Mechanisms of perinatal brain injury

Semin Neonatol 2000; 5: 3–16 doi:10.1053/siny.1999.0112, available online at http://www.idealibrary.com

Mechanisms of perinatal brain injury Terrie E. Inder* and Joseph J. Volpe† *Department of Paediatrics, Christchurch School of Medicine and Hospital, University of Otago, Christchurch, New Zealand; †Department of Neurology, Harvard Medical School and Children’s Hospital, Boston, MA, USA Key words: newborn, infant, prematurity, hypoxic–ischaemic injury, brain, periventricular leukomalacia, perinatal asphyxia, intraventricular haemorrhage

This article is focused on the mechanisms underlying primarily ischaemic/reperfusion brain injury in both the term and premature infant. Although the mechanisms involved include similar initiating events, principally ischaemia–reperfusion, and similar final common pathways to cell death, particularly free radical-mediated events, there are certain unique maturational factors influencing the type and pattern of cellular injury. We will therefore initially describe the physiological and cellular/molecular mechanisms of brain injury in the term infant, followed by the mechanisms in the premature infant.  2000 Harcourt Publishers Ltd

Mechanisms of brain injury in the term infant The major established cause of neurological morbidity in the term infant is hypoxic–ischaemic brain injury. The sequence of pathophysiological events leading to neonatal hypoxic-ischaemic brain injury is ultimately dominated by the occurrence of global cerebral ischaemia. In the following models we discuss the neuropathology and pathogenesis of this injury, and the opportunities for prevention or amelioration afforded by insights into mechanisms. Neuropathology The neuropathological features of neonatal hypoxic–ischaemic encephalopathy are determined by factors such as the severity, temporal pattern and type of insult, as well as the gestational age and metabolic status, including temperature, of the infant. The most common variety of neuronal Correspondence to: Joseph J. Volpe, Neurologist-in-Chief, Neurology Department, Fegan 1103, Children’s Hospital, 300 Longwood Avenue, Boston, MA 02115, USA. Tel: 617–355–6386; Fax: 617–730–0416; Email: [email protected]

1084–2756/00/010003+14 r35.00/0

injury resulting from hypoxic–ischaemic injury in the term infant is ‘selective neuronal necrosis’. Neurons demonstrate the greatest sensitivity to hypoxic–ischaemic injury in the term (vs the preterm) infant. The first observable change in the neuron is cytoplasmic vacuolation, caused by mitochondrial swelling, occurring within 5–30 min after the onset of the ischaemia [1,2]. Among term infants, neurons of the CA1 region of the hippocampus (Sommer’s sector), deeper layers of the cerebral cortex, putamen, thalamus and cerebellar Purkinje cells are injured most frequently by hypoxic–ischaemic insult [3–5]. Neuronal necrosis is most prominent in the watershed regions of the cerebral cortex and in the depths of sulci, reflecting the greater effect of ischaemia in these regions. ‘Parasagittal cerebral injury’ is an important ischaemic lesion of the full-term infant, and is characterized by injury to the cerebral cortex and subcortical white matter in the parasagittal and superomedial aspects of the cerebral convexities. The injury is bilateral, and usually symmetrical. ‘Status mamoratus’ describes a pattern of injury involving the basal ganglia and thalamus characterized initially by neuronal loss, gliosis and hypermyelination. In full form, this lesion develops in a much smaller © 2000 Harcourt Publishers Ltd

4

T. E. Inder and J. J. Volpe

Figure 1. A cascade of pathological processes leading to delayed neuronal death is set in motion by a reversible global hypoxic–ischaemic insult

number of infants than might be predicted by the relative frequency of neuronal injury in basal ganglia and thalamus observed in the neonatal period. The hypermyelination is apparent from approximately 8 months of age, and produces the characteristic marbled appearance of the basal ganglia. Other patterns of hypoxic-ischaemic injury include focal and multifocal ischaemic brain necrosis related to an interruption of blood flow in the distribution of a single vessel or multiple vessels. This ‘lesion’ usually takes the form of a single infarction and may be followed by porencephaly. Multiple infarctions may be followed by hydranencephaly or multicystic encephalomalacia [4]. Periventricular leukomalacia will be discussed in detail in a later section.

neuronal loss in the hippocampus and parasagittal cortex [6]. In contrast, three repeated episodes of 10 min of ischaemia, repeated at 1-h and 5-h intervals, result in marked neuronal damage in the striatum and less neuronal damage in the cortex [7]. Similarly, repeated umbilical cord occlusion results in damage almost confined to the striatum [8], whereas a single 10 min of umbilical cord occlusion causes hippocampal neuronal loss without striatal damage [9]. Cerebral temperature can also alter the pathological outcome of a hypoxic–ischaemic insult. Lowering cerebral temperature by 2–3C during hypoxic–ischaemic insults can be neuroprotective [10], while hyperthermia has been demonstrated to worsen outcome [11]. A reduction of temperature following the insult for 72 h after a hypoxic–ischaemic insult in the 21-day-old rat markedly protected the brain [12].

Pathogenesis Mechanisms of neuronal death Factors influencing the severity and regional distribution of damage

Factors such as the severity, pattern and type of insult, as well as the gestational age and metabolic status, including temperature, of the infant are crucial determinants of the neuropathology of hypoxic–ischaemic brain injury in the newborn. The majority of studies in animal models have investigated the effect of a single insult. Repeated insults appear to affect both the severity and the regional nature of neuronal injury. A single 30-min ischaemic insult in the fetal sheep causes severe

Following a reversible global insult, it appears that neuronal death occurs in two major phases [13–15] (Fig. 1). The mechanisms involved in each of the phases are different, and influenced by the nature and the severity of the insult (Fig. 2). The more severe the insult, the greater the proportion of cell death at that time. Nevertheless, it is now clear that secondary damage accounts for a significant proportion of final cell loss even after very severe insults. ‘Primary cell loss’ is related to cellular hypoxia, which leads to exhaustion of high-energy

Mechanisms of perinatal brain injury

5

Figure 2. A representative example of the time course of biophysical parameters following a single 30-min episode of cerebral hypoperfusion in a term fetal sheep (reproduced with permission Dr Chris Williams, Research Centre for Developmental Medicine and Biology, University of Auckland, New Zealand). During bilateral carotid occlusion, EEG activity is suppressed and there is a primary rise in cortical impedance (CI). The concentrations of both the total cerebral haemoglobin (tHB) and cytochrome oxidase (CytO2) as measured by near infrared spectroscopy (NIRS) fell significantly. Following release of the occlusion, EEG activity remained suppressed, while CI partially recovered to 106% above baseline. A significant but transient hyperaemia developed (a) following the insult. The onset of the secondary phase commenced with a secondary hyperaemia at 12.62.0 h. The rise in tHb preceded a secondary rise in CI at 290100 min. This period was associated with an increased amplitude in EEG, characterized by increased spike wave and poly-spike activity. During the secondary phase, CytO2 continued to fall reaching a minimum of
4.41.1 mmol/1, suggesting progressive cell death during this phase. At the end of the experiment (4 days) the EEG activity remained depressed and the CI returned to a level of 1083% above baseline. Histological analysis showed infarction of the parasagittal cortex and severe neuronal loss in the hippocampus with moderate loss in the thalamus.

metabolism (primary energy failure) and cellular depolarization. During primary energy failure, studies suggest that there are three closely interrelated mechanisms involved in the death of neurons. Firstly, depolarization due to hypoxia causes an influx of sodium and a lesser efflux of potassium with passive chloride entry along the electrochemical gradient. This, in turn, favours further cation and water entry leading to cell swelling and, if sufficiently severe, cell lysis [16]. Secondly, intracellular calcium accumulation occurs due to both excessive entry of calcium due to activation of the voltage-dependent and neurotransmitter-associated ion channels, and also the failure during hypoxia of the energy-dependent process of calcium removal

by the sodium–calcium pump [17–20]. Thirdly, extracellular glutamate accumulation (excitotoxicity) due to failure of energy-dependent re-uptake and excessive release because of membrane depolarization not only contributes to acute cell swelling through opening receptor mediated Na + /K + channels [21,22], but is also a key mechanism stimulating intracellular calcium accumulation through the N-methyl-D-aspartate (NMDA)receptor-channel complex [19,23]. Further cell membrane damage may occur [24] due to the action of free radicals in the immediate reperfusion phase. However, many neurons do not die during the primary phase of neuronal death. Rather, a cascade of pathologic processes is triggered and

6

leads to further loss of neurons, starting some hours later and extending over several days (Fig. 2). This secondary loss of neurons is termed secondary or delayed neuronal death. This phase may be associated with hyperexcitability and cytotoxic oedema from about 6–100 h after the injury [25]. The mechanisms involved in delayed neuronal death include excitotoxicity [26,27], apoptosis [28,29] and cytotoxic actions of activated microglia. The associated seizures may also contribute importantly to delayed neuronal loss. The delayed phase is associated with a marked hyperaemia, increased seizure activity, cytotoxic oedema, mitochondrial failure [30], NO synthesis [31] and an accumulation of excitotoxins. A prominent degree of neuronal injury has been associated with the development of intense seizures and changes in cerebral blood volume and flow in the near-term fetal sheep model of asphyxial brain injury [30–32]. In this model, two major components of cerebrovascular response were also confirmed; impairment during the immediate reperfusion period, primarily in the parasagittal cortex, followed by vasodilatation measured by NIRS [30,31], ultrasonic flow probes [33], and microspheres (unpublished observation) (Fig. 2). This secondary hyperaemia appears neuro-protective and is, in part, mediated by nitric oxide release [31]. Excitotoxins accumulate during both the primary and delayed phases of injury. Accumulation appears to be greatest in the secondary phase and appears secondary to the glial and energy failure. This accumulation of excitotoxins is associated with the development of intense NMDA receptor-mediated seizure activity. This secondary phase has been confirmed with magnetic resonance spectroscopy studies in infants with moderate-to-severe asphyxial insults. The infants showed normal cerebral oxidative metabolism shortly after birth, with some developing evidence of secondary energy failure associated with a severe mortality. In survivors, the degree of secondary energy failure at 24–48 h strongly predicted neurodevelopmental outcome at 1 year of age [15,34].

Mechanisms of brain injury in the premature infant Very low birth-weight (VLBW) infants now represent approximately 1.2% of all live-born infants,

T. E. Inder and J. J. Volpe

increasing slightly as a proportion of total births over the last three decades [35,36]. However, more notably, with recent improvements in obstetric and neonatal management, the survival of VLBW infants has increased to nearly 90% [36–38]. This increased survival has been greatest in the extremely low birth-weight (ELBW) cohort who weigh less than 1000 g and who are at greatest risk of later neurological deficits [37,39]. With this improved survival there has been no reduction in the documented rates of neurological disability occurring in premature infants, with a recognized increase in the absolute numbers of infants with long-term neurological sequelae [40]. For all VLBW survivors, 5–15% later exhibit major spastic motor deficits, while an additional 25–50% exhibit less prominent developmental disabilities, involving not only motility but also cognition and behaviour, with school disturbance the nearly uniform result [4,40–45]. Improvement of the detection, diagnosis and management of neurological injuries and, ultimately, formulation of strategies to prevent the accompanying morbidity require an understanding of the pathogenesis of the brain injury of the premature infant.

Neuropathological substrates The principal neuropathological substrates for the neurological disturbances in the premature infants involve the cerebral white matter. These lesions, periventricular haemorrhagic infarction and periventricular leukomalacia, are the focus of this article. The influence of white matter injury on subsequent cerebral cortical gray matter development will also be discussed. Posthaemorrhagic hydrocephalus, selective neuronal injury, and focal cerebral ischaemic lesions are additional neuropathologies observed in premature infants, but relative to the two lesions noted above, these abnormalities are less important and are discussed elsewhere [4].

Periventricular haemorrhagic infarction Periventricular haemorrhagic infarction refers to haemorrhagic necrosis of periventricular white matter that is usually large and almost invariably asymmetric. The lesion most often coexists with

Mechanisms of perinatal brain injury

7

Table 1. Incidence of intraventricular haemorrhage in premature infants Years of Study*

Total number studied

Incidence

Late 1970s–early 1980s Late 1980s 1995 1997

1200 infants 449 VLBW infants 2544 VLBW infants 19 581 VLBW infants

39–49% 16% 32% 25%

*See text for references.

Neuropathology

45 All IVH Grade III/IV IPE

40 35 30 25 20 15 10 5 0

500–750 g

the incidence of intraparenchymal echodensity 4% (maximal incidence of 10.5% in infants <750 g). Thus, although there has been a decline in the incidence of all IVH since the early 1980s, it appears that the lesion remains a significant problem, particularly in the extremely low birthweight infant with a very high risk of associated periventricular haemorrhagic infarction.

751–1000 g

1001–1250 g 1251–1500 g

Figure 3. The incidence of all intraventricular haemorrhage, Grade II/IV and intraparenchyal echodensity (IPE) (Vermont-Oxford Database 1997, n=19 581).

IVH and, indeed, approximately 15–20% of all infants with IVH also exhibit periventricular haemorrhagic infarction [37,38]. The incidence of IVH in premature infants has declined over the last 25 years to a variable degree in most neonatal centers (Table 1). In well-studied series of premature infants subjected to routine computed tomography or ultrasound scans and studied in the late 70s and early 80s, the incidence of IVH ranged from approximately 35 to 50% [46–49]. During the late 1980s, the incidence had fallen to approximately 20% [4,50,51]. However, the incidence in the 1990s has remained unchanged, as shown by data from one large series of 19 581 very low birth-weight infants in 250 centres collected in the Vermont–Oxford Neonatal Database for 1997, in which the incidence of all IVH was 25%. The overall incidence of intraparenchymal echodensity was 4%, with the highest value of 11% in infants born with a birth-weight of less than 750 g [37] (Fig. 3). The Australian and New Zealand Neonatal Network shows similar data on 2544 very low birth-weight infants born in 1995 [38]. The incidence of all IVH was 32%, with

The neuropathology of periventricular haemorrhagic infarction is striking, and consists of a relatively large region of haemorrhagic necrosis in the periventricular white matter, just dorsal and lateral to the external angle of the lateral ventricle. The necrosis is strikingly asymmetric — in the largest series reported, 67% of such lesions were exclusively unilateral, and in virtually all of the remaining cases, grossly asymmetric, even though bilateral [52]. Approximately one-half of the lesions are extensive and involve the periventricular white matter from frontal to parieto-occipital regions; the remainder are more localized. Approximately 80% of cases are associated with large IVH, and commonly (and mistakenly) the parenchymal haemorrhagic lesion is described as being an ‘extension’ of the haemorrhage. That simple extension of blood into cerebral white matter from germinal matrix or lateral ventricle does not account for the periventricular haemorrhagic necrosis has been shown by several neuropathological studies [46,52–57]. Microscopic study of this periventricular haemorrhagic necrosis indicates that the lesion is a ‘haemorrhagic infarction’ [46,52–54,56–58]. The careful studies of Gould et al. and of Takashima et al. [53,58] emphasize that the haemorrhagic component of the infarction tends to be most concentrated near the ventricular angle where the medullary veins draining the cerebral white matter become confluent and ultimately join the terminal vein in the subependymal region (Fig. 4). This topography and the microscopic appearance indicate that the periventricular haemorrhagic necrosis occurring in association with large IVH is, in fact, a venous infarction. This lesion is distinguishable neuropathologically from secondary haemorrhage into periventricular leukomalacia, the ischaemic, usually non-haemorrhagic, and symmetric lesion of periventricular white matter of the premature infant (see later). However, distinction of these two lesions in vivo often is very difficult, and indeed

8

Figure 4. Venous drainage of cerebral white matter (schematic appearance). The schematic diagram shows that the medullary veins, arranged in a fan-shaped distribution, drain blood from the cerebral white matter into the terminal vein which courses through the germinal matrix. (Adapted from Volpe JJ. Neurology of the Newborn, 3rd edition. Philadelphia: WB Saunders, 1995.)

overlap is common. (This overlap makes characterization of the fundamental lesions in vivo difficult.) In Table 2 we compare the basic features of these two periventricular white matter lesions of the premature infant. Pathogenesis The pathogenesis of periventricular haemorrhagic infarction appears in most cases to be related directly to the associated germinal matrix-IVH, on the basis of several observations from the largest reported study [52]. First, approximately 80% of the parenchymal lesions were observed in association with large (and usually asymmetric) IVH. Second, the parenchymal lesion invariably occurred on the same side as the larger amount of germinal matrix and intraventricular blood. Third, the parenchymal lesion developed and progressed after the occurrence of the IVH. The peak time of occurrence of the parenchymal lesion was the fourth postnatal day, i.e. when 90% of associated cases of IVH had already occurred [52]. The association of large asymmetric germinal matrix-IVH and the progression to ipsilateral periventricular haemorrhagic infarction has been confirmed [59,60]. These data raised the possibility that the IVH and/or its associated germinal matrix haemorrhage led to obstruction of the terminal veins and haemorrhagic venous infarction [4,52]. A similar conclusion had been suggested from a neuropathological study

T. E. Inder and J. J. Volpe

[53]. This pathogenetic formulation received strong support from Doppler determinations of blood flow velocity in the terminal vein during the evolution of the infarction in the living premature infant — obstruction of flow in the terminal vein by the ipsilateral germinal matrix haemorrhage was shown clearly [61]. Thus, the pathogenetic scheme that we consider to account for most examples of periventricular haemorrhagic infarction is shown in Figure 5. This scheme should be distinguished from that operative for haemorrhagic periventricular leukomalacia, in which secondary haemorrhage occurs into a prior region of leukomalacia. However, clearly the lesions could coexist. The frequency of coexistence of the two lesions is not known. Additionally, the two pathogenetic schemes could operate in sequence, i.e. periventricular leukomalacia could become haemorrhagic (and perhaps a larger area of injury) when germinal matrix or IVH subsequently causes venous obstruction. The prevention of periventricular haemorrhagic infarction thus relies on prevention of germinal matrix haemorrhage. The pathogenesis of germinal matrix haemorrhage is reviewed elsewhere [4], but the factors of importance are summarized in Table 3.

Periventricular leukomalacia Periventricular leukomalacia refers to necrosis of white matter in a characteristic distribution, i.e. in the white matter dorsal and lateral to the external angles of the lateral ventricles, involving particularly the centrum semiovale (frontal horn and body), and the optic (trigone and occipital horn) radiation [4]. The incidence of the disorder at autopsy varies considerably from one medical centre to another, i.e. from approximately 25 to 75%, but several facts remain clear: the lesion is observed particularly often: (1) in the premature infant; (2) in infants with postnatal survival of more than a few days; (3) in infants who also have IVH; (4) in infants with evidence of cardiorespiratory disturbance; and (5) in infants with evidence of antenatal/placental/ fetal infection [3,4,62–78]. The incidence of the disorder in living infants varies particularly according to the ultrasonographic criteria utilized, i.e. echodensities, echolucencies, ventriculomegaly, in different combinations. Because ultrasonography detects only the overt forms of periventricular white matter injury (see later discussion) and,

Mechanisms of perinatal brain injury

9

Table 2 Periventricular white matter lesions of the premature infant Term

Periventricular leukomalacia Periventricular haemorrhagic infarction

Symmetry

Grossly haemorrhagic

Probable site of circulatory disturbance

Predominantly symmetric Predominantly asymmetric

Uncommon Invariable

Arterial Venous

Table 3 Pathogenesis of germinal matrix haemorrhage* Intravascular factors (1) Fluctuating blood pressure (2) Increase in cerebral blood flow — systemic hypertension in the setting of a pressure-passive circulation; rapid volume expansion; hypercarbia (3) Increase in cerebral venous pressure — venous anatomical arrangement; labour and vaginal delivery; respiratory disturbance (4) Decrease in cerebral blood flow (followed by reperfusion) — systemic hypotension (5) Platelet and coagulation disturbance. Vascular factors (1) Tenous capillary integrity — the germinal matrix is an involuting remodelling capillary bed with deficient vascular lining and a large vascular and luminal area (2) Vulnerability of matrix capillaries to hypoxic–ischemic injury — the area has a high requirement for oxidative metabolism and lies at a vascular border zone. Extravascular factors (1) Deficient vascular support (2) Excessive fibrinolytic activity. Figure 5. Pathogenesis of periventricular haemorrhagic infarction. Formulation indicates a central role for germinal matrix-intraventricular haemorrhage in causation of the periventricular venous infarction. (Adapted from Volpe JJ. Neurology of the Newborn, 3rd edition. Philadelphia: WB Saunders, 1995.)

of course, because living infants as a group are different from autopsied infants, the reported incidences of periventricular leukomalacia based on ultrasonographic criteria in infants less than 1500 g birth-weight, i.e. approximately 5–15%, are much lower than those observed at autopsy. Incidences have appeared to be highest in premature infants of approximately 26–28 weeks of gestation [78,79]. Because rates of survival for more than a few days decline markedly in the most immature infants, and therefore the total number of infants of less than 28 weeks of gestation studied is small, it is unclear whether the incidence of periventricular leukomalacia peaks at approximately 28 weeks of gestation or continues to increase as gestational age

*Adapted from Volpe JJ. Neurology of the Newborn, 3rd edition. Philadelphia: W.B. Saunders, 1995.

decreases. The Vermont–Oxford Network data shows that the incidence of cystic PVL to be greatest in the infants of birth-weight 500–750 g with an incidence of 6%, compared with an incidence of 5, 3 and 2%, respectively, in the 751– 1000, 1001–1250 and 1251–1500 g infant [37]. The clinical consequences of periventricular leukomalacia are reviewed elsewhere [4]. A recent review of 12 studies assessing 272 infants with periventricular echolucency on cranial ultrasound to developmental outcome found that the more extensive the sonographically-defined white matter injury the greater the likelihood of severe motor and cognitive deficits [80]. Overall, 58% of the 272 infants developed cerebral palsy, compared with 2.6% of the 655 infants with normal head ultrasound scans. The risk of cerebral palsy was most closely associated with the involvement of the parieto-occipital region, and the risk for cerebral

10

T. E. Inder and J. J. Volpe

The pathological features of periventricular leukomalacia are distinctive and consist primarily of both ‘focal periventricular necrosis’ and more ‘diffuse’ cerebral white matter injury. The focal necrotic lesions occur primarily in the distribution of the end zones of the long penetrating arteries. The cellular neuropathology of the ‘focal component’ of PVL is characterized in the first 6–12 h after an acute hypoxic–ischaemic insult by coagulation necrosis with loss of all cellular elements at the sites of the focal periventricular lesion. Subsequent tissue dissolution and cavity formation occur over 1–3 weeks. These multiple, small cysts are demonstrable by cranial ultrasonography, if larger than 2–3 mm. The more diffuse regions of cerebral white matter injury have been emphasized particularly in studies with larger numbers of smaller infants with longer periods of postnatal survival [69,70,73,81– 84]. These more diffuse lesions less commonly undergo major cystic change and are more likely to be undetected by cranial ultrasonography during life [70]. The cellular hallmarks of the diffuse component of PVL are pyknotic glial nuclei (i.e. ‘acutely damaged glia’) and hypertrophic astrocytes [69,85,86].

[3,87–92]. The vessels penetrating the cerebral wall from the pial surface, the long penetrators, are derived from the middle cerebral artery and, to a lesser extent, the anterior or posterior cerebral artery, and terminate in the deep periventricular white matter. The pathogenesis for the more diffuse white matter injury, similarly, may relate in part to the development of the penetrating cerebral vasculature more peripherally [89–92]. Thus, as emphasized by Rorke [90], the penetrating cerebral arteries can be divided into the long penetrators that terminate deep in periventricular white matter, as discussed in relation to the periventricular focal necroses, and the short penetrators that extend only into subcortical white matter. At 24 to 28 weeks of gestation the long penetrators, have few side branches and infrequent intraparenchymal anastomoses with the short penetrators, which importantly also are few in number. Thus, end and border zones may exist at this time in the cerebral white matter relatively distant from the periventricular region. From 32 weeks to term, increases in the development of the short penetrators and in anastomoses between the long and short penetrators occur, and perhaps thereby lead to a decrease in border and end zones in both the subcortical and the periventricular white matter. A physiological correlate of these vascular anatomic factors appears to be the extremely low blood flow to cerebral white matter in the human premature newborn, first shown by positron emission tomography [93]. Xenon clearance studies also have documented low mean global cerebral blood-flow values in human premature infants of only approximately 10–12 ml/ 100 g/min [94–100].

Pathogenesis

Cerebral vascular regulation

The pathogenesis of periventricular leukomalacia is related to at least three major interacting factors that are postulated to result in ischaemia to the periventricular region and injury to the particularly vulnerable cerebral white matter. Other interacting factors, including maternal intrauterine infection/ inflammation and cytokine release and excitotoxicity, also appear to be important in the pathogenesis, as outlined below.

The vascular end zones just described thus render the premature infant’s brain particularly vulnerable to injury in the presence of cerebral ischaemia. Perhaps of particular importance in the genesis of impaired cerebral blood flow, and thereby cerebral ischaemia, is an impairment of cerebrovascular regulation in at least a subset of premature infants. Studies employing the invasive technique of radioactive Xenon clearance have shown that certain premature infants, generally clinically unstable, exhibit a pressure-passive cerebral circulation [98,101]. This pressure-passive circulation will produce a lowering of cerebral blood flow during any period of systemic hypotension. Clinically stable

palsy and an IQ<70 increased as the echolucencies replaced a greater proportion of the white matter. The finding that more than one-third of the infants with unilateral sonographic damage had diplegia indicates that cranial ultrasonography is a suboptimal indicator of white matter damage [80]. Neuropathology

Cerebrovascular anatomic factors The deep focal necrotic lesions of PVL occur in the areas that are considered arterial end zones

Mechanisms of perinatal brain injury

premature infants seem less likely to exhibit this apparent lack of cerebrovascular autoregulation [94,95,102,103]. Finally, it should be considered that even in the presence of intact cerebrovascular autoregulation, marked cerebral vasoconstriction or severe systemic hypotension could lead to sufficiently impaired cerebral blood flow to the periventricular vascular end zones to result in cerebral white matter injury. This explanation may account for the demonstrated relationship between marked hypocarbia or hypotension and periventricular leukomalacia in premature infants [95,104–107]. Moreover, hypotension has been shown to lead to cerebral white matter injury in the near term fetal lamb [108], and in related models of cerebral hypoperfusion in the preterm fetal lamb [109,110] and in the immature rat [111,112]. Intrinsic vulnerability of cerebral white matter of premature newborn An intrinsic vulnerability of the immature oligodendrocyte in cerebral white matter of the human infant, particularly regarding the more diffuse component of periventricular white matter injury, is suggested by experimental studies, by the rarity of the lesion at later ages, and by the relative cellular specificity of the diffuse injury, i.e. involving oligodendrocytes precursors or early differentiating oligodendroglia [35,63]. We have studied this cell type in a defined system in culture, demonstrating that the immature oligodendrocytes in the human premature brain are exquisitely vulnerable to free radical attack, whereas the mature oligodendrocyte is totally resistant [113–115]. Moreover, clinically safe free radical scavengers, e.g. vitamin E, totally prevented the oligodendroglial death caused by free radical attack. The mechanism of the vulnerability of the immature oligodendrocyte is not known. It may be due to an increased production of free radicals via iron catalysis with iron uptake in the differentiating oligodendroglia being necessary for subsequent myelination [116–118], or from release of iron from an accompanying IVH. There may also be a reduction in antioxidant defenses with documentation of a delay in the development of at least one crucial component of antioxidant enzymatic defense, i.e. catalase, in immature oligodendroglia. This maturation-dependent vulnerability thus may be critical for the predilection of this lesion for the human brain early in life and the

11

absence of the lesion in similar form after oligodendroglial maturation and myelination occur.

Intrauterine infection/inflammation and cytokine release A role for maternal/fetal infection, endotoxin and presumably endotoxin-mediated cytokine release in the pathogenesis of periventricular white matter injury was suggested initially by neuropathological/epidemiological studies of infant brain and related experimental studies of Gilles and coworkers [72,119,120]. A recent demonstration of cerebral white matter injury in fetal rabbits after the induction of maternal intrauterine infection is consistent with earlier observations [121]. Several recent human studies lend further support to a contributory role for such factors in the pathogenesis of PVL. Thus, the incidence of PVL and/or cerebral plasy in premature infants is increased in the presence of: (1) evidence for maternal/placental/ fetal infection [74,78,122]; (2) elevated levels of interleukin-6 in cord blood [76]; and (3) elevated levels of interleukin-6 and interleukin-1 in amniotic fluid [123]. Moreover, although potentially a secondary effect of ischaemia, also possibly supportive of a relation to intrauterine infection and cytokines is the demonstration within lesions of PVL of interleukin-6 and tumor necrosis factor (TNF-
) [77,124,125]. In general, the role of cytokines and inflammatory cells in the pathogenesis of periventricular leukomalacia remains unclear. If these factors do play a role, the relative importance of ischaemia/reperfusion or infection or both in activation of these factors requires elucidation.

The potential role of glutamate A possible role for excess extracellular glutamate in the pathogenesis of PVL is suggested by four basic observations. Firstly, the earliest neuropathological feature of focal periventricular leukomalacia is coagulation necrosis and disruption of axones [62,126]. Secondly, neurons, and presumably axones, contain millimolar concentrations of glutamate [114,127–129], which could leak into tissue upon disruption. Thirdly, glutamate causes glutathione depletion in oligodendrocytes and thereby free radical-mediated cell death [114], as described earlier. Fourthly, activation of the AMPA/kainate

12

T. E. Inder and J. J. Volpe

Table 4 Quantitative 3D-MRI volumes [mean  standard error (cc)] of cerebral tissues at term in preterm infants with evidence of white matter injury (WMI), preterm infants with no evidence of WMI and normal term infants Quantitative 3D- MRI volumes [mean  SEM (cc)] Preterm with WMI (n=11) Cortical GM Myelinated WM CSF

Preterm with no WMI (n=10)

152.115.7 16.8 2.3 62.8 6.1

Normal term (n=14)

222.1 9.7 22.6 3.0 56.010.4

type of glutamate receptor can lead to oligodendroglial death [130]. Moreover, our recent data indicate that this AMPA/kainate form of oligodendroglial death occurs only in the developing oligodendrocyte and not in the mature oligodendrocyte. Thus, glutamate may lead to toxicity to developing oligodendrocytes by two mechanism, one non-receptor-mediated and the other, receptor-mediated.

The influence of PVL on subsequent cerebral development The full extent of the role of periventricular white matter injury in the genesis of intellectual deficits in the premature infant, even the infant without prominent spasticity, remains to be clarified completely. Using an advanced quantitative volumetric 3D-MRI technique, we measured brain tissue volumes at term in preterm infants with early ultrasonographic and MRI evidence of periventricular white matter injury (WMI, mean GA at birth 28.71.7 weeks, n=11), in preterm infants with normal neonatal imaging studies (no WMI, mean GA at birth 28.60.7 weeks, n=10) and in control term infants (n=14). We documented that preterm infants with early periventricular white matter lesions had a marked reduction at term in volume of cerebral cortical grey matter, of similar magnitude to the reduction in myelinated white matter. A compensatory increase in CSF volume, both ventricular and extracerebral, was also found (Table 4). In this study, periventricular WMI in the premature infant was shown for the first time to impair both subsequent cerebral cortical development and myelination. These findings may provide insight into the anatomical correlate for the intellectual deficits associated with periventricular white matter injury in the premature infant. The reduction in cortical grey matter volume demonstrated

218.85.6 26.93.0 32.93.5

ANOVA p value 0.001 0.03 0.002

raises the possibility that the PVL is causally related to a disturbance in subsequent cortical neuronal development.

Conclusions The understanding of the pathogenesis of cerebral injury in hypoxic–ischaemic injury in the term infant has greatly advanced within the last decade, with trials of neuronal rescue therapies using cerebral hypothermia underway. In the premature infant there has been greatest progress toward prevention of brain injury in periventricular haemorrhagic infarction, but with the advent of new technologies, especially near-infrared spectroscopy and advanced magnetic resonance imaging techniques, and of new insights into the cellular basis for oligodendroglial vulnerability, there is improved hope for prevention of periventricular leukomalacia. References 1 Brown AW, Brierley JB. The earliest alterations in the rat neurons and astrocytes after anoxia-ischemia. Acta Neuropathol 1973; 23: 9–22. 2 Kim SU. Brain hypoxia studied in mouse central nervous system cultures. 1. Sequential cellular changes. Lab Invest, 1975; 33: 658–669. 3 Larroche JC. Developmental Pathology of the Neonate. New York: Excerpta Medica, 1977. 4 Volpe JJ. Neurology of the Newborn, 3rd edition. Philadelphia: WB Saunders, 1995. 5 Rivkin MJ. Hypoxic-ischemic brain injury in the term newborn. Neuropathology, clinical aspects, and neuroimaging. Clin Perinatol 1997; 24: 607–625. 6 Williams CE, et al. Outcome after ischemia in the developing sheep brain: an electroencephalographic and histological study. Ann Neurol 1992; 31: 14–21. 7 Mallard EC, et al. Frequent episodes of brief ischemia sensitize the fetal sheep brain to neuronal loss and induce striatal injury. Pediatr Res 1993; 33: 61–65.

Mechanisms of perinatal brain injury

8 Mallard EC, et al. Repeated asphyxia causes loss of striatal projection neurons in the fetal sheep brain. Neuroscience 1995; 65: 827–36. 9 Mallard EC, et al. Transient umbilical cord occlusion causes hippocampal damage in the fetal sheep. Am J Obstet Gynecol 1992; 167: 1423–1430. 10 Busto R, et al. The importance of brain temperature in cerebral ischemic injury. Stroke 1989; 20: 1113–1114. 11 Minamisawa H, Smith ML, Siesjo BK. The effect of mild hyperthermia and hypothermia on brain damage following 5, 10, and 15 minutes of forebrain ischemia. Ann Neurol 1990; 28: 26–33. 12 Sirimanne ES, et al. The effect of prolonged modification of cerebral temperature on outcome after hypoxicischemic brain injury in the infant rat. Pediatr Res 1996; 39: 591–597. 13 Gluckman PD, Williams CE. When and why do brain cells die? Dev Med Child Neurol 1992; 34: 1010–1014. 14 Lorek A, et al. Delayed (‘‘secondary’’) cerebral energy failure after acute hypoxia-ischemia in the newborn piglet: continuous 48-hour studies by phosphorus magnetic resonance spectroscopy. Pediatr Res 1994; 36: 699–706. 15 Penrice J, et al. Proton magnetic resonance spectroscopy of the brain in normal preterm and term infants, and early changes after perinatal hypoxia-ischemia. Pediatr Res 1996; 40: 6–14. 16 Williams CE, et al. Pathophysiology of perinatal asphyxia. Clin Perinatol 1993; 20: 305–325. 17 Orrenius S, et al. Calcium ions and oxidative cell injury. Ann Neurol 1992; 32(Suppl.): S33–S42. 18 du Plessis AJ, Johnston MV. Hypoxic-ischemic brain injury in the newborn. Cellular mechanisms and potential strategies for neuroprotection. Clin Perinatol 1997; 24: 627–654. 19 Choi DW. Calcium: still center-stage in hypoxicischemic neuronal death. Trends Neurosci 1995; 18: 58–60. 20 Fletcher EJ, Lodge D. New developments in the molecular pharmacology of alpha-amino-3-hydroxy-5-methyl4-isoxazole propionate and kainate receptors. Pharmacol Ther 1996; 70: 65–89. 21 Marret S, et al. Effect of ibotenate on brain development: an excitotoxic mouse model of microgyria and posthypoxic-like lesions. J Neuropathol Exp Neurol 1995; 54: 358–370. 22 Johnston MV. Neurotransmitters and vulnerability of the developing brain. Brain Dev 1995; 17: 301–306. 23 Gunn AJ, et al. Flunarizine, a calcium channel antagonist, is partially prophylactically neuroprotective in hypoxicischemic encephalopathy in the fetal sheep. Pediatr Res 1994; 35: 657–663. 24 Goplerud JM, Mishra OP, Delivoria-Papadopoulos M. Brain cell membrane dysfunction following acute asphyxia in newborn piglets. Biol Neonate 1992; 61: 33–41. 25 Williams CE, Gunn A, Gluckman PD. Time course of intracellular edema and epileptiform activity following prenatal cerebral ischemia in sheep. Stroke 1991; 22: 516–521. 26 Tan WK, et al. Suppression of postischemic epileptiform activity with MK-801 improves neural outcome in fetal sheep. Ann Neurol 1992; 32: 677–682.

13

27 Tan WK, et al. Accumulation of cytotoxins during the development of seizures and edema after hypoxicischemic injury in late gestation fetal sheep. Pediatr Res 1996; 39: 791–797. 28 MacManus JP, et al. Global ischemia can cause DNA fragmentation indicative of apoptosis in rat brain. Neurosci Lett 1993; 164: 89–92. 29 Beilharz EJ, et al. Mechanisms of delayed cell death following hypoxic-ischemic injury in the immature rat: evidence for apoptosis during selective neuronal loss. Brain Res Mol Brain Res 1995; 29: 1–14. 30 Marks KA, et al. Delayed vasodilation and altered oxygenation after cerebral ischemia in fetal sheep. Pediatr Res 1996; 39: 48–54. 31 Marks KA, et al. Nitric oxide synthase inhibition attenuates delayed vasodilation and increases injury after cerebral ischemia in fetal sheep. Pediatr Res 1996; 40: 185–191. 32 Mallard EC, et al. Neuronal damage in the developing brain following intrauterine asphyxia. Reprod Fertil Dev 1995; 7: 647–653. 33 Gunn AJ, et al. Dramatic neuronal rescue with prolonged selective head cooling after ischemia in fetal lambs. J Clin Invest 1997; 99: 248–256. 34 Roth SC, et al. Relation of deranged neonatal cerebral oxidative metabolism with neurodevelopmental outcome and head circumference at 4 years. Dev Med Child Neurol 1997; 39: 718–725. 35 Volpe JJ. Brain injury in the premature infant. Neuropathology, clinical aspects, pathogenesis, and prevention. Clin Perinatol 1997; 24: 567–587. 36 Hack M, Friedman H, Fanaroff AA. Outcomes of extremely low birth weight infants. Pediatrics 1996; 98: 931–937. 37 Horbar JD. Vermont-Oxford Network 1997 Database Summary. Burlington, Vermont: Vermont-Oxford Network, 1997. 38 Network AaNZN. Australian and New Zealand Neonatal Network Series number 2. Sydney: Australian Institute of Health and Welfare Perinatal Statistics Unit 1995. 39 The Victorian Infant Collaborative Study Group. Improved outcome into the 1990s for infants weighing 500–999 g at birth. Arch Dis Child Fetal Neonatal Ed, 1997; 77: F91–F94. 40 Pharoah PO, Platt MJ, Cooke T. The changing epidemiology of cerebral palsy. Arch Dis Child Fetal Neonatal Ed, 1996; 75: F169–F173. 41 Pharoah PO, et al. Birthweight specific trends in cerebral palsy. Arch Dis Child 1990; 65: 602–606. 42 Piecuch RE, et al. Outcome of extremely low birth weight infants (500 to 999 grams) over a 12-year period. Pediatrics 1997; 100: 633–639. 43 Powls A, et al. Visual impairment in very low birthweight children. Arch Dis Child Fetal Neonatal Ed, 1997; 76: F82–F87. 44 Skranes JS, et al. Cerebral magnetic resonance imaging and mental and motor function of very low birth weight children at six years of age. Neuropediatrics 1997; 28: 149–154. 45 Singer L, et al. A longitudinal study of developmental outcome of infants with bronchopulmonary dysplasia and very low birth weight. Pediatrics 1997; 100: 987–993.

14

46 McMenamin JB, Shackelford GD, Volpe JJ. Outcome of neonatal itravetricular haemorrhage with periventricular echodense lesions. Ann Neurol 1984; 15: 285–290. 47 Levene MI, Wigglesworth JS, Dubowitz V. Cerebral structure and intraventricular hemorrhage in the neonate: a real time ultrasound study. Arch Dis Child 1981; 56: 416–424. 48 Holt PJ, Allan WF. The natural history of ventrisular dilatation in neonatal intraventricular hemorrhage and its therapeutic implication. Ann Neurol 1981; 10: 293. 49 Ahmann PA, et al. Intraventricular hemorrhage in the high-risk preterm infant: incidence and outcome. Ann Neurol 1980; 7: 118–124. 50 Paneth N, et al. Incidence and timing of germinal matrix/intraventricular hemorrhage in low birth weight infants. Am J Epidemiol 1993; 137: 1167–1176. 51 Philip AG, et al. Intraventricular hemorrhage in preterm infants: declining incidence in the 1980s. Pediatrics 1989; 84: 797–801. 52 Guzzetta F, et al. Periventricular intraparenchymal echodensities in the premature newborn: critical determinant of neurologic outcome. Pediatrics 1986; 78: 995–1006. 53 Gould SJ, et al. Periventricular intraparenchymal cerebral haemorrhage in preterm infants: the role of venous infarction. J Pathol 1987; 151: 197–202. 54 Flodmark O, et al. Correlation between computed tomography and autopsy in premature and full-term neonates that have suffered perinatal asphyxia. Radiology 1980; 137: 93–103. 55 Armstrong DL, Sauls CD, Goddard-Finegold J. Neuropathologic findings in short-term survivors of intraventricular hemorrhage. Am J Dis Child 1987; 141: 617–621. 56 Volpe JJ, et al. Positron emission tomography in the newborn: extensive impairment of regional cerebral blood flow with intraventricular hemorrhage and hemorrhagic intracerebral involvement. Pediatrics 1983; 72: 589–601. 57 Rushton DI, Preston PR, Durbin GM. Structure and evolution of echo dense lesions in the neonatal brain. A combined ultrasound and necropsy study. Arch Dis Child 1985; 60: 798–808. 58 Takashima S, Mito T, Ando Y. Pathogenesis of periventricular white matter hemorrhages in preterm infants. Brain Dev 1986; 8: 25–30. 59 Perlman JM, et al. Relationship between periventricular intraparenchymal echodensities and germinal matrixintraventricular hemorrhage in the very low birth weight neonate. Pediatrics 1993; 91: 474–480. 60 Ment LR, et al. Risk factors for early intraventricular hemorrhage in low birth weight infants. J Pediatr 1992; 121: 776–783. 61 Taylor GA. Effect of germinal matrix hemorrhage on terminal vein position and patency. Pediatr Radiol 1995; 25 (Suppl. 1): S37–S40. 62 Banker BQ, Larroche JC. Periventricular leukomalacia of infancy. Arch Neurol 1962; 7: 386–410. 63 Back SA, Volpe JJ. Cellular and molecular pathogenesis of periventricular white matter injury. MRDD Res Rev 1997; 3: 96–107. 64 Barth PG, et al. On the relationship between germinal layer haemorrhage and telencephalic leucoencephalopathy in the preterm infant. Neuropadiatrie 1980; 11: 17–26.

T. E. Inder and J. J. Volpe

65 Carson SC, et al. Value of sonography in the diagnosis of intracranial hemorrhage and periventricular leukomalacia: a postmortem study of 35 cases. AJR Am J Roentgenol 1990; 155: 595–601. 66 Dammann O, Leviton A. Maternal intrauterine infection, cytokines, and brain damage in the preterm newborn. Pediatr Res 1997; 42: 1–8. 67 de Vries LS, Eken P, Dubowitz LM. The spectrum of leukomalacia using cranial ultrasound. Behav Brain Res 1992; 49: 1–6. 68 de Vries LS, et al. Correlation between the degree of periventricular leukomalacia diagnosed using cranial ultrasound and MRI later in infancy in children with cerebral palsy. Neuropediatrics 1993; 24: 263–268. 69 Golden JA, et al. Frequency of neuropathological abnormalities in very low birth weight infants. J Neuropathol Exp Neurol 1997; 56: 472–478. 70 Hope PL, et al. Precision of ultrasound diagnosis of pathologically verified lesions in the brains of very preterm infants. Dev Med Child Neurol 1988; 30: 457–471. 71 Iida K, Takashima S, Ueda K. Immunohistochemical study of myelination and oligodendrocyte in infants with periventricular leukomalacia. Pediatr Neurol 1995; 13: 296–304. 72 Leviton A, Gilles FH. Acquired perinatal leukoencephalopathy. Ann Neurol 1984; 16: 1–8. 73 Paneth, N, et al. White matter necrosis in very low birth weight infants: neuropathologic and ultrasonographic findings in infants surviving six days or longer. J Pediatr 1990; 116: 975–984. 74 Perlman, JM, Risser R, Broyles RS. Bilateral cystic periventricular leukomalacia in the premature infant: associated risk factors. Pediatrics 1996; 97: 822–827. 75 Takashima S, Iida K, Deguchi K. Periventricular leukomalacia, glial development and myelination. Early Hum Dev 1995; 43: 177–184. 76 Yoon BH, et al. Interleukin-6 concentrations in umbilical cord plasma are elevated in neonates with white matter lesions associated with periventricular leukomalacia. Am J Obstet Gynecol 1996; 174: 1433–1440. 77 Yoon BH, et al. High expression of tumor necrosis factor-alpha and interleukin-6 in periventricular leukomalacia. Am J Obstet Gynecol 1997; 177: 406–411. 78 Zupan V, et al. Periventricular leukomalacia: risk factors revisited. Dev Med Child Neurol 1996; 38: 1061–1067. 79 Hesser U, et al. Diagnosis of intracranial lesions in very-low-birthweight infants by ultrasound: incidence and association with potential risk factors. Acta Paediatr Suppl 1997; 419: 16–26. 80 Holling EE, Leviton A. Characteristics of cranial ultrasound white-matter echolucencies that predict disability: a review. Dev Med Child Neurol 1999; 41: 136–139. 81 Iida K, et al. Neuropathologic study of newborns with prenatal-onset leukomalacia. Pediatr Neurol 1993; 9: 45–48. 82 Dambska M, Laure-Kamionowska M, Schmidt-Sidor B. Early and late neuropathological changes in perinatal white matter damage. J Child Neurol 1989; 4: 291–298. 83 Rodriguez J, et al. Periventricular leukomalacia: ultrasonic and neuropathological correlations. Dev Med Child Neurol 1990; 32: 347–352.

Mechanisms of perinatal brain injury

84 Takashima S, et al. Relationship between periventricular hemorrhage, leukomalacia and brainstem lesions in prematurely born infants. Brain Dev 1989; 11: 121–124. 85 Gilles FH, Leviton A, Dooling EC. The Developing Human Brain: Growth and Epidemiologic Neuropathology. Boston: John Wright, 1983. 86 Rorke LB. Pathology of Perinatal Brain Injury. New York: Raven Press, 1982. 87 Reuck JD. The human periventricular arterial blood supply and the anatomy of cerebral infarctions. Eur Neurol 1971; 5: 321–334. 88 Reuck JD, Chatta AS, Richardson EP. Pathogenesis and evolution of periventricular leukomalacia in infancy. Arch Neurol 1972; 27: 229–236. 89 Reuck JD. Cerebral angioarchitecture and perinatal brain lesions in premature and full term infants. Acta Neurol Scand 1984; 70: 391–395. 90 Rorke LB. Anatomical features of the developing brain implicated in pathogenesis of hypoxic-ischemic injury. Brain Pathol 1992; 2: 211–221. 91 Takashima S, Tanaka K. Development of cerebrovascular architecture and its relationship to periventricular leukomalacia. Arch Neurol 1978; 35: 11–16. 92 Takashima S. Pathology on neonatal hypoxic brain damage and intracranial hemorrhage. Factors important in their pathogenesis. In: International Congress Series No. 579, Child Neurology. Amsterdam: Excerpta Medica, 1982. 93 Altman DI, et al. Cerebral blood flow requirement for brain viability in newborn infants is lower than in adults. Ann Neurol 1988; 24: 218–226. 94 Greisen G. Cerebral blood flow in preterm infants during the first week of life. Acta Paediatr Scand 1986; 75: 43–51. 95 Greisen G, Trojaborg W. Cerebral blood flow, PaCO2 changes, and visual evoked potentials in mechanically ventilated, preterm infants. Acta Paediatr Scand 1987; 76: 394–400. 96 Greisen G, Pryds O. Low CBF, discontinuous EEG activity, and periventricular brain injury in ill, preterm neonates. Brain Dev 1989; 11: 164–168. 97 Pryds O, Greisen G, Friis-Hansen B. Compensatory increase of CBF in preterm infants during hypoglycaemia. Acta Paediatr Scand 1988; 77: 632–637. 98 Pryds O, et al. Heterogeneity of cerebral vasoreactivity in preterm infants supported by mechanical ventilation. J Pediatr 1989; 115: 638–645. 99 Pryds O, Greisen G. Effect of PaCO2 and haemoglobin concentration on day to day variation of CBF in preterm neonates. Acta Paediatr Scand Suppl 1989; 360: 33–36. 100 Pryds O, Greisen G. Preservation of single-flash visual evoked potentials at very low cerebral oxygen delivery in preterm infants. Pediatr Neurol 1990; 6: 151–158. 101 Lou HC, et al. Pressure passive cerebral blood flow and breakdown of the blood-brain barrier in experimental fetal asphyxia. Acta Paediatr Scand 1979; 68: 57–63. 102 Younkin DP, et al. The effect of hematocrit and systolic blood pressure on cerebral blood flow in newborn infants. J Cereb Blood Flow Metab 1987; 7: 295–299.

15

103 Pryds O. Contorl of cerebral circulation in the high-risk neonate. Ann Neurol 1991; 30: 321–329. 104 Wiswell TE, et al. Effects of hypocarbia on the development of cystic periventricular leukomalacia in premature infants treated with high-frequency jet ventilation. Pediatrics 1996; 98: 918–924. 105 Graziani LJ, et al. Clinical antecedents of neurologic and audiologic abnormalities in survivors of neonatal extracorporeal membrane oxygenation. J Child Neurol 1997; 12: 415–422. 106 Fujimoto S, et al. Hypocarbia and cystic periventricular leukomalacia in premature infants. Arch Dis Child 1994; 71: F107–F111. 107 Calvert SA, et al. Periventricular leukomalacia: ultrasonic diagnosis and neurological outcome. Act Paed Scand 1986; 75: 489–496. 108 Ikeda T, et al. Physiologic and histologic changes in near-term fetal lambs exposed to asphyxia by partial umbilical cord occlusion. Am J Obstet Gynecol 1998; 178: 24–32. 109 Reddy K, et al. Maturational change in the cortical response to hypoperfusion injury in the fetal sheep. Pediatr Res 1998; 43: 674–682. 110 Williams CE, et al. Cerebral ischemic injury in the preterm fetal sheep leads ot specific white matter lesions. Pediatric research 1999; (in press). 111 Towfighi J, et al. Neuropathology of remote hypoxicischemic damage in the immature rat. Acta Neuropathol 1991; 81: 578–587. 112 Sheldon RA, Chuai J, Ferriero DM. A rat model for hypoxic-ischemic brain damage in very premature infants. Biol Neonate 1996; 69: 327–341. 113 Back SA, et al. Maturation-dependent vulnerability of oligodendrocytes to death induced by glutathione depletion. Soc Neurosci Abstr 1997. 114 Oka A, et al. Vulnerability of oligodendroglia to glutamate: pharmacology, mechanism and prevention. J Neuroscience 1993; 13: 1441–1453. 115 Yonezawa M, et al. Cystine deprivation induces oligodendroglial death: rescue by free radical scavengers and by a diffuse glial factor. J Neurochem 1996; 67: 566–573. 116 Connor JR, Fine RE. Development of transferrin-positive oligodendrocytes in the rat central nervous system. J Neurosci Res 1987; 17: 51–59. 117 Connor JR, Menzies SL. Altered cellular distribution of iron in thecentral nervous system of myelin deficient rats. Neuroscience 1990; 34: 265–271. 118 Connor JR, Menzies SL. Relationship of iron to oligodendrocytes and myelination. Glia 1996; 17: 83–93. 119 Gilles FH, Leviton A, Kerr CS. Suspectibility of the neonatal feline telencephalic white matter to a lipopolysaccharide. J Neurol Sci 1976; 27: 183–191. 120 Gilles FH, Averill, Jr DR, Kerr CS. Neonatal endotoxin encephalopathy. Ann Neurol 1977; 2: 49–56. 121 Yoon BH, et al. Experimentally induced intrauterine infection causes fetal brain white matter lesions in rabbits. Am J Obstet Gynecol 1997; 177: 797–802. 122 O’Shea TM, et al. Intrauterine infection and the risk of cerebral palsy in very low-birthweight infants. Paediatr Perinat Epidemiol 1998; 12: 72–83.

16

123 Yoon BH, et al. Amniotic fluid inflammatory cytokines (interleukin-6, interleukin-1 beta, and tumor necrosis factor-alpha), neonatal brain white matter lesions, and cerebral palsy. Am J Obstet Gynecol 1997; 177: 19–26. 124 Deguchi K, Mizuguchi M, Takashima S. Immunohistochemical expression of tumor necrosis factor alpha in neonatal leukomalacia. Pediatr Neurol 1996; 14: 13–16. 125 Deguchi K, Oguchi K, Takashima S. Characteristic neuropathology of leukomalacia in extremely low birth weight infants. Pediatr Neurol 1997; 16: 296–300. 126 Meng SZ, et al. Early detection of axonal and neuronal lesions in prenatal-onset periventricular leukomalacia. Brain Dev 1997; 19: 480–484.

T. E. Inder and J. J. Volpe

127 Beneveniste H. The excitotoxin hypothesis in relation to cerebral ischemia. Cerebrovasc Metab Rev 1991; 3: 213–245. 128 Battistin L, Grynbaum A, Lajtha A. The uptake of various amino acids by the mouse brain in vivo. Brain Res 1971; 29: 85–99. 129 Fonnum F. Glutamate: a neurotransmitter in mammalian brain. J Neurochem 1984; 42: 1–11. 130 Yoshioka A, Bacskai B, Pleasure D. Pathophysiology of oligodendroglial excitotoxicity. J Neurosci Res 1996; 46: 427–438.