Management of the Asphyxiated Full Term Infant

612 Archives of Disease in Childhood 1993; 68: 612-616

PERSONAL PRACTICE Arch Dis Child: first published as 10.1136/adc.68.5_Spec_No.612 on 1 May 1993. Downloaded from

Management of the asphyxiated full term infant

Malcolm I Levene

In Britain, approximately one full term baby the effect of making paediatricians very per thousand dies or is severely disabled as the cautious in the use of intravenous glucose result of birth asphyxia. It is arguably the most during neonatal resuscitation. More recent important avoidable cause of permanent data has shown conclusively that there is a neurological injury affecting the mature fundamental difference in the way the fetus/newborn. It is generally agreed that immature and the more mature brain clinical signs of hypoxic-ischaemic encephalo- responds to glucose infusion. This is probably pathy (HIE) are the best markers for a in the main related to the impaired rate of diagnosis of intrapartum 'asphyxia'. Unfor- glucose transport across the immature tunately the severity of HIE can only be blood-brain barrier. The immature brain diagnosed retrospectively after symptoms have appears to be protected by raised glucose con- developed. Early therapeutic intervention in centrations before asphyxial insult compared the asphyxiated baby may be important to with animals that had no additional glucose.3 modify cerebral injury (see below) and there- There is conflicting data concerning the brain fore there remains a need to have early markers protective effect of high glucose concen- of asphyxia such as depressed Apgar scores, trations after asphyxia. In an immature delay in establishing respiration, or evidence of animal model, administration of glucose significant metabolic acidosis on samples of immediately after a period of hypoxic- cord blood. ischaemic insult resulted in significant reduc- Although there appears to have been a fall in tion in cerebral infarction,4 but others have both the incidence of HIE and the number of shown that in a slightly different rat pup children disabled by this condition in recent model there was significant exacerbation of years,' there is little evidence that this has been damage in the presence of hyperglycaemia due to improvement in postnatal management. after injury.5

In this paper I will consider the standard man- Hypoglycaemia must be avoided during and http://adc.bmj.com/ agement of asphyxia, potentially useful new after resuscitation of asphyxiated babies, but methods for treating the asphyxiated brain, faced with conflicting reports on the role of and consider methods of deciding when to glucose infusion after hypoxic-ischaemic injury withdraw care. The brain ofthe full term infant it is not possible at the present time to give responds to asphyxia in a very different manner practical advice concerning the role of glucose to that of a premature baby and this paper only infusion after human birth asphyxia. This

considers the management of asphyxia in full question would be best answered in a double on September 24, 2021 by guest. Protected copyright. term infants. blind controlled clinical study.

Rational basis of standard management PREVENTION OF CEREBRAL OEDEMA Asphyxiated infants require expert and rapid It is widespread practice to anticipate cerebral resuscitation wherever they are born. All health oedema and manage the baby so as to reduce care professionals involved in the birth of the possibility that this complication may babies must be adequately trained and develop. This is done in two ways. Firstly, a retrained in resuscitation techniques. In most cases and mask' is sufficient to maintain 'bag Summary of management ofsevere birth asphyxia. See text ventilation until someone with advanced resus- for details citation skills arrives. Appropriate cardiovascu- lar support must be available for infants born Immediate management: 1. Establish effective ventilation with poor or absent circulation. The table 2. Assist circulation if necessary summarises an approach to the management of Early management: 1. Restrict fluids by 20% for first two days the asphyxiated full term infant. 2. Monitor blood pressure and treat hypotension vigorously 3. Assess respiratory effort and Academic Unit of (a) ventilate if baby breathing spontaneously with arterial Paediatrics and Child carbon dioxide tension >7 kPa GLUCOSE (b) if baby ventilated maintain arterial carbon dioxide Health, University of tension at 4-5 kPa Leeds, D Floor, It has been shown that a raised blood glucose 4. If clinical signs of raised intracranial pressure give Clarendon Wing, The concentration before hypoxic-ischaemic mannitol 1 g/kg over 20 minutes and repeat if necessary General Infirmary at every 4-6 hours Leeds, Leeds LS2 9NS injury in adolescent animals results in more Anticonvulsants if: extensive cerebral injury than in those with 1. Frequent convulsions >3 per hour Correspondence to: 2. Prolonged convulsions lasting .3 minutes Professor Levene. normal or low blood glucose.2 This has had Management of the asphyxiatedfull term infant 613

regimen of fluid restriction is often instituted asphyxiated newborn. Babies who have Arch Dis Child: first published as 10.1136/adc.68.5_Spec_No.612 on 1 May 1993. Downloaded from and secondly corticosteroids are administered. suffered significant birth asphyxia may spontaneously hypoventilate with resulting hyper- capnia and increased cerebral blood volume Fluid restriction which is probably undesirable. For this reason There have been no studies on the effect of all encephalopathic babies should have an fluid restriction in infants with cerebral arterial carbon dioxide tension measurement oedema. In general, measures to reduce and if this is >7 kPa (53 mm Hg) then the cerebral oedema probably have no effect on baby should be electively ventilated. The long term neurological outcome (see below) mechanical ventilator should be adjusted to and it is difficult to argue that routine fluid maintain the arterial carbon dioxide tension at restriction has any advantage in this respect. about 4-5 kPa (34 mm Hg). Fluid restriction may be important in asphyxiated infants who have complications such as inappropriate secretion of antidiuretic Osmotic agents hormone and renal compromise. Fluid reten- Osmotic agents (mannitol or glycerol) are used tion occurring as the result of these two con- to reduce cerebral oedema by increasing serum ditions may further compromise the infant and osmolality. In a neonatal animal model manni- for these reasons I recommend restricting tol significantly reduced brain water content fluids by 20% of the normal regimen for the when given immediately after an hypoxic- first two days of life or until such time as the ischaemic event,9 but it did not reduce the baby's renal function recovers. severity of distribution of brain damage in Hypotension is a relatively common compli- treated compared with untreated animals. cation of birth asphyxia and may be due to There have been no neonatal randomised con- reduction of the circulating blood volume and trolled studies of mannitol or glycerol in the therapeutic dehydration may exacerbate this management of intracranial hypertension. condition. Plasma infusion may be required in Marchal et al in an uncontrolled study gave hypotensive asphyxiated infants. Hypogly- mannitol to 225 babies with the diagnosis of caemia and overt dehydration as a result of asphyxia,1I although the precise indications for fluid restriction must be avoided. treatment were quite varied. Early treatment was defined as mannitol infusion (1 g/kg) before the baby was 2 hours of age. There were Corticosteroids significantly fewer deaths (p=0 005) and the There are no data to support the use of corti- survivors had better neurological outcome costeroids in the routine management of birth (p=0-014) in the early treatment group com- asphyxia. Studies in adults and in animal pared to those treated after 2 hours. Levene models have failed to show any benefit in and Evans showed that mannitol (1 g/kg over reducing brain swelling or improving outcome. 20 minutes) reduced ICP in a small number of

In animal models of asphyxia, corticosteroids severely asphyxiated infants with intracranial http://adc.bmj.com/ have shown either a detrimental effect6 or no hypertension (>1-33 kPa or >10 mm Hg).'1 benefit at all.7 Corticosteroids are associated There was a concomitant rise in cerebral per- with a number of actual or potential side fusion pressure 60 minutes after starting the effects including hyperglycaemia, hyperten- mannitol infusion. The effect of mannitol sion, susceptibility to infection, gastrointestinal lasted for approximately four hours. haemorrhage, and restriction of later brain Providing there is adequate renal function to

growth. In my view, corticosteroids should not allow excretion of mannitol, it appears to be a on September 24, 2021 by guest. Protected copyright. be used in the management of birth asphyxia. relatively safe agent in the management of cerebral oedema. Its efficacy remains in doubt, but I would recommend its use in infants MANAGEMENT OF CEREBRAL OEDEMA with a bulging fontanelle or in whom there In very severely asphyxiated infants, raised is the clinical impression of intracranial intracranial pressure (ICP, a sustained hypertension. increase to 1x33 kPa (>10 mm Hg)) lasting for There are very few studies to evaluate the 20 minutes or more occurs in 70% of cases.8 role of intracranial pressure monitoring and The management of intracranial hypertension management of intracranial hypertension. includes hyperventilation (lowering of the Non-invasive methods for measuring ICP are arterial carbon dioxide tension) and infusion of unreliable in giving accurate measurements of osmotic agents. pressure. Levene et al reported that in a group of 23 babies who had direct subarachnoid pressure recordings, knowledge of the actual ICP Hyperventilation reading might have altered management to the Hyperventilation lowers ICP by cerebral vaso- benefit of the patient in only 9%/o of cases.8 The constriction with reduction of cerebral blood routine use of ICP monitoring cannot be volume and consequent decrease in total recommended. intracranial volume. An important component of the pathophysiology of postasphyxial injury is cerebral hypoperfusion and it is possible that ANTICONVULSANT TREATMENT hyperventilation may exacerbate impaired Convulsions occur commonly after birth reperfusion to the brain. There have been no asphyxia and anticonvulsant drugs are the clinical trials of hyperventilation in the most widely used pharmacological agents in 614 Levene

the management of infants with birth asphyxia. suggest that there is progressive and perma- Arch Dis Child: first published as 10.1136/adc.68.5_Spec_No.612 on 1 May 1993. Downloaded from The immature brain appears to develop con- nent degradation of high energy ATP mole- vulsions at a significantly lower threshold than cules within the brain, but no abnormalities the more mature brain.12 Seizures cause a can be detected on magnetic resonance doubling in cortical metabolic rate and it has spectroscopy for up to 24 hours after the been suggested that this causes further insult.16 Measurement of cerebral blood flow neuronal injury as a result of relative substrate velocity by Doppler ultrasound has shown that depletion. In immature animal models, the an abnormal signal which accurately predicts induction of status epilepticus caused a severe cerebral injury only becomes abnormal marked reduction in eventual brain weight and 24 hours after the asphyxial event.17 These reduced cerebral DNA concentration.13 observations support the evidence from animal Despite these studies, controversy exists as models that asphyxia sets up a cascade ofintra- to whether neonatal seizures simply reflect cellular events which causes the neuron to die neuronal compromise or whether the convul- some hours after the acute insult. The process sions contribute to further neuronal loss. of reoxygenation and reperfusion after There are no human data to support the latter asphyxia appears to be the spark that ignites concept. Nevertheless, the use ofmultiple anti- the fuse to eventual neuronal death. Recent convulsants is very common in the asphyxiated research has suggested a number of pathways neonate and it is clear that exposure ofthe neo- by which this damage occurs and therapeutic natal brain to these drugs may be associated strategies are suggested by this work. Two with adverse effects.14 pathways that offer promise of protecting the There have been few controlled studies on perinatal human brain are (i) that which pre- the use of anticonvulsants in the neonatal vents free radical damage and (ii) that which period. Goldberg et al treated a group of causes the antagonism of excitatory neuro- asphyxiated infants with a continuous infusion transmitters. of thiopentone and compared the outcome with a similar group treated by standard anticonvulsants. 15 There were no significant differ- FREE RADICAL INJURY ences in outcome, but 14 of the 17 infants A free radical is a highly energetic substance given thiopentone treatment required inotrope which contains an uneven number of electrons support for hypertension compared with only in its outer ring. Two free radicals, the super- seven of 15 controls. oxide (02-) and the hydroxyl ('OH), are Although there is no scientific evidence that generated by oxygen metabolism. Their half anticonvulsants improve outcome after neo- lives are very short, but they may under certain natal convulsions due to birth asphyxia, it is circumstances generate chain reactions of free clinically difficult not to treat infants with neo- radicals which cause damage to cellular mem- natal convulsions. In my view, it is not neces- branes. Naturally occurring free radical sary to abolish all convulsions, but treatment scavengers exist to limit the production of

should be instituted for frequent (>3 convul- these substances, but these may be over- http://adc.bmj.com/ sions per hour) or prolonged convulsions (any whelmed after asphyxial compromise. fit lasting ¢ 3 minutes). I recommend pheno- During hypoxic-ischaemic insult, free barbitone as the first line anticonvulsant (20 radicals are produced by the process of mg/kg loading dose followed by 3 mg/kg every degradation of ATP. On reperfusion of the 12 hours). If frequent or prolonged convul- tissues xanthine oxidase metabolises molecular sions continue then a second half loading dose oxygen to produce oxygen free radicals. The

can be given (10 mg/kg). Clonazepam is used main source of xanthine oxidase in the brain is on September 24, 2021 by guest. Protected copyright. as a second line anticonvulsant (100 ,ug/kg the endothelial cell. Free radicals are also loading dose which can be followed by a con- produced in the cortex after asphyxia as the tinuous infusion of 10 pug/kg/hour ifnecessary). result of arachidonic acid metabolism. After an Anticonvulsants can be stopped once the asphyxial injury of the immature animal, infant is thought to be neurologically normal indomethacin has been shown to inhibit the on clinical examination. production of free radicals in the brain.'8 There are no data on the clinical significance Allopurinol is a free radical scavenger and also of electroconvulsive seizure activity in asphyxi- inhibits the enzyme xanthine oxidase. Palmer ated infants with reference to subsequent out- et al have shown that treatment with allopuri- come. Electroencephalographic monitoring of nol before an asphyxial insult reduces both infants may give very important prognostic brain swelling and structural damage in the information (see below), but I do not believe perinatal brain.'9 that it is yet a useful clinical technique for deciding which infants to treat or when to start anticonvulsant treatment. GLUTAMATE RELATED INJURY Glutamate is an excitatory neurotransmitter that is particularly ubiquitous in the developing New treatments brain. The glutamate neuroreceptor is stimu- It is of considerable interest that the neuron is lated by at least three ligands, of which N- remarkably resistant to asphyxial insult and methyl-D-aspartate (NMDA) opens a receptor may recover to generate electrical potentials operated channel which allows calcium to even after an hour of severe hypoxic-ischaemic enter the neuron. Asphyxia causes excessive insult. Evidence from magnetic resonance release of glutamate from the presynaptic spectroscopy studies in asphyxiated neonates vesicles and inhibits uptake of glutamate from Management of the asphyxiatedfull term infant 615

the synaptic cleft. This causes hyperstimula- handicap. Two studies have reported a Arch Dis Child: first published as 10.1136/adc.68.5_Spec_No.612 on 1 May 1993. Downloaded from tion of the glutamate receptors resulting in sensitivity of this finding of 90% and 91% early and late damage to the cell. Abnormal respectively.24 25 It is of interest that prognosis stimulation of the non-NMDA receptors using computed tomography is reliable only in causes entry of excess sodium and water into the second week of life.26 the cell shortly after the asphyxial injury lead- ing to cytotoxic oedema. Neuronal death occurs later and appears to be more closely Prognosis related to excessive stimulation of the NMDA The aggressive early management of the receptors which causes accumulation of toxic severely asphyxiated infant may be tempered concentrations of calcium within the neuron. by consideration of the eventual outcome of This leads to a cascade of biochemical events, the baby and the risk of severe disability. The including activation of intracellular proteases question relating to prognosis may arise at two and lipases with the secondary effect of genera- different times during the course of the infant's tion of oxygen free radicals which in turn cause management. further damage to intracellular membranes. A group of substances antagonise the NMDA channel and protect the perinatal (1) WHEN SHOULD RESUSCITATION STOP? brain after asphyxia.20 One such substance, If an infant has no cardiac output after 10 MK-801, when given after an NMDA insult minutes of effective resuscitation then treat- resulted in 95% protection ofthe brain.21 Even ment should be abandoned. There is no con- when given 120 minutes after the insult there sensus view as to how long resuscitation should was some protection.22 Unfortunately, MK- continue if the baby has a good cardiac output 801 and other NMDA receptor antagonists are yet has failed to breathe spontaneously. There highly toxic. Magnesium appears to be a are reports in the literature of babies not estab- naturally occurring antagonist which has a lishing spontaneous respiration by 20 minutes receptor site deep within the calcium channel. and surviving to be normal. In a Swedish study It has been suggested that increasing the extra- 25% of surviving babies who had not breathed cellular neuronal magnesium concentration spontaneously by 20 minutes were without may protect the brain against hypoxic- significant handicap.27 ischaemic insults by an action similar to The literature and individual experience NMDA receptor antagonists. It has been suggests that the prognosis for normal or near shown that treatment with magnesium normal survival ifthe baby has not breathed for sulphate up to an hour after excessive exposure 30 minutes after birth is poor. Peliowski and to a NMDA-like compound protected the Finer have reviewed the literature and report animal against neurological sequelae.23 the outcome of only 35 full term babies who Magnesium sulphate is a substance that has did not breathe spontaneously by 30 minutes been widely used for over 60 years in perinatal and 24 (80%) died or were significantly handi-

practice as a treatment of premature labour capped.28 Unfortunately, many of the data on http://adc.bmj.com/ and severe pre-eclamptic toxaemia. It seems to time to respiration are anecdotal and policies be well tolerated by the fetus and newborn based on time to achieve spontaneous respira- infant, although transient hypotonia and tion must be recognised as being prone to lethargy are commonly seen for a few days serious prejudice. afterwards. Depression of Apgar scores are also a There are no published data on the use of relatively poor predictor of adverse outcome.

NMDA antagonists in the human neonate, but An overview of three studies reporting mortal- on September 24, 2021 by guest. Protected copyright. the role of magnesium sulphate deserves active ity and morbidity in infants with Apgar scores consideration as a neuroprotective agent for of 3 or less at five minutes showed that this use after birth asphyxia. carried an overall risk of mortality of 16%, but only a 3% risk of handicap in surviving infants.28 Another study reported that the best The role ofbrain imaging predictor of death or handicap was an Apgar There are two indications for imaging the brain score of 5 or less at 10 minutes.29 of asphyxiated newborn infants. The first is to As it is usually inexperienced medical staff discover a treatable complication of the who are called to the delivery suite to resus- asphyxia; most notably a subdural collection. citate asphyxiated babies, I advise that if there In my experience, subdural haemorrhage is any doubt as to whether to continue resusci- severe enough to require surgical treatment tation the baby should be given the benefit of occurs in <5% offull term asphyxiated infants. the doubt and transferred to a neonatal inten- An early ultrasound scan within 12 hours of sive care unit for further management and birth will detect a midline shift if there is a observation. It may be possible to give a more significant lesion. As there appears to be no accurate prediction of outcome later than rational basis for the systematic management immediately after birth (see below). of cerebral oedema, frequent ultrasound or computed tomography to detect this are of no value. (2) WHEN SHOULD MECHANICAL VENTILATION Imaging the brain may be ofvalue in the pre- BE WITHDRAWN? diction of outcome. Extensive areas of Signs of irreversible cerebral injury may be decreased radiodensity have been shown to be delayed for many hours and accurate and a good prognostic indicator of death or severe honest prognostication may have to be delayed 616 Levene

up to 24 hours. The of HIE has been asphyxiated rats treated with dexamethasone. Biol Neonate Arch Dis Child: first published as 10.1136/adc.68.5_Spec_No.612 on 1 May 1993. Downloaded from severity 1973; 22: 388-97. shown to be the most accurate clinical predic- 8 Levene MI, Evans DH, Forde A, Archer LNJ. Value of tor of outcome after birth in full term intracranial pressure monitoring of asphyxiated newborn asphyxia infants. Dev Med Child Neurol 1987; 29: 311-9. infants.29 Unfortunately, the maximal degree 9 Mujsce DJ, Stern DR, Vannucci RC, Towfighi J, Hershey ofHIE not be determined until the PA. Mannitol therapy in perinatal hypoxic-ischemic brain may baby is damage. Ann Neurol 1988; 24: 338. a few days old and it therefore cannot be used 10 Marchal C, Costagliola P, Leveau P, Dulcq P, Steckler R, as an of outcome. number Rouquier F. Traitement de la souffrance cerebrale neo- early predictor A of natale d'origine anoxique par le mannitol. Revue Pediatrie tests have been shown accurately to predict 1974; 9: 581-90. outcome when within 24 hours 11 Levene MI, Evans DH. Medical management of raised performed of intracranial pressure after severe birth asphyxia. Arch Dis life. Child 1985; 60: 12-6. in the brain have 12 Jensen FE, Applegate CD, Holtzman D, Belin TR, Experiments developing Burchfiel JL. Epileptogenic effect of hypoxia in the shown that asphyxia causes a sequence of immature rodent brain. Ann Neurol 1991; 29: 629-37. abnormalities in the electroencephalogram. 13 Wasterlain CG. Effects of neonatal status epilepticus on rat brain development. Neurology 1976; 26: 975-86. This includes an initial period of depressed 14 Levene MI. Neonatal seizures. Neonatal neurology. Current electrical reviews in paediatrics. Edinburgh: Churchill Livingstone, activity lasting approximately eight 1987: 201-38. hours, followed by a period of epileptiform 15 Goldberg R, Moscoso P, Bauer C, et al. Use of barbiturate and in loss of at therapy in severe perinatal asphyxia: a randomized con- activity culminating intensity trolled trial. JPediatr 1986; 109: 851-6. all frequencies by 72 hours.30 The final pattern 16 Wyatt JS, Edwards AD, Azzopardi D, Reynolds EOR. of low correlated with severe neuronal Magnetic resonance and near infrared spectroscopy for activity investigation of perinatal hypoxic-ischaemic brain injury. damage.31 In human neonates, similar severe Arch Dis Child 1989; 64: 953-63. abnormalities on the or 17 Levene MI, Fenton AC,. Evans DH, Archer LNJ, electrocephalogram Shortland DB, Gibson NA. Severe birth asphyxia and cerebral function monitor including burst sup- abnormal cerebral blood-flow velocity. Dev Med Child pression, sustained low voltage, and isoelectric Neurol 1989; 31: 427-34. 18 Pourcyrous M, Leffler CW, Mirro R, Busija DW. Brain activity have also been shown to correlate with superoxide anion generation during asphyxia and reventi- an overall risk of 95% for adverse outcome.28 lation in newborn pigs. Pediatr Res 1990; 28: 618-21. 19 Palmer C, Vannucci RC, Towfighi J. Reduction of perinatal The time scale for these severe abnormalities hypoxic-ischemic brain damage with allopurinol. Pediatr to develop appears to be shorter in the human Res 1990; 27: 332-6. 20 Levene MI. Role of excitatory amino acid antagonists in the than in the animal model. management of birth asphyxia. Biol Neonate 1992; 62: Severely abnormal evoked potentials also 248-51. 21 McDonald JW, Roeser NF, Silverstein FS, Johnston MV. accurately predict adverse outcome, although Quantitative assessment of neuroprotection against of outcome in the first 24 NMDA-induced brain injury. Exp Neurol 1989; 106: prediction hours 289-96. using these techniques is not reliable.32 33 We 22 McDonald JW, Silverstein FS, Cardona D, Hudson C, have shown that abnormally high cerebral Chen R, Johnston MV. Systemic administration of MK- 801 protects against N-methyl-D-aspartate and blood flow velocity detected by duplex quisqualate-mediated neurotoxicity in perinatal rats. ultrasound has a positive predictive Neuroscience 1990; 36: 589-99. Doppler 23 Wolf G, Keilhoff G, Fischer S, Hass P. Subcutaneously value of 94% for adverse outcome (death or applied magnesium protects reliably against quinolinate- severe when in the first induced N-methyl-D-aspartate (NMDA)-mediated neuro- handicap) performed degeneration and convulsions in rats: are there 24 hours of life. 17 therapeutical implications? Neurosci Lett 1990; 117: 207-11. The of abnormal results http://adc.bmj.com/ presence severely 24 Adsett DB, Fitz CR, Hill A. Hypoxic-ischemic cerebral on electroencephalography or Doppler assess- injury in the term newborn: correlation of CT findings ments in 24 with neurological outcome. Dev Med Child Neurol 1985; the first hours appears to predict 27: 155-60. accurately bad outcome and I advise that these 25 Lipper EG, Voorhies TM, Ross G, Vannucci RC, Auld P. tests be twice in the first 24 hours and Early predictors at one-year outcome for infants asphyxi- repeated ated at birth. Dev Med Child Neurol 1986; 28: 303-9. if on both occasions the results are unequivo- 26 Lipp-Zwahlen AE, Deonna T, Chrzanowski R, Micheli JL, abnormal then the should Calame A. Temporal evolution of hypoxic-ischaemic cally poor prognosis brain lesions in asphyxiated full-term newboms assessed

be explained to the parents and consideration by computerized tomography. Neuroradiology 1985; 27: on September 24, 2021 by guest. Protected copyright. to if this 138-44. given withdrawing ventilatory support 27 Ergander U, Eriksson M, Zetterstrom R. Severe neonatal is appropriate. asphyxia: incidence and prediction of outcome in the Stockholm area. Acta Paediatr Scand 1983; 72: 321-5. 28 Peliowski A, Finer NN. Birth asphyxia in the term infant. 1 Hull J, Dodd KL. Falling incidence of hypoxic-ischaemic In: Sinclair JC, Bracken MB, eds. Effective care ofthe new- encephalopathy in term infants. Br J Obstet Gynaecol born infant. Oxford: Oxford University Press, 1992: 1992; 99: 386-91. 249-80. 2 Myers RE, Yamaguchi S. Nervous system effects of cardiac 29 Levene MI, Sands C, Grindulis H, Moore JR. Comparison arrest in monkeys. Preservation of vision. Arch Neurol of two methods of predicting outcome in perinatal 1977; 34: 65-74. asphyxia. Lancet 1986; i: 67-8. 3 Vannucci RC, Mujsce DJ. Effect of glucose on perinatal 30 Williams CE, Gunn AJ, Mallard C, Gluckman PD. hypoxic-ischemic brain damage. Biol Neonate 1992; 62: Outcome after ischemia in the developing sheep brain: an 215-24. electroencephalographic and histological study. Neurology 4 Hattori H, Wasterlain CG. Posthypoxic glucose supplement 1992; 31: 14-21. reduces hypoxic-ischemic brain damage in the neonatal 31 Williams CE, Gunn AJ, Synek B, Gluckman PD. Delayed rat. Ann Neurol 1990; 28: 122-8. seizures occurring with hypoxic-ischemic encephalopathy 5 Sheldon RA, Partridge JC, Ferriero DM. Postischemic in the fetal sheep. Pediatr Res 1990; 27: 561-5. hyperglycemia is not protective to the neonatal rat brain. 32 Gibson NA, Brezinova V, Levene MI. Somatosensory PediatrRes 1992; 32: 489-930. evoked potentials in the term newborn. Electroencephalogr 6 Altman DI, Young RS, Yagel SK. Effects of dexamethasone Clin Neurophysiol 1992; 84: 26-31. in hypoxic ischemic brain injury in the neonatal rat. Biol 33 Taylor MJ, Murphy WJ, Whyte HE. Prognostic reliability of Neonate 1984; 46: 149-56. somatosensory and visual evoked potentials of asphyxi- 7 De Souza SW, Dobbing J. Cerebral oedema in developing ated term infants. Dev Med Child Neurol 1992; 34: brain. III. Brain water and electrolytes in immature 507-15.