression.  Paediatr Anaesth  2004; 14:60.

Kwiatkowski and Manno, 1999. Kwiatkowski JL, Manno CS: Blood transfusion support in pediatric cardiovascular surgery.  Transfus Sci  1999; 21:63.

Lacour-Gayet et al., 2001. Lacour-Gayet F, Piot D, Zoghbi J, et al: Surgical management and indication of left ventricular retraining in arterial switch for transposition of the great arteries with intact ventricular septum.  Eur J Cardiothorac Surg  2001; 20:824-829.

Laishley et al., 1986. Laishley RS, Burrows FA, Lerman J, et al: Effects of anesthetic induction on oxygen saturation in cyanotic congenital heart disease.  Anesthesiology  1986; 65:673-679.

Lakier et al., 1974. Lakier J, Stanger P, Heymann M, et al: Tetralogy of Fallot with absent pulmonary valve. Natural history and hemodynamic considerations.  Circulation  1974; 50:167-173.

Lang et al., 1980. Lang P, Williams R, Norwood W: The hemodynamic effects of dopamine in infants after corrective cardiac surgery.  J Pediatr  1980; 96:630.

Langercrantz and Slotkin, 1986. Langercrantz H, Slotkin TA: The “stress” of being born.  Sci Am  1986; 1:100-108.

Langley et al., 2000. Langley SM, Chai PJ, Jaggers JJ, et al: Preoperative high dose methylprednisolone attenuates the cerebral response to deep hypothermic circulatory arrest.  Eur J Cardiothorac Surg  2000; 17:279-286.

Langley et al., 1999. Langley SM, Chai PJ, Miller SE, et al: Intermittent perfusion protects the brain during deep hypothermic circulatory arrest.  Ann Thorac Surg  1999; 68:4-12.

Lau et al., 1999. Lau CL, Posther KE, Stephenson GR, et al: Mini-circuit cardiopulmonary bypass with vacuum assisted venous drainage: Feasibility of an asanguineous prime in the neonate.  Perfusion  1999; 14:389-396.

Laussen et al., 2001. Laussen PC, Murphy JA, Zubrakowski D, et al: Bispectral index monitoring in children undergoing mild hypothermic cardiopulmonary bypass.  Paediatr Anaesth  2001; 11:567-573.

Lawless et al., 1988. Lawless S, Burckart G, Diven W: Amrinone pharmacokinetics in neonates and infants.  J Clin Pharmacol  1988; 28:283.

Lazar et al., 1980. Lazar H, Buckberg G, Managanaro A, et al: Reversal of ischemic damage with amino acid substrate enhancement during reperfusion.  J Thorac Surg  1980; 80:350-354.

Lazenby et al., 1981. Lazenby W, Ko W, Zelano J, et al: Effects of temperature and flow rate on regional blood flow and metabolism during cardiopulmonary bypass.  Ann Thorac Surg  1981; 54:449-459.

Lefkowitz, 1993. Lefkowitz RJ: G protein-coupled receptor kinases.  Cell  1993; 74:409-412.

Legato, 1975. Legato M: Ultrastructural changes during normal growth in the dog and rat ventricular myofiber. Developmental and physiological correlates of cardiac muscle,  New York: Raven Press; 1975:249.

Lerner et al., 2003. Lerner A, Dinardo JA, Comunale ME: Anesthetic management for repair of Ebstein's anomaly.  J Cardiothorac Vasc Anesth  2003; 17:232-235.

Lev and Eckner, 1964. Lev M, Eckner F: The pathologic anatomy of tetralogy of Fallot and its variants.  Dis Chest  1964; 45:251-261.

Levy et al., 2002. Levy JH, Bailey JM, Deeb GM: Intravenous milrinone in cardiac surgery.  Ann Thorac Surg  2002; 73:325-330.

Lillehei et al., 1999. Lillehei CW, Mayer Jr JE, Shamberger RC, et al: Pediatric lung transplantation and “lessons from Green Surgery.”.  Ann Thorac Surg  1999; 68(3 Suppl):S25-S27.

Lloyd et al., 1998. Lloyd TR, Rydberg A, Ludomirsky A, et al: Late fenestration closure in the hypoplastic left heart syndrome: Comparison of hemodynamic changes.  Am Heart J  1998; 136:302-306.

Lock et al., 1982. Lock J, Einjig S, Bass J, et al: Pulmonary vascular response to oxygen and its influence on operative results in children with ventricular septal defects.  Pediatr Cardiol  1982; 3:41.

Lock et al., 1989. Lock J, Rome J, Davis R, et al: Transatrial closure of atrial septal defects: Experimental studies.  Circulation  1989; 79:1091-1099.

Lodge et al., 1999. Lodge AJ, Chai PJ, Daggett CW, et al: Methylprednisolone reduces the inflammatory response to cardiopulmonary bypass in neonatal piglets: Timing of dose is important.  J Thorac Cardiovasc Surg  1999; 117:515-522.

Lundar et al., 1987. Lundar T, Lindberg H, Lindegaard KF, et al: Cerebral perfusion during major cardiac surgery in children.  Pediatr Cardiol  1987; 8:161-165.

Lyrene et al., 1985. Lyrene R, Welch K, Godoy G, et al: Alkalosis attenuates hypoxic pulmonary vasoconstriction in neonatal lambs.  Pediatr Res  1985; 19:126.

Maher et al., 2003. Maher KO, Pizarro C, Gidding SS, et al: Hemodynamic profile after the Norwood procedure with right ventricle to pulmonary artery conduit.  Circulation  2003; 108:782-784.

Mahle et al., 2001. Mahle WT, Cohen MS, Spray TL, et al: Atrioventricular valve regurgitation in patients with single ventricle: Impact of the bidirectional cavopulmonary anastomosis.  Ann Thorac Surg  2001; 72:831-835.

Mahle et al., 2003. Mahle WT, Cuadrado AR, Tam VK: Early experience with a modified Norwood procedure using right ventricle to pulmonary artery conduit.  Ann Thorac Surg  2003; 76:1084-1088.

Mahle et al., 2002. Mahle WT, Tavani F, Zimmerman RA, et al: An MRI study of neurologic injury before and after congenital heart surgery.  Circulation  2002; 106(Suppl):1109-1114.

Mahle et al., 2002. Mahle WT, McBride MG, Paridon SM: Exercise performance in tetralogy of Fallot: The impact of primary complete repair in infancy.  Pediatr Cardiol  2002; 23:224-229.

Mahle et al., 2000. Mahle WT, Spray TL, Wernovsky G, et al: Survival after reconstructive surgery for hypoplastic left heart syndrome: a 15-year experience from a single institution.  Circulation  2000; 102:III136.

Mair et al., 2003. Mair R, Tulzer G, Sames E, et al: Right ventricular to pulmonary artery conduit instead of modified Blalock-Taussig shunt improves postoperative hemodynamics in newborns after the Norwood operation.  J Thorac Cardiovasc Surg  2003; 126:1378-1384.

Malhotra et al., 2003. Malhotra SP, Thelitz S, Riemer RK, et al: Fetal myocardial protection is markedly improved by reduced cardioplegic calcium content.  Ann Thorac Surg  2003; 75:1937-1941.

Malenka et al., 1999. Malenka DJ, McGirth PD, Wennberg DE, et al: The relationship between operator volume and outcomes after percutaneous coronary interventions in high volume hospitals in 1994–1996: The northern New England experience. Northern New England Cardiovascular Diseases Study Group.  J Am Coll Cardiol  1999; 34:1471-1480.

Malik and Kidd, 1973. Malik A, Kidd S: Independent effects of changes in H+ and CO2 concentrations on hypoxic pulmonary vasoconstriction.  J Appl Physiol  1973; 34:318-323.

Malviya et al., 2003. Malviya S, Voepel-Lewis T, Siewer M, et al: Risk factors for adverse postoperative outcomes in children presenting for cardiac surgery with upper respiratory tract infections.  Anesthesiology  2003; 98:628-632.

Malviya et al., 1989. Malviya S, Burrows FA, Johnston A, et al: Anaesthetic experience with paediatric interventional cardiology.  Can J Anaesth  1989; 36:320-324.

Mann et al., 2002. Mann K, Berg RA, Nadkarni V: Beneficial effects of vasopressin in prolonged pediatric cardiac arrest: a case series.  Resuscitation  2002; 52:149.

Manno et al., 1991. Manno CS, Hedberg KW, Kim HC, et al: Comparison of the hemostatic effects of fresh whole blood, stored whole blood, and components after open heart surgery in children.  Blood  1991; 77:930-936.

Mansell and Bryan, 1972. Mansell A, Bryan A: Airway closure in children.  J Appl Physiol  1972; 33:711.

Marshall et al., 2003. Marshall AC, Love BA, Lang P, et al: Staged repair of tetralogy of Fallot and diminutive pulmonary arteries with a fenestrated ventricular septal patch.  J Thorac Cardiovasc Surg  2003; 126:1427-1433.

Martindale et al., 1996. Martindale SJ, Shayevitz JR, E'Errico D: The activated coagulation time: suitability for monitoring heparin effect and neutralization during pediatric cardiac surgery.  J Cardiothorac Vasc Anesth  1996; 10:458.

Mass et al., 1997. Mass C, Kok R, Segers P, et al: Intermittent antegrade/selective cerebral perfusion during circulatory arrest for repair of the aortic arch.  Perfusion  1997; 12:127-132.

Masuda et al., 1998. Masuda H, Moriyama Y, Yamaoka A, et al: Preoperative autologous donation of blood in cardiac surgery—age related factors.  Jpn J Thorac Cardiovasc Surg  1998; 46:267.

Matsota et al., 2001. Matsota P, Livanios S, Marinopoulou E: Intercostal nerve block with bupivacaine for post-thoracotomy pain relief in children.  Eur J Pediatr Surg  2001; 11:219-222.

Matsumato et al., 1971. Matsumato T, Wolferth CJ, Perlman M: Effects of pulsatile and non-pulsatile perfusion upon cerebral and conjunctival microcirculation in the dog.  Am Surg  1971; 37:61-64.

Mault et al., 1993. Mault J, Ohtake S, Klingensmith M, et al: Cerebral metabolism and circulatory arrest: Effects of duration and strategies for protection.  Ann Thorac Surg  1993; 55:57-63.

Mault et al., 1992. Mault J, Whitaker E, Heinle J, et al: Intermittent perfusion during hypothermic circulatory arrest: A new and effective technique for cerebral protection,  New York, American College of Surgeons Clinical Congress, Surgical Forum, 1992.

Mavroudis et al., 1984. Mavroudis C, Brown G, Katzmark S, et al: Blood flow distribution in infant pigs subjected to surface cooling, deep hypothermia, and circulatory arrest.  J Thorac Cardiovasc Surg  1984; 87:665-672.

Mayer et al., 1986. Mayer J, Helgason H, Jonas R, et al: Extending the limits of modified Fontan procedures.  J Thorac Cardiovasc Surg  1986; 92:1021-1028.

Mayer et al., 1992. Mayer JEJ, Bridges ND, Lock JE, et al: Factors associated with marked reduction in mortality for Fontan operations in patients with single ventricle.  J Thorac Cardiovasc Surg  1992; 103:444-451.

Mazzera et al., 1989. Mazzera E, Corno A, Picardo S, et al: Bidirectional cavopulmonary shunts: Clinical applications as a staged or definitive palliation.  Ann Thorac Surg  1989; 47:415-420.

McGoon and Mair, 1990. McGoon DC, Mair DD: On the unmuddling of shunting, mixing and streaming.  J Thorac Cardiovasc Surg  1990; 100:77.

McGovern et al., 2000. McGovern JJ, Cheifetz IM, Craig DM, et al: Right ventricular injury in young swine: Effects of catecholamines on right ventricular function and pulmonary vascular mechanics.  Pediatr Res  2000; 48:763-769.

McGrath and Gonzales-Lavin, 1987. McGrath L, Gonzales-Lavin L: Actuarial survival, freedom from reoperation and other events after repair of atrioventricular septal defects.  Thorac Cardiovasc Surg  1987; 94:582-586.

Mehta et al., 2000. Mehta PA, Cunningham CK, Colella CB, et al: Risk factors for sternal wound and other infections in pediatric cardiac surgery patients.  Pediatr Infect Dis J  2000; 19:1000-1004.

Meliones et al., 1991. Meliones J, Bove E, Dekeon M, et al: High-frequency jet ventilation improves cardiac function after the Fontan procedure.  Circulation  1991; 84(Suppl III):III-364-III-368.

Meyer et al., 1972. Meyer R, Schwartz D, Benzing G, et al: Ventricular septum in right ventricular volume overload.  Am J Cardiol  1972; 30:349-354.

Michelakis et al., 2002. Michelakis E, Tymchak W, Lien D, et al: Oral sildenafil is an effective and specific pulmonary vasodilator in patients with pulmonary arterial hypertension: Comparison with inhaled nitric oxide.  Circulation  2002; 105:2398-2403.

Michenfelder, 1988. Michenfelder J: The hypothermic brain. Anesthesia and the brain,  New York: Churchill Livingstone; 1988:23-34.

Michenfelder and Theye, 1968. Michenfelder J, Theye R: Hypothermia: Effect of canine brain and whole-body metabolism.  Anesthesiology  1968; 29:1107.

Mies and Paschen, 1984. Mies G, Paschen W: Regional changes of blood flow, glucose, and ATP content determined on brain sections during a single passage of spreading depression in rat brain cortex.  Exp Neurol  1984; 84:249.

Mietus-Snyder et al., 1987. Mietus-Snyder M, Lang P, Mayer J, et al: Childhood systemic-pulmonary shunts: Subsequent suitability of Fontan operation.  Circulation  1987; 76(Suppl III):III-39-III-46.

Milde, 1992. Milde LN: Clinical use of mild hypothermia for brain protection—A dream revisited.  J Neurosurg Anesthesiol  1992; 4:211-215.

Miller et al., 1997. Miller BE, Mochizuki T, Levy JH, et al: Predicting and treating coagulopathies after cardiopulmonary bypass in children.  Anesth Analg  1997; 85:1196.

Mohiddin and Fananapazir, 2002. Mohiddin SA, Fananapazir L: Systolic compression of epicardial coronary and intramural arteries in children with hypertrophic cardiomyopathy.  Tex Heart Inst J  2002; 29:290.

Miyamoto et al., 1986. Miyamoto K, Kawashima Y, Matsuda H, et al: Optimal perfusion flow rate for the brain during deep hypothermic cardiopulmonary bypass at 20°C.  J Thora c Cardiovasc Surg  1986; 92:1065-1070.

Moore, 1981. Moore R: Anesthesia for the pediatric congenital heart patient for noncardiac surgery.  Anesthesiol Rev  1981; 8:23.

Moore et al., 1985. Moore RA, Yang SS, McNicholas KW, et al: Hemodynamic and anesthetic effects of sufentanil as the sole anesthetic for pediatric cardiovascular surgery.  Anesthesiology  1985; 63:725-732.

Mori et al., 1987. Mori A, Tabata R, Nakamura Y, et al: Effects of pulsatile cardiopulmonary bypass on carbohydrate and lipid metabolism.  J Cardiovasc Surg  1987; 28:621-626.

Mortier et al., 1996. Mortier E, Ongenae M, Vermassen F, et al: Operative closure of patent ductus arteriosus in the neonatal intensive care unit.  Acta Chir Belg  1996; 96:266.

Moskowitz et al., 1984. Moskowitz M, Kiwak K, Hekimian K: Synthesis of compounds with properties of leukotrienes C4 and D4 in gerbil brains after ischemia and reperfusion.  Science  1984; 224:886.

Mochizuki et al., 1999. Mochizuki Y, Patel AK, Banerjee A, et al: Intraoperative transesophageal echocardiography: Correlation of echocardiographic findings and surgical pathology.  Cardiol Rev  1999; 7:270-276.

Mossinger et al., 2003. Mossinger H, Kietrich W, Braun SL, et al: High-dose aprotinin reduces activation of hemostasis, allogeneic blood requirement, and duration of postoperative ventilation in pediatric cardiac surgery.  Ann Thorac Surg  2003; 75:430.

Mott et al., 2001. Mott AR, Spray TL, Gaynor JW, et al: Improved early results with cavopulmonary connections.  Cardiol Young  2001; 11:3-11.

Muhiudeen et al., 1991. Muhiudeen IA, Kuecherer HF, Lee E, et al: Intraoperative estimation of cardiac output by transesophageal pulsed Doppler echocardiography.  Anesthesiology  1991; 74:9-14.

Muhiudeen et al., 1990. Muhiudeen IA, Roberson DA, Silverman NH, et al: Intraoperative echocardiography in infants and children with congenital cardiac shunt lesions: Transesophageal versus epicardial echocardiography.  J Am Coll Cardiol  1990; 16:1687-1695.

Muhiudeen et al., 1992. Muhiudeen IA, Roberson DA, Silverman NH, et al: Intraoperative echocardiography for evaluation of congenital heart defects in infants and children.  Anesthesiology  1992; 76:165-172.

Muller et al., 2003. Muller M, Akinturk H, Schindler E, et al: Combined state 1 and 2 repair for hypoplastic left heart syndrome: Anaesthetic considerations.  Paediatr Anaesth  2003; 13:360-365.

Mullins, 1989. Mullins C: Pediatric and congenital therapeutic cardiac catheterization.  Circulation  1989; 79:1153-1159.

Murkin et al., 1987. Murkin JM, Farrar JK, Tweed WA, et al: Cerebral autoregulation and flow/metabolism coupling during cardiopulmonary bypass: The influence of PaCO2.  Anesth Analg  1987; 66:825-832.

Murray et al., 1986. Murray A, Glaria AP, Pearson DT, et al: The protective effect of profound hypothermia on the canine central nervous system during one hour of circulatory arrest.  Ann Thorac Surg  1986; 41:255-259.

Mutch and Hansen, 1984. Mutch W, Hansen AJ: Extracellular pH changes during spreading depression and cerebral ischemia: Mechanisms of brain pH regulation.  J Cereb Blood Flow Metab  1984; 4:17.

Nadroo et al., 2002. Nadroo AM, Glass RB, Lin J, et al: Changes in upper extremity position cause migration of peripherally inserted central catheters in neonates.  Pediatrics  2002; 110:131.

Naik et al., 1991. Naik SK, Knight A, Elliott M: A prospective randomized study of a modified technique of ultrafiltration during pediatric open-heart surgery.  Circulation  1991; 84(Suppl III):III-422-III-431.

Nagy et al., 2003. Nagy ZL, Collins M, Sharpe T, et al: Effect of two different bypass techniques on the serum troponin-T levels in newborns and children: Does pH-stat provide better protection?.  Circulation  2003; 108:577-582.

Nakanishi et al., 1987. Nakanishi T, Seguchi M, Takao A: Intracellular calcium concentrations in the newborn myocardium.  Circulation  1987; 76(Suppl IV):IV-455-IV-461.

Nasser et al., 1987. Nasser R, Reedy MC, Anderson P: Developmental changes in the ultrastructure and sarcomere shortening of the isolated rabbit myocardium.  Circ Res  1987; 61:465.

Nelson, 1966. Nelson N: Neonatal pulmonary function.  Pediatr Clin North Am  1966; 13:769.

Newburger et al., 1993. Newburger JW, Jonas RA, Wernovsky G, et al: A comparison of the perioperative neurologic effects of hypothermic circulatory arrest versus low-flow cardiopulmonary bypass in infant heart surgery.  N Engl J Med  1993; 329:329-336.

Newburger et al., 2003. Newburger JW, Wypij D, Bellinger DC, et al: Length of stay after heart surgery is related to cognitive outcome at age 8 years.  J Pediatr  2003; 143:67-73.

Ng et al., 2004. Ng C, Franklin O, Vaidya M, et al: Adenosine infusion for the management of persistent pulmonary hypertension of the newborn.  Pediatr Crit Care Med  2004; 5:10-13.

Ng et al., 2001. Ng PC, Lam CW, Fok T, et al: Refractory hypotension in preterm infants with adrenocortical insufficiency.  Arch Dis Child Fetal Neonatal Educ  2001; 84:F122-F124.

Nishioka et al., 1981. Nishioka K, Kamiya T, Ueda T, et al: Left ventricular volume characteristics in children with tricuspid atresia before and after surgery.  Am J Cardiol  1981; 47:1105-1107.

Nishinaka et al., 2002. Nishinaka T, Tatsumi E, Taenaka Y, et al: At least thirty-four days of animal continuous perfusion by a newly developed extracorporeal membrane oxygenation system without systemic anticoagulants.  Artif Organs  2002; 26:548-551.

Nollert et al., 1999. Nollert G, Nagashima M, Bucerius J, et al: Oxygenation strategy and neurologic damage after deep hypothermic circulatory arrest. II. Hypoxic versus free radical injury.  J Thorac Cardiovasc Surg  1999; 117:1172-1179.

Normal, 1978. Normal MG: Perinatal brain damage.  Perspect Pediatric Pathol  1978; 4:41-92.

Norstrom and Siesko, 1978. Norstrom C, Siesko B: Influence of phenobarbital on changes in the metabolites of the energy reserve of the cerebral cortex following complete ischemia.  Acta Physiol Scand  1978; 104:271.

Norwood, 1991. Norwood W: Hypoplastic left heart syndrome.  Ann Thorac Surg  1991; 52:688-695.

Norwood et al., 1983. Norwood W, Lang P, Hansen D: Physiologic repair of aortic atresia-hypoplastic left syndrome.  N Engl J Med  1983; 308:23.

Norwood et al., 1979a. Norwood W, Norwood C, Castaneda A: Cerebral anoxia: Effect of deep hypothermia and pH.  Surgery  1979; 86:203-209.

Norwood et al., 1979b. Norwood W, Norwood C, Ingwall J, et al: Hypothermic circulatory arrest: 31-Phosphorus nuclear magnetic resonance of isolated perfused neonatal rat brain.  J Thorac Cardiovasc Surg  1979; 78:823-830.

Norwood and Murphy, 1990. Norwood W, Murphy J: Hypoplastic left heart syndrome.   In: Sabiston DC, Spencer FC, ed. Surgery of the chest,  5th ed.. Philadelphia, PA: WB Saunders; 1990:1493-1502.

Norwood et al., 1992. Norwood WIJ, Jacobs MI, Murphy JD: Fontan procedure for hypoplastic left heart syndrome.  Ann Thorac Surg  1992; 54:1025-1029.

Ogata et al., 1960. Ogata T, Ida Y, Nonoyama A, et al: A comparative study of the effectiveness of pulsatile and nonpulsatile flow in extracorporeal circulation.  Arch Jpn Clin  1960; 29:59.

Okada et al., 1988. Okada Y, Tanimoto M, Yoneda K: The protective effect of hypothermia on reversibility in the neuronal function of the hippocampal slice during long lasting anoxia.  Neurosci Lett  1988; 84:277-282.

Ootaki et al., 2002. Ootaki Y, Yamaguchi M, Oshima Y, et al: Effects of modified ultrafiltration on coagulation factors in pediatric cardiac surgery.  Surg Today  2002; 32:203-206.

Olsson et al., 2000. Olsson C, Siegbahn A, Henze A, et al: Heparin-coated cardiopulmonary bypass circuits reduce circulating complement factors and interleukin-6 in paediatric heart surgery.  Scand Cardiovasc J  2000; 34:33-40.

Oshita et al., 1989. Oshita S, Uchimoto R, Oka H, et al: Correlation between arterial blood pressure and oxygenation in tetralogy of Fallot.  J Cardiovasc Anesth  1989; 3:597-600.

Ozawa et al., 2000. Ozawa T, Yoshihara K, Koyama N, et al: Clinical efficacy of heparin-bonded bypass circuits related to cytokine responses in children.  Ann Thorac Surg  2000; 69:584-590.

Pacifico et al., 1990. Pacifico A, Kirklin J, Kirklin J: Surgical treatment of ventricular septal defect.  Surgery of the chest,  5th ed.. Philadelphia, PA: WB Saunders; 1990:1314-1331.

Pacifico et al., 1987. Pacifico AD, Sand ME, Bargeron LM, et al: Transatrial-transpulmonary repair of tetralogy of Fallot.  J Thorac Cardiovasc Surg  1987; 93:919-926.

Padayachee et al., 1987. Padayachee TS, Parsons S, Theobold R, et al: The detection of microemboli in the middle cerebral artery during cardiopulmonary bypass: A transcranial Doppler ultrasound investigation using membrane and bubble oxygenators.  Ann Thorac Surg  1987; 44:298-302.

Pearl et al., 1991. Pearl J, Laks H, Drinkwater D, et al: Repair of truncus arteriosus in infancy.  Ann Thorac Surg  1991; 52:780-786.

Pearl, 2003. Pearl JM: Right ventricular-pulmonary artery connection in Stage 1 palliation of hypoplastic left heart syndrome.  J Thorac Cardiovasc Surg  2003; 126:1268-1270.

Pearl et al., 2002. Pearl JM, Nelson DP, Schwartz SM, et al: Glucocorticoids reduce ischemic-reperfusion induced myocardial apoptosis in immature hearts.  Ann Thorac Surg  2002; 74:830-836.

Perloff, 1982. Perloff J: Development and regression of increased ventricular mass.  Am J Cardiol  1982; 50:605.

Picardo et al., 1995. Picardo S, Testa G, La Vigna G, et al: Post-thoracotomy analgesia in pediatric heart surgery: Comparison of 2 different techniques.  Minerva Anesthesiol  1995; 61:277-282.

Pigott et al., 1988. Pigott J, Murphy J, Barber G, et al: Palliative reconstructive surgery for hypoplastic left heart syndrome.  Ann Thorac Surg  1988; 45:122-128.

Pizarro et al., 2003. Pizarro C, Malec E, Maher KO, et al: Right ventricle to pulmonary artery conduit improves outcome after stage 1 Norwood for hypoplastic left heart syndrome.  Circulation  2003; 108(Suppl III):III-55-III-60.

Pizarro et al., 2004. Pizarro C, Mroczek T, Malec E, Norwood WI: Right ventricle to pulmonary artery conduit reduces interim mortality after Stage 1 Norwood for hypoplastic left heart syndrome.  Ann Thorac Surg  2004; 78:1959.

Pizarro and Norwood, 2003. Pizarro C, Norwood WI: Right ventricle to pulmonary artery conduit has a favorable impact on postoperative physiology after Stage I Norwood: preliminary results.  Eur J Cardiothorac Surg  2003; 23:991.

Portman et al., 2000. Portman MA, Fearneyhough C, Ning XH, et al: Triiodothyronine repletion in infants during cardiopulmonary bypass for congenital heart disease.  J Thorac Cardiovasc Surg  2000; 120:604-608.

Pozzi et al., 2000. Pozzi M, Trivedi DB, Kitchner D, et al: Tetralogy of Fallot: What operation, at which age?.  Eur J Cardiothorac Surg  2000; 17:631-636.

Quattlebaum et al., 1976. Quattlebaum T, Varghese J, Barber G, et al: Sudden death among postoperative patients with tetralogy of Fallot.  Circulation  1976; 54:671.

Rabinovitch et al., 1981. Rabinovitch M, Herrera-DeLeon V, Castaneda A, et al: Growth and development of the pulmonary vascular bed in patients with tetralogy of Fallot with and without pulmonary atresia.  Circulation  1981; 64:1234.

Radford, 1985. Radford I: The level of induced DNA double strand breakage correlates with cell killing after x-radiation.  Int J Radiat Biol  1985; 48:45.

Ramakrishnan and Kothari, 2004. Ramakrishnan S, Kothari SS: Preoperative balloon dilation of obstructed total anomalous pulmonary venous connection in a neonate.  Cathet Cardiovasc Interv  2004; 61:128-130.

Rappaport et al., 1998. Rappaport LA, Wypij D, Bellinger DC, et al: Relation of seizures after cardiac surgery in early infancy to neurodevelopmental outcome. Boston Circulatory Arrest Study Group.  Circulation  1998; 97:773-779.

Rebeyka et al., 1990. Rebeyka I, Yeh TJ, Hanan S, et al: Altered contractile response in neonatal myocardium to citrate-phosphate-dextrose infusion.  Circulation  1990; 82(Suppl IV):IV-367-IV-370.

Reddy et al., 1999. Reddy VM, McElhinney DB, Sagrado T, et al: Results of 102 cases of complete repair of congenital heart defects in patients weighing 700 to 2500 grams.  J Thorac Cardiovasc Surg  1999; 117:324-331.

Reddy and Hanley, 2000. Reddy VM, Hanley FI: Cardiac surgery in infants with very low birth weight.  Semin Pediatr Surg  2000; 9:91-95.

Reid et al., 1997. Reid RW, Zimmerman A, Laussen PC, et al: The efficacy of tranexamic acid versus placebo in decreasing blood loss in pediatric patients undergoing repeat cardiac surgery.  Anesth Analg  1997; 84:990.

Reinhartz et al., 2003. Reinhartz O, Copeland JG, Farrar DJ: Thoratec ventricular assist devices in children with less than 1.3 m2 of body surface area.  ASAIO J  2003; 49:727-730.

Rich and Langer, 1982. Rich T, Langer G: Calcium depletion in rabbit myocardium: Calcium paradox protection by hypothermia and cation substitution.  Circ Res  1982; 51:131-141.

Richardson, 1996. Richardson P: Task Force on the Definition and Classification of the Cardiomyopathies.  Circulation  1996; 93:41.

Richardson and Clark, 1976. Richardson JP, Clark CP: Tetralogy of Fallot: Risk factors associated with complete repair.  Br Heart J  1976; 38:926-932.

Rigby-Jones et al., 2002. Rigby-Jones AE, Nolan JA, Priston MJ, et al: Pharmacokinetics of propofol infusions in critically ill neonates, infants, and children in an intensive care unit.  Anesthesiology  2002; 97:1393-1400.

Rinne and Kaukinen, 1998. Rinne T, Kaukinen S: Does lidocaine protect the heart during coronary revascularization?.  Acta Anaesthesiol Scand  1998; 42:936-940.

Riordan et al., 1996. Riordan CJ, Randsbaek F, Storey JH, et al: Inotropes in the hypoplastic left heart syndrome: Effects in an animal model.  Ann Thorac Surg  1996; 62:83-90.

Roberts et al., 1993a. Roberts J, Chen T, Kawai N, et al: Inhaled nitric oxide reverses pulmonary vasoconstriction in the hypoxic and acidotic newborn lamb.  Circ Res  1993; 72:246-254.

Roberts et al., 1993b. Roberts J, Lang P, Bigatello L, et al: Inhaled nitric oxide in congenital heart disease.  Circulation  1993; 87:447-453.

Roodpeyma and Sadeghian, 2000. Roodpeyma S, Sadeghian N: Acute pericarditis in children: A 10-year experience.  Pediatr Cardiol  2000; 21:363-367.

Rosenberg et al., 1982. Rosenberg A, Jones M, Traystman R, et al: Response of cerebral blood flow to changes in PCO2 in fetal, newborn and adult sheep.  Am J Physiol  1982; 242:H862-H866.

Rosenzweig et al., 1999. Rosenzweig EB, Starc TJ, Chen JM, et al: Intravenous arginine-vasopressin children with vasodilatory shock after cardiac surgery.  Circulation  1999; 100(Suppl II):II-182-II-186.

Rothman et al., 1990. Rothman A, Perry S, Keane J, et al: Early results and follow-up of balloon angioplasty for branch pulmonary artery stenoses.  J Am Coll Cardiol  1990; 15:1109-1117.

Royston et al., 1987. Royston D, Taylor KM, Bidstrup BP, et al: Effect of aprotinin on need for blood transfusion after repeat open-heart surgery.  Lancet  1987; 2:1289-1291.

Rozencrantz et al., 1982. Rozencrantz E, Okamoto F, Buckberg G, et al: Advantages of glutamate-enriched cold blood cardioplegia in energy-depleted hearts.  Circulation  1982; 66(Suppl II):II-151-II-156.

Rudolph, 2001. In: Rudolph AM, ed. Congenital diseases of the heart: clinical-physiological considerations,  2nd ed.. New York Futura: Armonk; 2001.

Rudolph, 1985. Rudolph A: Distribution and regulation of blood flow in the fetal and newborn lamb.  Circ Res  1985; 57:811.

Rudolph and Yuan, 1966. Rudolph A, Yuan S: Response of the pulmonary vasculature to hypoxia and H+ ion concentration changes.  J Clin Invest  1966; 45:399.

Russell et al., 2001. Russell IA, Miller Hance WC, Gregory G, et al: The safety and efficacy of sevoflurane anesthesia in infants and children with congenital heart disease.  Anesth Analg  2001; 92:1152-1158.

Rychik et al., 1991. Rychik J, Norwood W, Chil A: Doppler color flow mapping assessment of residual shunt after closure of large ventricular septal defects.  Circulation  1991; 84(Suppl III):III-153-III-161.

Sabiston et al., 1955. Sabiston DJ, Theilen E, Gregg D: The relationship of coronary blood flow and cardiac output and other parameters in hypothermia.  Surgery  1955; 38:498-505.

Sahn, 1985. Sahn DJ: Real-time two-dimensional Doppler echocardiographic flow mapping.  Circulation  1985; 71:849-853.

Salem and Griffin, 1981. Salem M, Griffin A: Anesthetic considerations in patients with congenital heart disease for noncardiac surgery.  Pediatric anesthesia: Current practice,  New York, Academic Press, 1981.

Salem et al., 1977. Salem M, Toyama T, Wong A, et al: Hemodynamic response to antagonism of tubocurarine block with atrophine-neostigmine mixture in pediatric patients.  Br J Anaesth  1977; 49:901.

Salem et al., 1970. Salem M, Ylagan L, Angel J, et al: Reversal of curarization with atrophine-neostigmine mixture in patients with congenital heart disease.  Br J Anaesth  1970; 42:991.

Saltzman et al., 1986. Saltzman E, Weinstein M, Weintraub R, et al: Treatment with desmopressin acetate to reduce blood loss after cardiac surgery: A double-blind randomized trial.  N Engl J Med  1986; 314:1402-1406.

Sanders et al., 1982. Sanders S, Wright G, Keane J, et al: Clinical and hemodynamic results of the Fontan operation for tricuspid atresia.  Am J Cardiol  1982; 49:1733-1740.

Sandhu and Pettitt, 2002. Sandhu SK, Pettitt TW: Interrupted aortic arch.  Curr Treat Options Cardiovasc Med  2002;337-340.

Sano et al., 2002. Sano S, Ishino K, Kawanda M, et al: Total right ventricular exclusion procedure: An operation for isolated congestive right ventricular failure.  J Thorac Cardiovasc Surg  2002; 123:640-647.

Sano et al., 2003. Sano S, Ishino K, Kawada M, et al: Right ventricle-pulmonary artery shunt in first-stage palliation of hypoplastic left heart syndrome.  J Thorac Cardiovasc Surg  2003; 126:504-509.

Schechter et al., 1995. Schechter WS, Kim C, Martinez M, et al: Anesthetic induction in a child with end-stage cardiomyopathy.  Can J Anesth  1995; 42:404-408.

Scheinberg and Stead, 1949. Scheinberg P, Stead EA: The cerebral blood flow in male subjects as measured by the nitrous technique: Normal values for blood flow, oxygen utilization, glucose utilization, and peripheral resistance with observations on the effect of lilting and anxiety.  J Clin Invest  1949; 28:1163-1168.

Schindler et al., 2003. Schindler E, Muller M, Kwapisz M, et al: Ventricular cardiac-assist devices in infants and children: Anesthetic considerations.  J Cardiothorac Vasc Anesth  2003; 17:617-621.

Schmid et al., 1999. Schmid ER, Burki C, Engel MH, et al: Inhaled nitric oxide versus intravenous vasodilators in severe pulmonary hypertension after cardiac surgery.  Anesth Analg  1999; 89:1108-1115.

Schrantz and Droege, 1993. Schrantz D, Droege A, et al: Uncoupling of human cardiac beta-adrenoceptors during cardiopulmonary bypass with cardioplegic cardiac arrest.  Circulation  1993; 87:422-426.

Schrantz et al., 1990. Schrantz D, Oelert H, Iverson S, et al: Congenital mitral regurgitation caused by a perforation in the anterior leaflet.  Pediatr Cardiol  1990; 11:93-95.

Schreiber et al., 2000. Schreiber C, Eicken A, Vogt M, et al: Repair of interrupted aortic arch: Results after more than 20 years.  Ann Thorac Surg  2000; 70:1896-1900.

Schulze-Neick et al., 2002. Schulze-Neick I, Li J, Reader JA, et al: The endothelin antagonist BQ123 reduces pulmonary vascular resistance after surgical intervention for congenital heart disease.  J Thorac Cardiovasc Surg  2002; 124:435-441.

Schwartz, 1985. Schwartz A: Anesthesia for non-cardiac surgery in the pediatric patient with congenital heart disease,  Park Ridge, IL, ASA Refresher Course Lectures, 1985.

Schwartz et al., 2003. Schwartz SM, Duffy JY, Pearl JM, et al: Glucocorticoids preserve calpastatin and troponin I during cardiopulmonary bypass in immature pigs.  Pediatr Res  2003; 54:91-97.

Schwartz and Jobes, 1982. Schwartz A, Jobes D: Congenital heart disease-specific anesthetic considerations.   In: Kaplan JA, ed. Cardiac anesthesia,  Menlo Park, CA: Addison-Wesley; 1982.

Schwinn, 1994. Schwinn DA: Cardiac pharmacology.  Cardiac anesthesia,  Philadelphia, PA: JB Lippincott; 1994:21-60.

Schwinn et al., 1991. Schwinn D, Leone B, Spahn D, et al: Desensitization of myocardial beta-adrenergic receptors during cardiopulmonary bypass.  Circulation  1991; 84:2559-2567.

Sell et al., 1988. Sell JE, Jonas RA, Mayer JE, et al: The results of a surgical program for interrupted aortic arch.  J Thorac Cardiovasc Surg  1988; 96:864-877.

Senzaki et al., 2002. Senzaki H, Masutani S, Kobayashi J, et al: Ventricular afterload and ventricular work in Fontan circulation: Comparison with normal two-ventricle circulation and single-ventricle circulation with Blalock-Taussig shunts.  Circulation  2002; 105:2885-2892.

Severinghaus and Spellman, 1990. Severinghaus J, Spellman B: Pulse oximeter failure thresholds in hypotension and vasoconstriction.  Anesthesiology  1990; 73:532-537.

Shah et al., 1992. Shah P, Stewart S, Calalang CC, et al: Transesophageal echocardiography and the intraoperative management of pediatric congenital heart disease: Initial experience with a pediatric esophageal 2D color flow echocardiographic probe.  J Cardiothorac Vasc Anesth  1992; 6:8-14.

Sharpe et al., 1994. Sharpe M, Dobkowski D, Murkin J, et al: The electrophysiologic effects of volatile anesthetics and sufentanil on the normal atrioventricular conduction system and accessory pathways in Wolff-Parkinson-White syndrome.  Anesthesiology  1994; 80:63-70.

Shigeaki and Mault, 1991. Shigeaki O, Mault JR, et al: Effect of a systemic-pulmonary artery shunt on myocardial function and perfusion in a piglet model.  Surg Forum  1991; 42:200-203.

Skaryak et al., 1995. Skaryak L, Chai P, Kern F, et al: Blood gas management and degree of cooling: Effects on cerebral metabolism before and after circulatory arrest.  J Thorac Cardiovasc Surg  1995; 110:1649-1657.

Skaryak et al., 1993. Skaryak L, Chai P, Kern F, et al: Combining alpha and pH stat blood gas strategy prior to circulatory arrest provides optimal recovery of cerebral metabolism.  Circulation  1993;

Skaryak et al., 1995. Skaryak LA, Kirshbom PM, DiBernardo LR, et al: Modified ultrafiltration improves cerebral metabolic recovery after circulatory arrest.  J Thorac Cardiovasc Surg  1995; 109:744-751.

Smallhorn, 2002. Smallhorn JF: Intraoperative transesophageal echocardiography in congenital heart disease.  Echocardiography  2002; 19:709-723.

Smith et al., 2002. Smith PC, Powell KR, Chang RKR, et al: Can regionalization decrease the number of deaths for children who undergo cardiac surgery? A theoretical analysis.  Pediatrics  2002; 110:849-850.

Somero and White, 1985. Somero G, White F: Enzymatic consequences under alpha stat regulation,  Boston, Nijhoff, 1985.

Soongswang et al., 2001. Soongswang J, McCrindle BW, Jones TK, et al: Outcomes of transcatheter balloon angioplasty of obstruction in the neo-aortic arch after the Norwood operation.  Cardiol Young  2001; 11:54-61.

Spach et al., 1980. Spach M, Serwer G, Anderson P, et al: Pulsatile aortopulmonary pressure-flow dynamics of patent ductus arteriosus in patients with various hemodynamic states.  Circulation  1980; 61:110-122.

Stamm et al., 2002. Stamm C, Friehs I, Duebener LF, et al: Improving results of the modified Fontan operation in patients with heterotaxy syndrome.  Ann Thorac Surg  2002; 74:1967-1977.

Stark, 1989. Stark J: Do we really correct congenital heart defects?.  J Thorac Cardiovasc Surg  1989; 97:1-9.

Stern et al., 1985. Stern DH, Gerson J, Allen FB: Can we trust the direct radial artery pressure immediately following cardiopulmonary bypass?.  Anesthesiology  1985; 62:557-563.

Stevenson, 1995. Stevenson JG: Role of intraoperative transesophageal echocardiography during repair of congenital cardiac defects.  Acta Paediatr Suppl  1995; 410:23-33.

Stevenson, 2003. Stevenson JG: Utilization of intraoperative transesophageal echocardiography during repair of congenital cardiac defects: A survey of North American centers.  Clin Cardiol  2003; 26:132-134.

Stocker et al., 1986. Stocker F, Herschkowitz N, Bossi E, et al: Cerebral blood flow during cardiopulmonary bypass in man: Effect of arterial filtration.  Thorax  1986; 41:386-395.

Stocker et al., 2003. Stocker C, Penny DJ, Brizard CP, et al: Intravenous sildenafil and inhaled nitric oxide: A randomized trial in infants after cardiac surgery.  Intensive Care Med  2003; 29:1996-2003.

Strauc et al., 2003. Strauc A, Spielvogel JT, Haldenwang PL, et al: Cerebral physiology and outcome after hypothermic circulatory arrest followed by selective cerebral perfusion.  Ann Thorac Surg  2003; 76:1972-1981.

Stullken and Michenfenlder, 1977. Stullken Jr EH, Michenfenlder JD, et al: The non-linear responses of cerebral metabolism to low concentrations of halothane, enflurane, isoflurane and thiopental.  Anesthesiology  1977; 46:28.

Sunamori et al., 1982. Sunamori M, Amano J, Okamura T, et al: Superior action of magnesium-lidocaine-1-aspartate cardioplegia in experimental myocardial protection.  Jpn J Surg  1982; 12:372-380.

Sungurtekin et al., 1999. Sungurtekin H, Plochl W, Cook DJ: Relationship between cardiopulmonary bypass flow rate and cerebral embolization in dogs.  Anesthesiology  1999; 91:1387-1393.

Sutton et al., 1991. Sutton L, Clark B, Norwood C: Global cerebral ischemia in piglets under conditions of mild and deep hypothermia.  Stroke  1991; 22:1567-1573.

Svensson et al., 2001. Svensson LG, Nadolny EM, Penney DL, et al: Prospective randomized neurocognitive and S-100 study of hypothermic circulatory arrest, retrograde brain perfusion and antegrade brain perfusion for aortic arch operations.  Ann Thorac Surg  2001; 71:1905-1912.

Swain et al., 1991. Swain JA, McDonald TJJ, Balaban RS, et al: Metabolism of the heart and brain during hypothermic cardiopulmonary bypass.  Ann Thorac Surg  1991; 51:105-109.

Swain et al., 1991. Swain JA, McDonald Jr TJ, Robbins RC, Balaban RS: Relationship of cerebral and myocardial intercellular pH to blood pH during hypothermia.  Am J Physiol  1991; 260:H1640.

Swain et al., 1990. Swain JA, Robbins RC, Balaban RS, et al: The effect of cardiopulmonary bypass on brain and heart metabolism: a 31P NMR study.  Magn Reson Med  1990; 15:446.

Swan, 1984. Swan H: The importance of acid-base management for cardiac and cerebral preservation during open heart operations.  Surg Gynecol Obstet  1984; 158:391-414.

Syzmonowitcz et al., 1990. Syzmonowitcz W, Walker AM, Yu YH, et al: Regional cerebral blood flow after hemorrhagic hypotension in the preterm, near term and term lamb.  Pediatr Res  1990; 162:977-982.

Tabbutt et al., 2001. Tabbutt S, Ramanoorthy C, Montenegro L, et al: Impact of inspired gas mixtures on preoperative infants with hypoplastic left heart syndrome during controlled ventilation.  Circulation  2001;I-159-I-164.

Taeed et al., 2001. Taeed R, Schwartz SM, Pearl JM, et al: Unrecognized pulmonary venous desaturation early after Norwood palliation confounds Gp:Gs assessment and compromises oxygen delivery.  Circulation  2001; 103:2699-2709.

Tanaka et al., 1988. Tanaka J, Shiki K, Asou T, et al: Carbon dioxide, brain damage, and cardiac surgery [letter].  Lancet  1988; 1:353.

Tang and Sun, 1985. Tang W, Sun GY: Effects of ischemia on free fatty acids and diacylglycerols in developing rat brain.  J Dev Neurosci  1985; 3:51-56.

Tavani et al., 2003. Tavani F, Zimmerman RA, Clancy RR, et al: Incidental intracranial hemorrhage after uncomplicated birth: MRI before and after neonatal heart surgery.  Neuroradiology  2003; 45:253-258.

Teitel et al., 1985. Teitel DF, Sidi D, Chin T, et al: Developmental changes in myocardial contractile reserve in the lamb.  Pediatr Res  1985; 19:948-955.

Teitel et al., 1991. Teitel DF, Klautz R, Steendijk P, et al: The end-systolic pressure-volume relationship in the newborn lamb: Effects of loading and inotropic interventions.  Pediatr Res  1991; 29:473-482.

Tempe and Kirmani, 2002. Tempe DK, Kirmani S: Coagulation abnormalities in patients with cyanotic congenital heart disease.  J Cardiothorac Vasc Anesth  2002; 16:752.

Thanopoulos et al., 2003. Thanopoulos BD, Karanassios E, Tsaousis G, et al: Catheter closure of congenital/acquired muscular VSDs and perimembranous VSDs using the Amplatzer devices.  J Interv Cardiol  2003; 16:399.

Tlaskal et al., 1999. Tlaskal T, Hucin B, Hruda J, et al: Results of primary and two-stage repair of interrupted aortic arch.  Eur J Thorac Surg  1999; 15:227-228.

Tobias, 2000. Tobias JD: Etomidate: Applications in pediatric care and pediatric anesthesiology.  Pediatr Crit Care Med  2000; 1:100-106.

Todd, 1998. Todd MM: EEGs, EEG processing, and the bispectral index.  Anesthesiology  1998; 89:815-817.

Tullis and Rubin, 1982. Tullis R, Rubin H: Calcium protects DNase I from proteinase K: A new method for the removal of spin-trapped radicals in aqueous solution of pyrimidine nucleosides and nucleotides. Reaction of the hydroxyl radical.  Int J Radiat Biol  1982; 41:241.

Tweddell et al., 2000. Tweddell JS, Hoffman GM, Fedderly RT, et al: Patients at risk for low systemic oxygen delivery after the Norwood procedure.  Ann Thorac Surg  2000; 69:1893-1899.

Ungerleider et al., 1989a. Ungerleider R, Greeley W, Sheikh K, et al: The use of intraoperative echo with Doppler color flow imaging to predict outcome after repair of congenital cardiac defects.  Ann Surg  1989; 210:526.

Ungerleider et al., 1992. Ungerleider R, Greeley W, Kanter R, et al: The learning curve for intraoperative echocardiography during congenital heart surgery.  Ann Thorac Surg  1992; 54:691-696.

Ungerleider et al., 1989b. Ungerleider R, Greeley W, Kisslo J: Intraoperative echocardiography in congenital heart disease surgery: Preliminary report on a current study.  Am J Cardiol  1989; 63:3F-8F.

Ungerleider et al., 1989c. Ungerleider R, Kisslo J, Greeley W, et al: Intraoperative prebypass and postbypass epicardial color flow imaging in the repair of atrioventricular septal defects.  J Thorac Cardiovasc Surg  1989; 98:90-99.

Ungerleider et al., 1995. Ungerleider R, Kisslo J, Greeley W, et al: Intraoperative echocardiography during congenital heart operation: Experience from 1,000 cases.  Ann Thorac Surg  1995; 60:539-542.

Ungerleider, 1998. Ungerleider R: Effects of cardiopulmonary bypass and use of modified ultrafiltration.  Ann Thorac Surg  1998; 65(6 Suppl):S35-S38.

Urata et al., 1989. Urata H, Healy B, Stewart R, et al: Angiotensin II receptors in normal and failing human hearts.  J Clin Endocrinol Metab  1989; 69:54-61.

van der Linden and Casimir, 1991. van der Linden J, Casimir AH: When do cerebral emboli appear during open heart operations? A transcranial Doppler study.  Ann Thorac Surg  1991; 51:237-241.

van Dongen et al., 2003. van Dongen EI, Glansdorp AG, Mildner RJ, et al: The influence of perioperative factors on outcomes in children aged less than 18 months after repair of tetralogy of Fallot.  J Thorac Cardiovasc Surg  2003; 126:703-710.

Van Praagh and Van Praagh, 1965. Van Praagh R, Van Praagh S: The anatomy of common aortopulmonary trunk (truncus arteriosus communis) and its embryologic implications. A study of 57 necropsy cases.  Am J Cardiol  1965; 16:406-425.

Vetter et al., 1986. Vetter R, Will H, Kuttner I, et al: Developmental changes of Ca++ transport systems in chick heart.  Biomed Biochem Acta  1986; 45:219.

Vieira et al., 1991. Vieira A, Berry L, Ofoso F, et al: Heparin sensitivity and resistance in the neonate: An explanation.  Thrombosis Res  1991; 63:85-98.

Vlahakes et al., 1980. Vlahakes GJ, Turley K, Hoffman JI: The physiology of failure in acute right ventricular hypertension.  Circulation  1980; 63:87-95.

Waldo, 1980. Waldo A: Modes and methods of recording electrograms. Diagnosis and treatment of cardiac arrhythmias following open heart surgery,  Mt Kisco, NY: Futura Publishing; 1980:21.

Ward et al., 1985. Ward J, Blakeky W, Joner E: Mammalian cells are not killed by DNA single-strand breaks caused by hydroxyl radicals from hydrogen peroxide.  Radic Res  1985; 103:383.

Watanabe et al., 1989. Watanabe T, Hrita H, Kobayashi M, et al: Brain tissue pH, oxygen tension, and carbon dioxide tension in profoundly hypothermic cardiopulmonary bypass.  J Thorac Cardiovasc Surg  1989; 97:396-401.

Wauthy and Abdel Kafi, 2003. Wauthy P, Abdel Kafi S, et al: Inhaled nitric oxide versus prostacyclin in chronic shunt-induced pulmonary hypertension.  J Thorac Cardiovasc Surg  2003; 126:1434-1441.

Weinstein et al., 2003. Weinstein S, Cua C, Chan D, et al: Outcome of symptomatic patients undergoing extracardiac Fontan conversion and cryoablation.  J Thorac Surg  2003; 126:529-536.

Weisfeldt and Halperin, 1986. Weisfeldt ML, Halperin HR: Cardiopulmonary resuscitation: Beyond cardiac massage.  Circulation  1986; 74:443-448.

Weller et al., 2002. Weller RJ, Weintraub R, Addonizio L, et al: Outcome of idiopathic restrictive cardiomyopathy in children.  Am J Cardiol  2002; 90:501-506.

Wernovsky et al., 1992. Wernovsky G, Jonas R, Newburger J: The Boston Circulatory Arrest Study: Hemodynamics and hospital course after the arterial switch operation.  Circulation  1992; 86(Suppl I):237A.

Wernovsky et al., 1988. Wernovsky G, Hougen T, Walsh E, et al: Mid-term results following the arterial switch operation for transposition of the great arteries with intact ventricular septum: Clinical, hemodynamic, echocardiographic, and electrophysiologic data.  Circulation  1988; 77:1333.

Wernovsky and Newburger, 2003. Wernovsky G, Newburger JW: Neurologic and developmental morbidity in children with complex congenital heart disease.  J Pediatr  2003; 142:6-8.

Wessel et al., 1997. Wessel DL, Adatia I, Van Marter LJ, et al: Improved oxygenation in a randomized trial of inhaled nitric oxide for persistent pulmonary hypertension of the newborn.  Pediatrics  1997; 100:E7.

Wessel et al., 1989. Wessel D, Triedman J, Wernovsky G, et al: Pulmonary and systemic hemodynamics of amrinone in neonates following cardiopulmonary bypass.  Circulation  1989;II-488.

Wieloch et al., 1984. Wieloch T, Harris R, Symon L, et al: Influence of severe hypoglycemia on brain extracellular calcium and potassium activities, energy, and phospholipid metabolism.  J Neurochem  1984; 43:160.

Williams et al., 1999. Williams GD, Jones TK, Hanson KA, et al: The hemodynamic effects of propofol in children with congenital heart disease.  Anesth Analg  1999; 89:1411-1416.

Williams et al., 2003. Williams WG, McBride BW, Ashburn DA, et al: Outcomes of 829 neonates with complete transposition of the great arteries 12–17 years after repair.  Eur J Cardio Surg  2003; 24:1-10.

Wirrel et al., 1993. Wirrel EC, Pelausa EO, Allen AC, et al: Massive pericardial effusion as a cause for sudden deterioration in a very low birth weight infant.  Am J Perinat  1993; 10:419-423.

Wong et al., 1974. Wong A, Salem M, Mani M, et al: Glycopyrrolate as a substitute for atropine in reversal of curarization in pediatric cardiac patients.  Anesth Analg  1974; 53:412.

Woodman and Harker, 1990. Woodman RC, Harker LA: Bleeding complications associated with cardiopulmonary bypass.  Blood  1990; 76:1680-1697.

Wypij et al., 2003. Wypij D, Newburger JW, Rappaport LA, et al: The effect of duration of deep hypothermic circulatory arrest in infant heart surgery on late neurodevelopment: The Boston Circulatory Arrest Trial.  J Thorac Cardiovasc Surg  2003; 126:1397-1403.

Yoshii et al., 2003. Yoshii S, Akashi O, Kobayashi M, et al: Preliminary results of intermittent retrograde cerebral perfusion during proximal aortic arch surgery.  Jpn J Thorac Cardiovasc Surg  2003; 51:588-593.

Zanchetta et al., 2003. Zanchetta M, Onorato E, Rigatelli G, et al: Intracardiac echocardiography-guided transcatheter closure of secundum atrial septal defect: a new efficient device selection method.  J Am Coll Cardiol  2003; 42:1677.

Zapletal et al., 1976. Zapletal A, Paul T, Samenek M: Pulmonary elasticity in children and adolescents.  J Appl Physiol  1976; 40:953.