Neurology of Congenital Cardiovascular Disease: Brain

Neonatal Cardiology, 3rd Ed. Michael Artman

Chapter 14. Neurology of Congenital Cardiovascular Disease: Brain

Development, Acquired Injury, and Neurodevelopmental Outcome

■ INTRODUCTION

■ GENETIC CONTRIBUTION TO ADVERSE OUTCOME

Shared Genetic Pathways in Brain and Heart Development

Common Genetic Syndromes with Congenital

Cardiovascular Disease

Patient-Specific Genetic Risk Modifiers

■ BRAIN DEVELOPMENT IS DELAYED IN PATIENTS

WITH CONGENITAL CARDIOVASCULAR DISEASE

Human Brain Developmental Time Line

Fetal Circulation in Congenital Cardiovascular

Disease: Effects on Cerebral Blood Flow

Magnetic Resonance Imaging Identifies Delayed Brain Development Before Surgery

Fetal Brain MRI

■ ACQUIRED BRAIN INJURY: WHITE MATTER INJURY AND STROKE

Imaging Characteristics of Acquired Brain Injury Risk Factors for Preoperative Brain Injury Risk Factors for Intraoperative Brain Injury Risk Factors for Postoperative Brain Injury

■ NEURODEVELOPMENTAL OUTCOME Transposition of the Great Arteries Hypoplastic Left Heart Syndrome Neurodevelopmental Signature of Complex

Congenital Cardiovascular Disease

■ CONCLUSIONS

■ SUGGESTED READINGS

■ INTRODUCTION

Dramatic advances in noninvasive imaging, cardiopulmonary bypass, surgical technique, and intensive care now allow most patients with congenital cardiovascular defects to undergo surgery during the neonatal period or in later infancy. With these advances, mortality has declined, but many patients are at risk for neurodevelopmental impairment. A natural assumption was that adverse neurologic outcome was directly related to brain injury sustained during neonatal surgical intervention, leading to a seminal study in the late 1980s. The Boston Circulatory Arrest Trial compared two methods of vital organ support in infants undergoing open-heart surgery to repair d-transposition of the great arteries. Much of what is known about the relationship between early intervention for complex cardiovascular disease and neurodevelopmen- tal outcome has been gleaned from this study. However, it is now apparent that injury to the brain may occur during fetal life, at birth, preoperatively, intraoperatively, and postoperatively. The interplay between brain development and the circulation is complex and occurs at many levels. This chapter reviews mechanisms influencing neurologic outcome, including (1) shared genetic and developmental pathways, (2) physiologic effects of congenital cardiovascular lesions on brain blood flow, and (3) the timing, appearance, and mechanism of acquired brain injuries. We summarize how these pathogenic mechanisms result in a neurodevelopmental “signature” of congenital cardiovascular disease. Finally, we will speculate on how these mechanisms suggest strategies of neuroprotection, repair, and recovery that may improve outcome.

■ GENETIC CONTRIBUTION TO ADVERSE OUTCOME

Shared Genetic Pathways in Brain and Heart Development

Certain aspects of heart and brain development occur simultaneously in the human fetus (summarized in Chapter 1 for heart and below for brain). Many vertebrate organs undergo related developmental events (eg, cell fate determination, cell migration, dorsal/ventral patterning, left/right asymmetry, area specification, etc.). Thus, it is not surprising that similar genes share important and similar developmental roles in both organs (Table 14-1). This includes genes such as those in the Ras-MAPK pathway, members of the transforming growth factor β family, fibroblast growth factor family members, notch and notch ligands, and vascular endothelial growth factor. Disruption of shared fundamental genetic pathways that result in cardiac defects will affect brain development as well.

Common Genetic Syndromes with Congenital Cardiovascular Disease

Congenital cardiovascular disease frequently occurs within the context of a genetic syndrome. Often but not exclusively, these syndromes also include abnormal neurodevelopment. Notable examples of genetic syndromes with both congenital cardiovascular disease and neurode- velopmental impairment include trisomy 21 (Down syndrome), 22q11 deletion (DiGeorge and velocardiofacial syndrome), monosomy X (Turner syndrome), and Jacobsen and Williams syndromes. Despite well-characterized chromosomal microdeletions, identifying a specific gene or limited set of genes that accounts for any one component (eg, cardiac or brain) of the syndrome is challenging. The 22q11 deletion syndrome is a good example. The causative microdeletion encompasses three megabases of DNA, representing 30 to 40 genes. The syndrome is notably highly heterogeneous with regard to all potential components of the syndrome (eg, parathyroid, immune deficiency, conotruncal cardiac defects, and neurodevel- opmental outcome). Even when known genetic or malformation syndromes are excluded, a large prospective study of the determinants of 1-year neurodevelopmental outcome identified genetic syndromes, unsuspected at birth, as among the most potent predictors of adverse outcome. The list of single-gene defects contributing to congenital cardiovascular disease and/or neurodevelopmental impairment is likely to grow quickly with the advent of large-scale genomics projects.

Patient-Specific Genetic Risk Modifiers

Beyond the direct effects of single genes or shared genetic pathways in genetic syndromes, patient genotype can influence neurodevelopmental outcome as “risk modifiers” of the response to brain injury. This is perhaps best described for alleles of apolipoprotein E, in which the Apo ε4 allele is associated with adverse outcome in many conditions in adults (eg, Alzheimer’s disease, traumatic brain injury, stroke, and subarachnoid hemorrhage). In infants with congenital cardiovascular disease, however, the Apo ε2 allele is associated with worse outcome. Interestingly, polymorphisms of inflammatory cytokines, including interleukin-6, have been associated with cerebral palsy in term newborns. This issue has not been examined in patients with congenital cardiovascular disease despite observations that inflammatory cytokines, including interleukin-6, are increased after surgery and are associated with postoperative cardiovascular morbidity.

■ BRAIN DEVELOPMENT IS DELAYED IN PATIENTS WITH CONGENITAL CARDIOVASCULAR DISEASE

Human Brain Developmental Time Line

Human cardiac development begins during the first post- conceptual days with rhythmic contractions of a primitive heart tube beginning by embryonic day 23 and a morphologically mature heart formed by day 50, or gestational week 7 (reviewed in Chapter 1). In contrast, brain development extends over a much longer time period, with morphologic events (cell proliferation, migration, axon pathfinding, and target selection) occurring in the first two trimesters, followed by a prolonged period of refinement of connections that occurs in both the third trimester and the early postnatal period.

TABLE 14-1. Select Genes with Identified Role(s) in Both Heart and Brain Development

Gene	Function in cardiac development	Function in brain development	Syndrome or isolated CHD
Nkx2.5	Cardioblast cell fate commitment, chamber septation	Neural cell fate commitment	Holt-Oram, ASD, VSD, TOF
TBX5	Left ventricular specification	Cortical area specification, axon guidance	Holt-Oram
Lefty1	Left/right asymmetry	Neural cell fate commitment, left/right asymmetry	Heterotaxy
ZIC1	Left/right asymmetry	Neural progenitors proliferation, neural crest and roof plate specification, holoprosencephaly, cerebellar development	Heterotaxy
GATA4 NF-1	Heart tube formation Myocardial growth	Astrocyte proliferation Glial differentiation	ASD, VSD
TFAP2B	Neural crest	Regulates monoaminergic gene expression in neural crest cells in midbrain, hindbrain, and spinal cord	Charge
Notch/Jagged	Cardiac progenitor cell fate determination	Neural cell fate commitment	Alagille, aortic stenosis/ bicuspid aortic valve
TBX1	Outflow tract, pharyngeal arch development, aortic arch patterning	Learning and memory (prepulse inhibition)	DiGeorge
Ras-MAPK	Pulmonary valve development, cardiomyopathy	Pleomorphic: learning memory, neuronal survival/death, plasticity	Noonan cardio- facio-cutaneous, Costello
Elastin/Limk1, Cyln2, Fzd9; GtfII	Elastin: valve formation	Limk1, Cyln2: regulation of neuronal cytoskeleton—growth cone motility, dendrite formation, synaptogenesis; Fxd9: hippocampal development; Gtf2i: visuospatial processing	Williams-Beuren
PROSIT240	Ventricular septation	Cerebellar development	D-TGA

A primitive neural tube forms by gestational week 5 with identifiable radial glia, the neural stem cell. Cortico- genesis, the process of production and migration of neurons from regions of proliferation to their targets in specific neocortical layers, occurs from week 7 through week 18, although a mature six-layered cortex does not appear until week 26. Axon outgrowth, pathfinding, target selection, and innervation all take place after production and migration of neurons. Formation of connections between the thalamus and cortex requires a transient population of early born neurons, referred to as subplate neurons because of their location as a discrete layer below the cortical plate. In humans, thalamocortical pathfinding takes place between the histological emergence of the subplate at the end of the first trimester and the appearance of cholinesterase positive fibers in the subplate at weeks 17 to 20 (earlier for somatosensory vs. visual cortex). Thalamocortical fibers accumulate during a “waiting period” from week 17 through weeks 22 to 26 before innervating the cortical plate.

Most morphologic events of brain development are completed by the end of the second trimester. The third trimester involves a period of dramatic brain growth and refinement of connections that are dependent on endogenous and spontaneous neuronal activity arising at multiple levels. In the visual system, this endogenous activity takes the form of spontaneous waves of neuronal activation that sweep across and tile the retina and are then transmitted to the thalamus and cortex. This patterned activity sculpts developing neural circuits into mature, precise patterns. Ocular dominance columns in the visual system, representing nonoverlapping eye-specific innervation, are a representative example of such a patterned circuit and form the basis of binocular vision. In higher primates, ocular dominance columns have fully formed by birth before onset of visually driven activity. At a gross morphologic level, the third trimester is characterized by development of secondary and tertiary gyri. At a cellular level, brain growth involves elaboration of dendritic arbors and formation of corticocortical connections. Myelination of fiber tracts also begins during fetal life, with a characteristic caudal- to-cranial pattern beginning with deep structures, such as the tegmentum and cerebellar peduncles. By the end of the third trimester, myelination extends to the posterior limb of the internal capsule and involves the motor fibers of the pyramidal tract. In neocortex, myelination begins in the optic radiations and occipital white matter after birth before extending to the frontal lobes by 9 months of age.

Brain development continues after birth, with sensory driven activity now influencing the refinement of connections. For a brief period of time (critical period), soon after the onset of vision, deprivation of visually evoked activity leads to permanent loss of visual acuity in the deprived eye, a process referred to in humans as amblyopia, which represents a form of neural plasticity. In humans, amblyopia is usually observed only after 6 months of age. Weeks of deprivation can lead to substantial loss of visual acuity between 6 and 18 months, and months of deprivation can have an effect until 8 years of age.

Fetal Circulation in Congenital Cardiovascular Disease: Effects on Cerebral Blood Flow

The fetal circulation is unique in a number of respects that impact cerebral blood flow. As reviewed in detail in Chapter 3, vascular and cardiac blood flow patterns in the normal fetus direct the most highly saturated blood from the ductus venosus and left hepatic veins via the foramen ovale to the left heart and, subsequently, to the cerebral circulation. In contrast, in d-transposition of the great arteries, the aorta arises from the right ventricle and thus receives the relatively desaturated blood from the superior vena cava, lower body, and coronary sinus. In hypoplastic left heart syndrome, the fetal circulation is characterized by admixture of all venous streams in the right atrium and ventricle. The ascending aorta is a very small vessel, delivering blood in a retrograde direction to the coronary arteries. The aortic arch is also hypoplastic and shows flow reversal to supply the brain and upper body (see Chapters 3 and 8). The effects of these abnormal flow patterns on brain development are uncertain but may be very different despite the fact that both decrease the oxygen content of the blood delivered to the brain. In d-transposition of the great arteries, the pulsatility and perfusion pressure of the cerebral arterial circulation are normal. On the other hand, in hypoplastic left heart syndrome, the hypoplastic isthmus and aortic arch may function as resistors, decreasing the pulsatility and perfusion pressure to the cerebral circulation. This cannot easily be overcome by the cerebral circulation, particularly early in gestation, and may explain the finding of a recent study showing that ascending aortic diameter predicts the degree of microcephaly in newborns with hypoplastic left heart syndrome.

Blood flow to the fetal brain is estimated to be almost one-quarter of the combined ventricular output in the third trimester. Local regulation of fetal cerebral blood flow is thought to redistribute blood flow to the brain in the setting of placental insufficiency, a phenomenon referred to as “brain sparing,” a pattern of in which overall somatic growth is restricted but head growth is relatively preserved. In the setting of congenital cardiovascular disease with decreased oxygen content of the cerebral arterial blood, similar mechanisms may be invoked; for example, the ratio of cerebral to placental resistance can decrease so that cerebral oxygen delivery can be preserved. However, low cerebral blood flow has been measured in newborns with hypoplastic left heart syndrome and d-transposition of the great arteries using magnetic resonance imaging, and newborns with certain forms of congenital cardiovascular disease have smaller head circumferences, which may be an indicator of impaired brain growth. The issue is complex, however. The different patterns of alterations in cerebral blood flow are associated with different patterns of growth disturbance. Newborns with d-transposition of the great arteries tend to have small head circumference with normal birth weight, whereas those with hypoplastic left heart syndrome are smaller in all dimensions but head volume is disproportionately decreased. Interestingly, infants with isolated aortic coarctation have a greater head volume relative to birth weight.

Magnetic Resonance Imaging Identifies Delayed Brain Development Before Surgery

Advanced magnetic resonance imaging (MRI) provides the highest-resolution conventional images of brain anatomy and acquired brain lesions. These techniques can also be used to measure brain development and to investigate brain metabolism and microstructure. Proton MR spectroscopic imaging measures resonance from N-acetyl groups (predominantly N-acetylaspartate), lactate, creatine, and tetramethylamines (predominantly choline- containing compounds). N-acetylaspartate is found predominantly in neurons (cell body and axon) so that changes in N-acetylaspartate reflect neuronal metabolic integrity. Particularly relevant to studies of brain development is the observation that N-acetylaspartate increases consistently with advancing cerebral maturity, providing a developmental brain “growth chart.”

Another advanced technique is diffusion tensor imaging, which provides a sensitive measure of regional brain microstructural development. Three-dimensional spatial distribution of water diffusion is characterized in each voxel of the MR image. With increasing brain maturation, brain water content diminishes, and developing neuronal and glial cell membranes increasingly restrict proton diffusion, resulting in a consistent decrease in average diffusivity over time in gray and white matter regions (Figure 14-1).

Recently, using both advanced imaging modalities, brain metabolism and microstructure were characterized as measures of brain maturation in term newborns with d-transposition of the great arteries or single-ventricle physiology before heart surgery and compared to control infants. Relative to the normal control newborns, newborns with congenital cardiovascular disease had 10% lower N-acetylaspartate/choline ratios and 4.5% higher average diffusivity. Comparing these data to values obtained from normal fetuses suggests that term newborns with congenital cardiovascular disease have a delay in brain development of approximately 1 month, equivalent to an infant born prematurely at 34 to 36 weeks. Thus, data from this study suggest that abnormal brain development is present at birth in these newborns.

Brain development can also be assessed at a macroscopic, morphologic level. In a similar study examining structural brain development, preoperative MRIs from term newborns with d-transposition of the great arteries or hypoplastic left heart syndrome were reviewed to assign a “total maturation score” describing myelination, cortical folding, involution of glial cell migration bands, and the presence of germinal matrix tissue (Figure 14-2). Consistent with the data described above, the maturation score was delayed, corresponding to a delay of 1 month in structural brain development compared to normative data in infants without congenital cardiovascular defects. These in vivo observations using quantitative MR methods are consistent with neuropathology data showing that newborns with congenital cardiovascular disease are more likely to be microcephalic and to have an immature cortical mantel.

Fetal Brain MRI

Fetal brain ultrasound studies suggest that the decline in head growth begins after mid-gestation in fetuses with hypoplastic left heart syndrome. A study comparing brain volumes and proton MR spectroscopic images from fetuses with congenital cardiovascular disease and normal fetuses between 25 and 37 gestational weeks showed definitive evidence for delayed fetal brain development. No differences were found between controls and fetuses with cardiovascular defects during the second trimester. During the third trimester, a progressive impairment of brain volumes was observed, particularly in those fetuses with left-sided obstruction. Additionally, larger delays in the expected increase in N-acetylaspartate/choline ratio and greater impairment of growth in brain volume were noted in fetuses with aortic atresia and no antegrade blood flow in the aortic arch. These observations support the concept that brain development is impaired during fetal life because of impaired fetal cerebral blood flow, resulting in abnormal oxygen and substrate delivery. Furthermore, these findings suggest that compensatory mechanisms (brain-sparing effects) are inadequate to compensate for impaired fetal cerebral blood flow.

■ ACQUIRED BRAIN INJURY: WHITE MATTER INJURY AND STROKE

Neonates with congenital cardiovascular disease are also at risk of discrete acquired brain injury in the perioperative period, and that injury may be exacerbated by delayed brain development.

FIGURE 14-1. Fetal MRI and diffusion imaging detects delayed brain development in congenital cardiovascular disease. A-C. Sagittal, axial, and coronal T2 images at 31 weeks in a fetus with prenatal diagnosis of hypoplastic left heart syndrome. The brain is morphologically normal. D and E. Average diffusivity coefficient (ADC) is plotted against gestational age for periatrial white matter (D) or thalamus (E). Values for three fetuses with hypoplastic left heart syndrome (orange) are compared with a control cohort without congenital cardiovascular disease (green). The regression line was fit to the control population. Average diffusion is higher in fetuses with hypoplastic left heart syndrome, indicating a relative delay in development.

Imaging Characteristics of Acquired Brain Injury

Focal brain abnormalities (or “injuries”) in the term newborn can be clearly and reliably detected with conventional MRI and with greater resolution than with either ultrasound or computed tomography. Further, the extent of MRI abnormalities corresponds closely to histopathological changes found on postmortem examination. The most common brain injuries observed in newborns with congenital cardiovascular disease are white matter injury (defined in more detail below) and small focal strokes (less than one- third to two-thirds of the arterial distribution; Figure 14-3).

It is important to recognize that the strokes and white matter injuries identified by MRI in these research studies are largely clinically silent and overlooked by routine clinical screening cranial ultrasounds. These patterns of brain injury are uncommon for term newborns experiencing hypoxic ischemic injury. Instead, newborns with birth asphyxia present with two different patterns of brain injury: parasag- gital watershed injury or the “basal ganglia pattern” with injury to deep gray nuclei, indicative of partial or total ischemia, respectively. Hypoxic ischemic injury is rarely seen in newborns with congenital cardiovascular disease, usually occurring only after cardiac arrest or severe shock.

FIGURE 14-2. Macroscopic brain development: cortical folding. Three-dimensional reconstructions of cortical white matter volumes from MRI scans taken at 24 weeks to term gestation showing the development of cortical folding and surface curvature. The white matter surface from a term newborn with hypoplastic left heart syndrome displays noticeably less cortical folding and smoother surface curvature than a term newborn without heart disease.

Stroke is defined as “a focal area of diffusion restriction in an arterial territory,” a definition that incorporates both imaging characteristics and a presumed embolic mechanism. Stroke is distinguished from white matter injury based on imaging characteristics, the latter characterized by “punctate periventricular lesions associated with T1 hyperintensity (brightness) with or without restriction of water diffusion.” Larger injuries (>5 mm) in the white matter may represent either focal embolic strokes or large, atypical confluent white matter injuries.

Identification of white matter injury in term newborns with cardiac defects was unexpected because this pattern was thought to be restricted to premature newborns with brain injury (periventricular leukomalacia). The early descriptions of periventricular leukomalacia referred to extensive cystic necrotic lesions that were frequently accompanied by white matter atrophy reflected by enlargement of the lateral ventricles. The cystic lesions were characteristically symmetric and located adjacent to the anterior and posterior horns of the lateral ventricles. However, with the increasing use of MRI tools, focal or diffuse noncystic white matter injuries are emerging as the predominant lesions, whereas cystic lesions account for less than 5% of injuries in some series.

Advanced MRI techniques such as diffusion tensor imaging and spectroscopy are useful for evaluating acute injury as well as measuring brain development. Quantitative and qualitative morphometric techniques (eg, deformation morphometry and analysis of curvature) have been useful for predicting outcomes following premature birth. Studies evaluating these techniques in newborns with congenital cardiovascular disease are in progress.

Risk Factors for Preoperative Brain Injury

Several relatively large prospective studies have been performed using preoperative and postoperative brain MRI to determine the frequency of acquired brain injury and associated risk factors in newborns with congenital cardiovascular disease. Preoperative brain injury in the form of white matter injury or stroke is present in 28% to 39% of such newborns. Risk factors for preoperative brain injury are summarized in Table 14-2 and include hypoxemia and time to surgery, preoperative base deficit, preoperative cardiac arrest, the presence of aortic atresia in HLHS, and balloon atrial septostomy. The risk of balloon atrial septostomy has not been observed in all studies, perhaps because of confounding association with hypoxemia, a common indication for the procedure. As expected, these risk factors reflect brain injury occurring as the consequence of hypoxemia and ischemia related to intracardiac shunting or impaired pulmonary or systemic perfusion, respectively, prior to diagnosis and initiation of therapy. Recent studies suggest that neonates with a prenatal diagnosis of aortic arch obstruction demonstrate less preoperative brain injury than those diagnosed after birth, likely because of earlier initiation of PGE1 and other therapies. Delayed brain development has also been implicated in susceptibility to brain injury, particularly in the preoperative period.

FIGURE 14-3. MRI patterns of injury. A and B. Moderate white matter injury in a newborn with hypoplastic left heart syndrome is seen on sagittal T1 images in the postoperative scan. White matter injuries appear as small, focal areas of T1 hyperintensity (brightness). C and D. Term newborn with hypoplastic left heart syndrome imaged postoperatively at day of life 17, after a modified Norwood procedure. A small middle cerebral artery distribution infarct is seen as cortical T2 hyperintensity (white arrows in C) and corresponding reduced diffusion (white arrows in D) in the right parietal- occipital lobe. E and F. Term newborn with transposition of the great arteries imaged preoperatively after a balloon atrial septostomy. A single focus of T1 hyperintensity is seen in the periatrial white matter on the coronal SPGR sequence (E). This same focus has reduced water diffusivity on the average diffusivity (Dav) map (F, dark spot). This spot is larger than the typical solitary white matter lesion and may represent a small embolic stroke.

Risk Factors for Intraoperative Brain Injury

Proposed risk factors for intraoperative brain injury relate predominantly to the method of cardiopulmonary bypass and/or hypothermic total circulatory arrest (Table 14-2).

Most neuroprotective trials have involved manipulating some component of cardiopulmonary bypass. The Boston Circulatory Arrest Trial compared two methods of vital organ support in infants undergoing open-heart surgery to repair d-transposition of the great arteries; d-transpo- sition of the great arteries is an ideal lesion to study the effects of acquired perioperative brain injury because the cardiac anatomy is relatively homogeneous and is rarely associated with genetic or malformation syndromes that might contribute independently to neurodevelopmental outcome. Neonates with this condition undergo complete repair in the neonatal period; normal cardiovascular physiology is established with low mortality and excellent long-term cardiac functional outcome. Although deep hypothermic circulatory arrest, which provides the surgeon with an empty and relaxed heart, allows intricate surgery to be performed most easily, there was concern that late adverse neurologic outcome could occur because the “safe” duration of circulatory arrest was unknown. As an alternative, low-flow cardiopulmonary bypass maintains some amount of brain oxygen delivery while still allowing the surgeon a relatively bloodless field. Equipoise emerged from concerns that low-flow bypass prolonged the exposure to pump-related sources of injury, including embolism and inflammation.

This study enrolled 171 infants into a single-center, randomized clinical trial comparing deep hypothermic total circulatory arrest with low-flow cardiopulmonary bypass. All of the early outcome variables pointed to superiority of low-flow bypass compared to circulatory arrest. The variables suggesting greater neurologic injury in the circulatory arrest group included more frequent postoperative seizures, higher serum levels of brain-specific enzymes (creatine kinase), worse 1-year motor outcome (Bayley Scales of Infant Development-Psychomotor Development Index), and abnormalities on neurologic exam. On the other hand, no differences were found in cognitive development (Bayley-Mental Development Index) or MRI at 1 year of age. Importantly, differences between the groups disappeared when the patients were assessed at 16 years of age, and both groups remained below population norms for performance on standardized tests. An important observation that has implications for all studies of neurodevelopment in patients with congenital cardiovascular disease was that testing at 1 year of age was only modestly predictive of outcome at 8 years of age with poor sensitivity and positive predictive value.

A number of other differences emerged as predictors of adverse outcome, including seizures, when the patient cohort was 16 years of age. The presence of a ventricular septal defect was associated with worse outcome at all ages and, apart from socioeconomic status, explained the largest percentage of variance in the scores (3.2% compared with 0.3% for the treatment group assignment). Interestingly, surgery was performed at older ages in infants with transposition of the great arteries and ventricular septal defect (2 to 3 weeks of age) compared with d-transposition of the great arteries and intact ventricular septum (1 week of age), which might, at least in part, explain this finding.

TABLE 14-2. Risk Factors for Brain Injury

Preoperative	Intraoperative	Postoperative
Low arterial hemoglobin saturation	Prolonged total circulatory arrest (>40 min)	Low blood pressure
Length of time to surgery	Decreased cerebral oxygen saturation (NIRS)	Low arterial PaO2
Catheter-based procedure (eg, balloon atrial septostomy)	Cardiopulmonary bypass strategy (regional cerebral perfusion)	Prolonged cerebral regional oxygen saturation (NIRS <45% for >3 h)
Preoperative base deficit	Air or particulate emboli	Morphologically immature brain (total maturation score)
Preoperative cardiac arrest	Inflammation	Single-ventricle physiology
Brain immaturity		Age at the time of surgery
Aortic atresia in HLHS
Abbreviations: HLHS, hypoplastic left heart syndrome; NIRS, near-infrared spectroscopy.

Other variables examined include circulatory arrest versus low-flow bypass, hypothermic blood pH management (alpha stat versus pH stat), hemodilution/hema- tocrit (25% vs. 35%), and maintaining regional cerebral perfusion during aortic arch reconstruction. None of these studies has identified definitively improved neurologic outcomes. Patients who underwent regional cerebral perfusion tended to have the worse outcome, and this technique was associated with new postoperative injury on brain MRI. These results suggest that although risks remain during the intraoperative period, a major burden of risk for acquired injury occurs outside of the operative period, but the possibility that unidentified intraoperative risk factors contribute cannot be excluded.

Risk Factors for Postoperative Brain Injury

Risk factors for postoperative brain injury (Table 14-2) include hypotension and hypoxemia related to low cardiac output syndrome, defined as a combination of clinical signs (tachycardia, oliguria, cold extremities, or cardiac arrest) and a >30% difference in arterial-mixed venous oxygen saturation or lactic acidosis (>4 mg/dL). Multiple studies have identified hypotension as a risk for new postoperative white matter injury, including low systolic blood pressure on admission, low mean blood pressure during postoperative day 1, and low diastolic blood pressure during postoperative days 1 to 2. One study showed that low regional cerebral oxygen saturation (<45%) measured by cerebral near-infrared spectroscopy for longer than 3 hours in the postoperative period was a risk for new ischemic injury. However, this has not been confirmed in subsequent studies. In many series, singleventricle physiology carries a higher risk of postoperative brain injury, which correlates with the higher postoperative hemodynamic instability, morbidity, and mortality seen in these patients. Decreased postoperative systemic venous oxygen saturation, indicative of low cardiac output, is a risk factor for poor neurodevelopmental outcome at 1 year. Studies are emerging suggesting that older age at the time of surgery for complex cardiovascular defects, such as hypoplastic left heart syndrome, leads to increased postoperative new white matter injury. Persistence of sub- optimal hemodynamics likely contributes to these findings; however, further prospective, randomized studies are needed to assess the true impact of timing to surgery on brain injury and neurodevelopmental outcome.

■ NEURODEVELOPMENTAL OUTCOME

Transposition of the Great Arteries

Formal intelligence testing of children with d-transposi- tion of the great arteries repaired in the neonatal period shows IQ scores that are in the normal range and only slightly below population norms (8-year WISC-III FullScale mean IQ 97.1 ± 15.3). Despite relatively modest differences in IQ from the population average and overall good general health status, a high percentage of children (one in five) in the Boston Circulatory Arrest Study were judged to have behavioral problems by parents and teachers: 37% required remedial education services, and 10% had repeated a grade. A more recent analysis of the same cohort in adolescence demonstrated continued deficits in memory, executive function, and attention with a high percentage requiring remedial services: one in three received tutoring; one in four received special education, occupational therapy, or psychotherapy; and one in six had repeated a grade. Not surprisingly, a lower socioeconomic status was significantly related to lower neuropsychiatric test scores, and parental IQ was a significant predictor of math and reading scores.

Hypoplastic Left Heart Syndrome

Patients with single-ventricle physiology usually require a three-stage surgical approach with exposure to cyanosis and embolic risk over the first few years of life. As one might expect, reported outcomes are worse than for d-transposition patients but still fall within the normal range. In the largest series reported to date (N = 83), the Mental Development Index was 90 (range 50 to 129) at 1 year of age. The median Psychomotor Development Index was lower at 73 (range 50 to 117). This pattern is seen in virtually every study of neurodevelopmental outcome at 1 year of age and likely represents a performance assessment heavily weighted toward the ability to walk. Other, smaller studies report similar values with a trend toward slight improvement with more contemporary series. Risk factors for poor neurodevelopmental outcome are similar to those discussed above and include the presence of a genetic syndrome, younger gestational age at birth, preoperative instability (increased base deficit, need for preoperative intubation), and postoperative instability (lower superior vena cava saturation). As the children are assessed at older ages, the prevalence of subtle abnormalities increases, with many children (~30%) needing special education services.

Neurodevelopmental Signature of Complex Congenital Cardiovascular Disease

Mortality rates for most congenital cardiac defects have fallen below 10%, with hospital mortality for many lesions <3%. These children grow up with overall physical and psychosocial health status similar to the general population. However, despite intelligence testing in the normal population range, many have a prevalence of pervasive but subtle cognitive problems that some have termed a “neurodevelopmental signature of complex congenital heart disease.” These children show behavioral and attention problems that are often not detected on standardized testing but result in poor school performance. This developmental signature in many ways resembles problems observed in survivors of premature birth. Among a large cohort of infants who were followed prospectively after undergoing surgery as infants, abnormalities on neurologic exam at school entry were present in 28%, although less than 5% were severe. Most of the abnormalities involved fine motor coordination and tone. Cognitive difficulty and behavioral problems were identified in 30%, with a large percentage requiring remedial services.

■ CONCLUSIONS

In this chapter, we have reviewed the major risk factors for adverse neurologic outcome, noting that they vary in both timing and mechanism. Some patient-specific factors, such as genetic defects that result in both abnormal heart and brain development and parental education and socioeconomic status, may not be amenable to intervention. However, other factors have been identified that may provide opportunities for intervention at different stages, including during fetal life, and before, during, and after surgery. Trials are under way to assess fetal interventions to improve left ventricular growth in prenatally diagnosed aortic stenosis and to prevent progression to hypoplastic left heart syndrome and the resulting cerebral perfusion abnormalities during fetal life. Another potential opportunity for investigation involves determining the optimal time for surgery. The identification of delayed brain development as a risk for perioperatively acquired injury suggests the possibility of delaying nonemergency surgery. The issue is complicated, however, because the longer surgery is delayed in newborns with d-transposition of the great arteries, the greater the likelihood of white matter injury. Other studies have shown that both younger and older ages at surgery are risk factors for injury. The combination of brain immaturity and vulnerability to white matter injury suggests that neuroprotection must be tailored toward the immature brain. In animal models, the preoligodendrocytes, precursors to the cells responsible for forming white matter and myelinating the brain, are uniquely vulnerable to hypoxia ischemia through specific mechanisms of glutamate excitotoxicity and oxidative stress. Despite well-performed trials with long-term follow-up, no cardiopulmonary bypass strategy has emerged as being clearly superior.

Although a wealth of information is emerging from sensitive advanced MRI studies, the relationship of perioperative acquired brain injury to long-term neurodevelopmental outcome has yet to be determined. In the premature infant and term infant with birth asphyxia, advanced MRI is the most sensitive study for detection of acquired injuries and is predictive of neurodevelopmental outcome. Studies examining whether perioperative MRI findings predict outcome in large cohorts of newborns with congenital cardiovascular disease are urgently needed. At the same time, for MRI to become sufficiently sensitive for establishing early outcome variables for use in interventional trials, investigators must determine which MRI method and findings are most sensitive to late outcome, when to perform MRIs in the course of the infant’s treatment, and how to describe the findings in a coherent, reproducible, and uniform manner.

Finally, although many patients show subtle or even pronounced neurodevelopmental deficits, it is clear that many are not receiving appropriate rehabilitation services (speech, occupational, or physical therapy). Caretakers should ensure that at-risk patients are evaluated and advocate for provision of appropriate services, as early recognition and intervention will improve late functional outcome.