DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY
ORIGINAL ARTICLE
Does aetiology of neonatal encephalopathy and hypoxic– ischaemic encephalopathy influence the outcome of treatment? SARAH MCINTYRE 1
| NADIA BADAWI 1,2 | EVE BLAIR 3 | KARIN B NELSON4,5
1 Cerebral Palsy Alliance, University of Notre Dame, Darlinghurst, NSW; 2 Grace Centre for Newborn Care, The Children’s Hospital at Westmead, University of Sydney, Sydney, NSW; 3 Telethon Institute for Child Health Research, University of Western Australia, Perth, WA, Australia. 4 Department of Neurology, Children’s National Medical Centre, Washington, DC; 5 National Institute of Neurological Disorder and Stroke, National Institutes of Health, Bethesda, MD, USA. Correspondence to Karin B Nelson at 5524 Charles Street, Bethesda, MD 20814, USA. Department of Neurology, Children’s National Medical Center 111 Michigan Avenue NW Washington, DC 20010 USA; and the National Institutes of Health 9000 Rockville Pike Bethesda, MD 20892 USA. E-mail: [email protected]
PUBLICATION DATA
Accepted for publication 12th September 2014. Published online ABBREVIATION
HIE
Hypoxic–ischaemic encephalopathy
Neonatal encephalopathy, a clinical syndrome affecting term-born and late preterm newborn infants, increases the risk of perinatal death and long-term neurological morbidity, especially cerebral palsy. With the advent of therapeutic hypothermia, a treatment designed for hypoxic or ischaemic injury, associated mortality and morbidity rates have decreased. Unfortunately, only about one in eight neonates (95% confidence interval) who meet eligibility criteria for therapeutic cooling apparently benefit from the treatment. Studies of infants in representative populations indicate that neonatal encephalopathy is a potential result of a variety of antecedents and that asphyxial complications at birth account for only a small percentage of neonatal encephalopathy. In contrast, clinical case series suggest that a large proportion of neonatal encephalopathy is hypoxic or ischaemic, and trials of therapeutic hypothermia are specifically designed to include only infants exposed to hypoxia or ischaemia. This review addresses the differences, definitional and methodological, between infants studied and investigations undertaken, in population studies compared with cooling trials. It raises the question if there may be subgroups of infants with a clinical diagnosis of hypoxic–ischaemic encephalopathy (HIE) in whom the pathobiology of neonatal neurological depression is not fundamentally hypoxic or ischaemic and, therefore, for whom cooling may not be beneficial. In addition, it suggests approaches to future trials of cooling plus adjuvant therapy that may contribute to further improvement of care for these vulnerable neonates.
Neonatal encephalopathy is a clinical syndrome of disordered neurological function occurring in the first days of life in term-born and late preterm neonates and is characterized by difficulty initiating and maintaining respiration, an abnormal level of consciousness, depression of tone and reflex responses, and often seizures. This symptom complex affects about 3 in every 1000 births, and is an important predictor of perinatal death and a major contributor to long-term adverse neurological outcomes, particularly cerebral palsy (CP).1,2 Evidence from clinical and experimental studies agrees that some instances of neonatal encephalopathy are related to hypoxic–ischaemic injury, but the proportions of neonatal encephalopathy attributed to recent asphyxial events vary with the definitions used and the methodology of the investigations. Studies of infants in representative populations indicate that asphyxial complications at birth account for only a minority of neonatal encephalopathy. These population-based studies also identify other factors that contribute to increased risk of neonatal encephalopathy, such as intrauterine exposure to inflammation or fetal growth restriction, and note that some non-asphyxial factors can produce a clinical picture that closely mimics hypoxia–ischaemia.3–5 2 DOI: 10.1111/dmcn.12725
In contrast, studies of neonatal cooling have approached this topic from the point of view of identifying neonates who might benefit from a treatment designed and proven to ameliorate the effects of acute asphyxial injury; they wish to identify and treat infants with hypoxic–ischaemic encephalopathy (HIE). The clinical literature pertaining to therapeutic cooling claims that acute asphyxial insult is the predominant cause of brain injury in neonates who are cooled (as anticipated if selection were appropriate), with little consideration given to the potential role of antenatal factors. However, throughout the cooling literature, the terms neonatal encephalopathy and HIE are used interchangeably, and inconsistencies in terminology are not limited to the cooling literature. In total population studies, HIE is considered a subtype of neonatal encephalopathy. These differences in terminology give rise to some of the differences in opinion with respect to the proportion of infants that develop neonatal encephalopathy as a result of events in the antepartum and intrapartum periods, and, specifically, as a result of intrapartum asphyxial events. Definitional problems of encephalopathy aside, evidence from 11 randomized clinical trials indicates that therapeutic cooling decreases mortality and long-term neurological © The Authors. Journal compilation © 2015 Mac Keith Press
morbidity rates by 15% (95% confidence interval [CI] 10–20%) and reduces the risk of CP by 12% (95% CI 6–18%), with a corresponding number needed to treat of 8 (95% CI 6–17).6 It is unclear why so many eligible neonates fail to benefit from therapeutic cooling. We examine the possibility that the underlying aetiology of neonatal encephalopathy in infants selected for cooling influences their response to treatment, and that differences in study objectives, design, and definitions contribute to the differing reports of the attribution of aetiology throughout the literature.
SOURCES OF DIFFERENCE BETWEEN CONCLUSIONS DRAWN FROM POPULATION STUDIES AND COOLING TRIALS Subject selection Population-based studies are designed to include all subjects in a given geographical area and time period, and are therefore broadly representative of individuals with a particular condition. Population-based studies are highly convergent, agreeing that less than 25% of neonatal encephalopathy is HIE (i.e. arises from birth asphyxia). In contrast, cooling studies from therapeutic referral centres, whose subject samples may differ from centre to centre, give estimates of the proportions of subjects with potentially asphyxial birth events, ranging from 40% to 60%.7–9 Since these trials were published, reports of cooling for the treatment of HIE, outside a clinical trial setting, indicate that between 22%10 and 51%11 of those treated experienced a potentially asphyxial birth event, suggesting substantial differences between different referral centre groups. These findings also suggest that, in the case of some infants being treated with therapeutic hypothermia, the timing and nature of injury are not obviously hypoxic–ischaemic. Study focus The purpose of epidemiological studies is to identify antecedents that may cause or predispose to, or be on a causal path to, neonatal encephalopathy or its subtype, HIE. The purpose of cooling studies is to identify neonates whose neurological illness results from intrapartum asphyxial injury (regardless of antecedents), which is the group of infants that therapeutic cooling was developed to treat, and to test cooling efficacy. These different goals lead to differences in exclusion and inclusion criteria. However, it is uncertain whether or not the inclusion and exclusion criteria of the cooling trials are suitable for specifically identifying asphyxial aetiologies, as discussed below. Cooling studies exclude infants with major birth defects identifiable in the first hours of life, and many exclude infants with severe fetal growth restriction (often defined by birthweight below an arbitrary value). A minority of cooling studies have excluded infants with inflammation, but have defined inflammation by highly insensitive measures, such as documented sepsis. Epidemiological studies
have shown that both later-discovered birth defects and fetal growth restriction are considerably more common in infants with neonatal encephalopathy than in neurologically asymptomatic neonates, and more common than potentially asphyxial events and markers of inflammation combined.12 Indeed, the chief evidence for the antepartum origin of neonatal encephalopathy and CP in many infants is the increased frequency of aberrant fetal growth and birth defects associated with these adverse neurological outcomes. Thus, the populations of infants included in population-based studies differ in important ways from those in cooling studies, the latter seldom being fully characterized in published reports. Furthermore, their analytical approaches also differ. A recent paper investigating antepartum and intrapartum factors associated with HIE13 excluded severe congenital malformations. Low birthweight for dates was included as an antepartum factor in multivariate analysis, along with hypertension and demographic characteristics. Intrapartum factors included abnormal electronic fetal monitoring, sentinel events, and thick meconium. This multivariate analysis indicated that antepartum factors were not significantly related to HIE when intrapartum factors were included in the model. Might this be misleading? The use of multivariate analysis, with entry depending on statistical significance, means that, if there are a variety of possible predisposing factors, and a small range of intrapartum events to which they predispose (especially when all sentinel events are combined into one variable), then there is greater statistical power to identify a ‘final common pathway’. However, the identity of specific predisposing factors will be lost, and only the intrapartum characteristics, which may result from a number of specific aetiological factors, will be identified as the sole causal characteristics. For example, abnormal cardiotocography may result from placental dysfunction caused by inflammation, autoimmune status, thrombotic tendency, etc. Inclusion of abnormal cardiotocography – the non-specific result of a number of prior causes – gives the impression of a single specific cause, and information will be lost on antecedents that contribute to cardiotocography abnormality and, consequently, information that may suggest early preventative strategies, will also be lost.
Criteria for birth asphyxia Differences in definitions of and criteria for birth asphyxia are important and often-acknowledged problems in perinatal medicine. Definitional issues contribute more than differences in study design to the variation in estimates of the proportion of infants with CP attributable to birth asphyxia.14 As already mentioned, some of these difficulties relate to the non-specific nature of many of the characteristics employed in the cooling literature to identify hypoxia–ischaemia, such as meconium in the amniotic fluid, abnormal fetal heart rate patterns, low Apgar scores, and neonatal seizures, all of which are characteristics that Does Aetiology of HIE Influence Outcome? Sarah McIntyre et al. 3
can result from a variety of causes often related to underlying abnormalities identifiable in the placenta. Moreover, diagnostic tools, such as those used to measure cord blood gas levels, are markedly underutilized in clinical practice. Hypoxia and ischaemia are possible endpoints of several different pathologies, and the effects of hypoxia and ischaemia are greatly dependent on the condition of the fetus when it meets an insult in the peripartum period. ‘If the role of birth asphyxia as an initiating factor . . . is to be correctly assessed, a surrogate must be used that is relatively specific to birth asphyxia and not itself an early symptom of the developing disorder’.14 Pin et al.,15 when reviewing the cooling literature, concluded that ‘researchers are using very loose diagnostic criteria of perinatal asphyxia and post-asphyxial neonatal encephalopathy, making the study samples heterogeneous’ and making it difficult to compare studies with one another. Rather than attempt to define birth asphyxia, some recent population studies have identified complications, such as major abruption of the placenta or uterine rupture, that are potentially capable of interrupting oxygen supply or blood flow to the fetal brain (overt ‘sentinel events’), and examined the frequency of their occurrence in children with adverse neurological outcomes compared with typically developing children. Such sentinel events are more frequent in neonates with a clinical diagnosis of HIE (with reports ranging from 22–60%) than in those without HIE, but still do not account for a large proportion of infants with HIE. In addition to hypoxia–ischaemia, recent evidence indicates that possible antecedents include inflammation (related to infection or immune processes), coagulation defects, maternal thyroid disorders, and other factors; however, much remains unknown.
How antecedents are identified There are important differences in how strenuously antecedents of neonatal encephalopathy or HIE are sought. Most population-based studies derive information on mother, family, pregnancy, and birth chiefly from medical records generated for clinical care. Such studies sometimes include statutorily collected data from registers of birth and death, or data from specialized registers, such as those for CP or birth defects. For uncommon outcomes, prospective studies are seldom performed because of their prohibitive cost. In retrospective use of records, investigators have no control over what information was collected or how it was recorded, and cannot impose study definitions, but must depend upon diagnostic impressions recorded by clinicians caring for patients. It is seldom possible to link these data to data from other sources, to check data accuracy (which is, for example, important for outliers), or to extend the range of variables. Despite these limitations, one of the major strengths of many population studies is the strenuous effort made to collect information relevant to antecedents prior to, or blinded to, the long-term outcomes. In many population studies, abstracters of medical records are specifically 4 HIE supplement 2015, 57 (Suppl. 3): 2–7
trained and random samples of records are checked for accuracy and completeness of data extraction. Statutorily maintained registers of births can provide an excellent sampling frame for selection of controls for total population research. The amount of detail available from record review is usually directly related to the effort spent in obtaining it, though quality of information is, of course, related to the care with which the records were created. In contrast, none of the cooling studies have described methods for the systematic and rigorous examination of medical records to identify antenatal factors, or what studies were undertaken to rule out aetiological factors other than hypoxia–ischaemia. Rigorous aetiological diagnosis is seldom complete before entry into cooling trials, which usually occurs within 6 hours of birth, since the necessary review of maternal and birth records, family history, placental examination, neonatal examination for internal birth defects, and possibly metabolic or genetic testing, cannot be completed within this time frame. Many infants who meet criteria for neonatal encephalopathy are found, on closer examination, to possess more than one associated risk factor.16 Many disorders that can cause neonatal encephalopathy, malformations of internal organs, or metabolic or other genetic disorders, commonly evade diagnosis in the neonatal period. None of these disorders would be expected to respond markedly to therapeutic hypothermia, so these disorders probably contribute to the observed majority of eligible neonates who do not benefit from cooling. Except for a few that include postnatal neuroimaging, no study has yet re-analysed the outcomes of therapeutic cooling, stratified by aetiologically pertinent information available post randomization; therefore, evidence supporting the presence or absence of significant antenatal factors, and the impact of those on outcome, is lacking. In a recent retrospective referral-based study whose specific objective was to determine the proportion of neonatal HIE caused by antepartum factors, alone or in combination with intrapartum factors, data on important antenatal variables (such as growth restriction, thyroid disorders, hypertension, and infections throughout pregnancy) were missing for 12% to 20% of infants with HIE, compared with only 3% to 4% of healthy infants.13 Does aetiology alter outcome? Given the lack of antepartum evidence, the literature on therapeutic hypothermia does not allow consideration of whether or not antecedents in selected infants alters their response to treatment, yet underlying aetiologies commonly do influence outcome: smallpox and chickenpox have different prognoses despite similarities at presentation, and different causes of prematurity are associated with differing outcomes of birth at the same gestational age.17,18 It has been observed that, among term-born infants selected for cooling, those with inflammatory placental histology19 or reduced fetal movements20 respond less well to therapeutic hypothermia. It seems likely that, as with other disease-associated outcomes, the antecedents of neonatal neurological abnormality will influence both long-term outcomes and response to therapies.
SOURCES OF INFORMATION The aetiological investigation of neonatal encephalopathy should include a full review of medical records of mother, pregnancy, birth, family history, and medical workup in the neonatal period and beyond. Other sources of information likely to play a more important role in the future are placental histological examination, multiplexed microassays, and genetic or genomic evaluation. The placenta In population-based studies, macroscopic examination of the placenta has identified that placental abnormalities, such as infarction, are more common in infants with neonatal encephalopathy or HIE than in healthy neonates.19,21 However, many important features of placental pathology cannot be identified by macroscopic examination. Abnormal placental histology has been reported in more than 75% of infants with neonatal encephalopathy (whether or not it is considered to be HIE) in clinical samples,11,22 with the most common placental lesions being vasculopathy, inflammation, or both. Placental histological examination is the definitive method for establishing the presence of inflammation, its locus, severity, and acuteness or chronicity. Importantly, histological examination is the only available means of differentiating two major subcategories of placental inflammation: chorioamnionitis or funisitis caused by infection, and chronic villitis, which is most commonly an immunological response related to the breakdown of maternal–fetal tolerance.23 The prevalence of chronic villitis increases with gestational age, and is associated with a marked increase in the risk of neonatal encephalopathy.24,25 Chronic villitis may coexist with, or be difficult to distinguish from, fetal thrombotic vasculopathy of the placenta,26 which is also associated with the risk of adverse neurological outcomes,24,25 and with stillbirth and cardiac malformations.26 Risk factors for both chronic villitis and adverse neurological outcomes include prior pregnancy loss, fetal growth restriction, abnormal fetal heart rate patterns, need for emergency surgical delivery, severe acidosis, and neuroimaging evidence of basal ganglia–thalamic injury.27 Some of these characteristics are sometimes assumed to be indicators of birth asphyxia but, since they are not specific to birth asphyxia, the possibility of misattribution of aetiology is high without a placental histological examination carried out by an experienced perinatal pathologist. Correct identification of chronic villitis is also important for its treatment implications: agents under consideration are steroids and intravenous immunoglobulins. Therapeutic hypothermia is not known to be an effective therapy in infants with placental villitis. Another important condition for which the placenta provides key evidence is fetal growth restriction, present in a significantly higher proportion of neonates with neonatal encephalopathy than those without, as indicated by population studies. Fetal growth restriction is also associated with
a range of other adverse neurological outcomes including stillbirth, epilepsy, intellectual disability, autism, and schizophrenia. Fetal growth restriction is itself a condition with many aetiologies, including aberrant expression of a variety of growth factors, genetic variants, and co-occurring birth defects. Fetal growth restriction contributes more to neonatal encephalopathy than sentinel events, and much more than sentinel events to total CP.12 Placental lesions indicative of disorders of maternal and fetal blood flow, and chronic villitis, are associated with fetal growth restriction. In future, studies of placental morphology and function are likely to become increasingly important. Such data will need to be linked with information on mother, pregnancy, birth, neonatal period, and outcome in longitudinal studies. Despite limitations in the quantity and quality of evidence, current reports indicate that placental abnormality is often associated with neonatal encephalopathy, perinatal stroke,28 low Apgar scores,29–31 and term stillbirth.32,33 However, the nature of the placental abnormality differs in several ways, supporting the aetiological diversity of these clinical syndromes. As there appears to be an important role for the placenta in the illnesses underlying the majority of neonatal encephalopathy, current animal models of asphyxial HIE may be inadequate for studying neonatal encephalopathy in human infants.
Multiplexed micro-assays In addition to inflammation and hypoxia–ischaemia, risk factors for neonatal encephalopathy include coagulation disorders, such as acquired thrombophilias, autoimmune conditions of the mother or fetus, and specific infections. Aetiological diagnosis is a matter of urgent concern for immediate medical care, but has been constrained by the small volumes of blood that can be drawn from a newborn infant. Technology is now available for miniaturized multiplexed assays that can provide answers to many questions from a tiny sample of neonatal blood or cerebrospinal fluid, and which could be assembled in one or more panels for immediate investigation of the neurologically symptomatic neonate. Among analytes that might be useful are markers of severity of neonatal illness and predictors of outcome, such as S100b, neuron-specific enolase, and ubiquitin; inflammatory indicators, including C-reactive protein, cytokines and chemokines; and immune factors including complement fractions and systematic lupus markers, coagulation indicators, hormones, and identifiers of specific infective organisms. It is hoped that clinicians and medical industries can cooperate to bring such diagnostic capabilities into practical use in the near future. Genetic and genomic studies Recent investigations of CP reveal copy number variants and mutations with possible relevance to propensity for inflammation, coagulation abnormalities, vulnerability to hypoxia and, at least in the case of preterm infants, to thyroid disorders.34 Longitudinal studies indicate that hypoxia Does Aetiology of HIE Influence Outcome? Sarah McIntyre et al. 5
and inflammation are related to CP via a pathway that includes neonatal encephalopathy,12,35 and maternal thyroid disorders are also associated with risk of neonatal encephalopathy.19,36 Vulnerability to neonatal encephalopathy is likely, in future, to be a fruitful subject for genetic and genomic investigation.
WHY IS KNOWLEDGE OF THE AETIOLOGY OF NEONATAL ENCEPHALOPATHY IMPORTANT? To improve understanding of the pathobiology underlying neonatal encephalopathy, and to develop more selective strategies for prevention and treatment, it is important that we understand the aetiology of neonatal encephalopathy. Knowledge of aetiology may suggest therapies, in addition or as an alternative to hypothermia. There may be subgroups of infants with neonatal encephalopathy, with certain aetiologies (anticipated to be those with HIE, perhaps best identified as those with sentinel events the timing of which is known), who are likely to respond relatively well to cooling, while those with different aetiologies may respond less well, or may even be harmed by this treatment. WHAT NEEDS TO BE DONE? A minimal dataset should be established, to be gathered post randomization in future neonatal neuroprotective trials, seeking to maximize information of possible aetiological relevance without excessively burdening the trials. Secondary analyses of cooling trials, with stratification that incorporates information gathered after randomization, should be undertaken. Significant collaboration will be required to obtain sample sizes sufficient to support such
stratification, along with adequate funding and careful design. As criteria for cooling expand, the pool of subjects for neonatal treatment is likely to become even more heterogeneous, and the relationships among antecedents, therapies, and outcomes even less clear. Therefore, it is hoped that the next generation of trials for neonatal neuroprotection will identify major aetiological subgroups in their samples, and provide a basis for secondary analysis of study results stratified by those subgroups, in order to maximize the interpretability of their results. Investigators involved in population-based studies, and those in neonatal care units providing therapeutic hypothermia to infants with neonatal encephalopathy, have different experiences and assumptions, but knowledge from both settings is needed for further improvements to the care of these infants. How much of the outcome of therapeutic hypothermia is related to the treatment, and how much to the pathologies that led to the treatment? Are there some encephalopathic newborn infants who need additional or alternative therapy? To date, the literature of therapeutic cooling does not fully address these questions, but it is important that future trials attempt to do so. A CK N O W L E D G E M E N T S Funding sources: the Castang Foundation supported SM to travel to the meeting; the Cerebral Palsy Research Foundation supported SM; Macquarie Group Foundation and Cerebral Palsy Research Foundation supported NB; NHMRC program grant No 353514 supported EB. The authors have stated that they had no interests that might be perceived as posing a conflict or bias.
REFERENCES 1. Leviton A, Nelson KB. Problems with definitions and
8. Jacobs SE, Morley CJ, Inder TE, et al. Whole-body
14. Ellenberg JH, Nelson KB. The association of cerebral
classifications of newborn encephalopathy. Pediatr Neurol
hypothermia for term and near-term newborns with
palsy with birth asphyxia: a definitional quagmire. Dev
1992; 8: 85–90.
hypoxic-ischemic encephalopathy: a randomized con-
2. Kurinczuk JJ, White-Koning M, Badawi N. Epidemiology of neonatal encephalopathy and hypoxic-ischaemic encephalopathy. Early Hum Dev 2010; 86: 329–38.
trolled trial. Arch Pediatr Adolesc Med 2011; 165: 692– 700. 9. Simbruner G, Mittal RA, Rohlmann F, Muche R. Sys-
3. Keogh JM, Badawi N, Kurinczuk JJ, Pemberton PJ,
temic hypothermia after neonatal encephalopathy: out-
Stanley FJ. Group B streptococcus infection, not birth
comes of neo.nEURO.network RCT. Pediatrics 2010;
asphyxia. Aust N Z J Obstet Gynaecol 1999; 39: 108–
126: e771–8.
10.
10. Hayes BC, McGarvey C, Mulvany S, et al. A case–con-
Med Child Neurol 2013; 55: 210–6. 15. Pin TW, Eldridge B, Galea MP. A review of developmental outcomes of term infants with post-asphyxia neonatal encephalopathy. Eur J Paediatr Neurol 2009; 13: 224–34. 16. Nelson KB, Grether JK. Selection of neonates for neuroprotective therapies: one set of criteria applied to a population. Arch Pediatr Adolesc Med 1999; 153: 393–8.
4. Grether JK, Nelson KB. Maternal infection and cerebral
trol study of hypoxic-ischemic encephalopathy in new-
17. Basso O, Wilcox A. Mortality risk among preterm
palsy in infants of normal birth weight. J Am Med Assoc
born infants at >36 weeks gestation. Am J Obstet Gynecol
babies: immaturity versus underlying pathology. Epidemi-
1997; 278: 207–11.
2013; 209: 29.e1–29.e19.
ology 2010; 21: 521–7.
11. Chang T, Reyes C, Teng J, Placette J, Massaro AN,
18. Gagliardi L, Rusconi F, Da Fre M, et al. Pregnancy dis-
for hypoxic-ischemic encephalopathy. Ann Neurol 2012;
Nelson KB. Neonatal encephalopathy, sentinel events
orders leading to very preterm birth influence neonatal
72: 156–66.
and the placenta. J Neonatal Perinatal Med 2012; 5: 1–8.
outcomes: results of the population-based ACTION
5. Volpe JJ. Neonatal encephalopathy: an inadequate term
cohort study. Pediatr Res 2013; 73: 794–801.
6. Jacobs SE, Berg M, Hunt R, Tarnow-Mordi WO, Inder
12. McIntyre S, Blair E, Badawi N, Keogh J, Nelson KB.
TE, Davis PG. Cooling for newborns with hypoxic is-
Antecedents of cerebral palsy and perinatal death in
19. Badawi N, Kurinczuk JJ, Keogh JM, et al. Antepartum
chaemic encephalopathy. Cochrane Database Syst Rev
term and late preterm singletons. Obstet Gynecol 2013;
risk factors for newborn encephalopathy: the Western
2013; 1: CD003311.
122: 869–77.
Australian case–control study. Br Med J 1998; 317:
7. Shankaran S, Laptook AR, Ehrenkranz RA, et al.
13. Martinez-Biarge M, Diez-Sebastian J, Wusthoff CJ,
Whole-body hypothermia for neonates with hypoxic-
Mercuri E, Cowan FM. Antepartum and intrapartum
20. Bonifacio SL, Glass HC, Vanderpluym J, et al. Perinatal
ischemic encephalopathy. N Engl J Med 2005; 353:
factors preceding neonatal hypoxic-ischemic encephalop-
events and early magnetic resonance imaging in thera-
1574–84.
athy. Pediatrics 2013; 132: e952–9.
peutic hypothermia. J Pediatr 2011; 158: 360–5.
6 HIE supplement 2015, 57 (Suppl. 3): 2–7
1549–53.
21. Blair E, de Groot J, Nelson KB. Placental infarction
27. Harteman JC, Nikkels PG, Benders MJ, Kwee A, Gro-
identified by macroscopic examination and risk of cere-
enendaal F, de Vries LS. Placental pathology in full-
bral palsy in infants at 35 weeks of gestational age and
term infants with hypoxic-ischemic neonatal encephalop-
33. Flenady V, Koopmans L, Middleton P, et al. Major risk
over. Am J Obstet Gynecol 2011; 205: e1–7.
athy and association with magnetic resonance imaging
factors for stillbirth in high-income countries: a system-
22. Wintermark P, Boyd T, Gregas MC, Labrecque M,
pattern of brain injury. J Pediatr 2013; 163: 968–95.
Hansen A. Placental pathology in asphyxiated newborns
28. Elbers J, Viero S, MacGregor D, DeVeber G, Moore
meeting the criteria for therapeutic hypothermia. Am J
AM. Placental pathology in neonatal stroke. Pediatrics
Obstet Gynecol 2010; 203: 579.e1–9.
2011; 127: e722–9.
ropathology and placental pathology. Pediatr Dev Pathol 2011; 14: 345–52.
atic review and meta-analysis. Lancet 2011; 377: 1331– 40. 34. Ho NT, Furge K, Fu W, et al. Gene expression in archived newborn blood spots distinguishes infants who
23. Tamblyn JA, Lissauer DM, Powell R, Cox P, Kilby
29. Naeye RL. Underlying disorders responsible for the
MD. The immunological basis of villitis of unknown ati-
neonatal deaths associated with low Apgar scores. Biol
ology – review. Placenta 2013; 34: 846–55.
Neonate 1979; 35: 150–5.
will later develop cerebral palsy from matched controls. Pediatr Res 2013; 73: 450–6. 35. Ellenberg JH, Nelson KB. Cluster of perinatal events
24. McDonald DGM, Kelehan P, McMenamin JB, et al.
30. Locatelli A, Incerti M, Ghidini A, Greco M, Villa E,
Placental fetal thrombotic vasculopathy is associated
Paterlini G. Factors associated with umbilical artery
with neonatal encephalopathy. Hum Pathol 2004; 35:
acidemia in term infants with low Apgar scores at
36. Ellis M, Manandhar N, Shrestha PS, Shrestha L, Man-
5 min. Eur J Obstet Gynecol Reprod Biol 2008; 139: 146–
andhar DS, Costello AM. Outcome at 1 year of neonatal
50.
encephalopathy in Kathmandu, Nepal. Dev Med Child
875–80. 25. Hayes BC, Cooley S, Donnelly J, et al. The placenta in infants >36 weeks gestation with neonatal encephalopa-
31. Wong L, MacLennan AH. Gathering the evidence: cord
thy: a case control study. Arch Dis Child Fetal Neonatal
gases and placental histology for births with low Apgar
Ed 2013; 98: F233–9. 26. Faye-Petersen OM. The placenta in preterm birth. J Clin Pathol 2008; 61: 1261–75.
identifying infants at high risk for death or disability. J Pediatr 1988; 113: 546–52.
Neurol 1999; 41: 689–95.
scores. Aust N Z J Obstet Gynaecol 2011; 51: 17–21. 32. Chang KT, Keating S, Costa S, Machin G, Kingdom J, Shannon P. Third-trimester stillbirths: correlative neu-
Does Aetiology of HIE Influence Outcome? Sarah McIntyre et al. 7
ORIGINAL ARTICLE
Does aetiology of neonatal encephalopathy and hypoxic– ischaemic encephalopathy influence the outcome of treatment? SARAH MCINTYRE 1
| NADIA BADAWI 1,2 | EVE BLAIR 3 | KARIN B NELSON4,5
1 Cerebral Palsy Alliance, University of Notre Dame, Darlinghurst, NSW; 2 Grace Centre for Newborn Care, The Children’s Hospital at Westmead, University of Sydney, Sydney, NSW; 3 Telethon Institute for Child Health Research, University of Western Australia, Perth, WA, Australia. 4 Department of Neurology, Children’s National Medical Centre, Washington, DC; 5 National Institute of Neurological Disorder and Stroke, National Institutes of Health, Bethesda, MD, USA. Correspondence to Karin B Nelson at 5524 Charles Street, Bethesda, MD 20814, USA. Department of Neurology, Children’s National Medical Center 111 Michigan Avenue NW Washington, DC 20010 USA; and the National Institutes of Health 9000 Rockville Pike Bethesda, MD 20892 USA. E-mail: [email protected]
PUBLICATION DATA
Accepted for publication 12th September 2014. Published online ABBREVIATION
HIE
Hypoxic–ischaemic encephalopathy
Neonatal encephalopathy, a clinical syndrome affecting term-born and late preterm newborn infants, increases the risk of perinatal death and long-term neurological morbidity, especially cerebral palsy. With the advent of therapeutic hypothermia, a treatment designed for hypoxic or ischaemic injury, associated mortality and morbidity rates have decreased. Unfortunately, only about one in eight neonates (95% confidence interval) who meet eligibility criteria for therapeutic cooling apparently benefit from the treatment. Studies of infants in representative populations indicate that neonatal encephalopathy is a potential result of a variety of antecedents and that asphyxial complications at birth account for only a small percentage of neonatal encephalopathy. In contrast, clinical case series suggest that a large proportion of neonatal encephalopathy is hypoxic or ischaemic, and trials of therapeutic hypothermia are specifically designed to include only infants exposed to hypoxia or ischaemia. This review addresses the differences, definitional and methodological, between infants studied and investigations undertaken, in population studies compared with cooling trials. It raises the question if there may be subgroups of infants with a clinical diagnosis of hypoxic–ischaemic encephalopathy (HIE) in whom the pathobiology of neonatal neurological depression is not fundamentally hypoxic or ischaemic and, therefore, for whom cooling may not be beneficial. In addition, it suggests approaches to future trials of cooling plus adjuvant therapy that may contribute to further improvement of care for these vulnerable neonates.
Neonatal encephalopathy is a clinical syndrome of disordered neurological function occurring in the first days of life in term-born and late preterm neonates and is characterized by difficulty initiating and maintaining respiration, an abnormal level of consciousness, depression of tone and reflex responses, and often seizures. This symptom complex affects about 3 in every 1000 births, and is an important predictor of perinatal death and a major contributor to long-term adverse neurological outcomes, particularly cerebral palsy (CP).1,2 Evidence from clinical and experimental studies agrees that some instances of neonatal encephalopathy are related to hypoxic–ischaemic injury, but the proportions of neonatal encephalopathy attributed to recent asphyxial events vary with the definitions used and the methodology of the investigations. Studies of infants in representative populations indicate that asphyxial complications at birth account for only a minority of neonatal encephalopathy. These population-based studies also identify other factors that contribute to increased risk of neonatal encephalopathy, such as intrauterine exposure to inflammation or fetal growth restriction, and note that some non-asphyxial factors can produce a clinical picture that closely mimics hypoxia–ischaemia.3–5 2 DOI: 10.1111/dmcn.12725
In contrast, studies of neonatal cooling have approached this topic from the point of view of identifying neonates who might benefit from a treatment designed and proven to ameliorate the effects of acute asphyxial injury; they wish to identify and treat infants with hypoxic–ischaemic encephalopathy (HIE). The clinical literature pertaining to therapeutic cooling claims that acute asphyxial insult is the predominant cause of brain injury in neonates who are cooled (as anticipated if selection were appropriate), with little consideration given to the potential role of antenatal factors. However, throughout the cooling literature, the terms neonatal encephalopathy and HIE are used interchangeably, and inconsistencies in terminology are not limited to the cooling literature. In total population studies, HIE is considered a subtype of neonatal encephalopathy. These differences in terminology give rise to some of the differences in opinion with respect to the proportion of infants that develop neonatal encephalopathy as a result of events in the antepartum and intrapartum periods, and, specifically, as a result of intrapartum asphyxial events. Definitional problems of encephalopathy aside, evidence from 11 randomized clinical trials indicates that therapeutic cooling decreases mortality and long-term neurological © The Authors. Journal compilation © 2015 Mac Keith Press
morbidity rates by 15% (95% confidence interval [CI] 10–20%) and reduces the risk of CP by 12% (95% CI 6–18%), with a corresponding number needed to treat of 8 (95% CI 6–17).6 It is unclear why so many eligible neonates fail to benefit from therapeutic cooling. We examine the possibility that the underlying aetiology of neonatal encephalopathy in infants selected for cooling influences their response to treatment, and that differences in study objectives, design, and definitions contribute to the differing reports of the attribution of aetiology throughout the literature.
SOURCES OF DIFFERENCE BETWEEN CONCLUSIONS DRAWN FROM POPULATION STUDIES AND COOLING TRIALS Subject selection Population-based studies are designed to include all subjects in a given geographical area and time period, and are therefore broadly representative of individuals with a particular condition. Population-based studies are highly convergent, agreeing that less than 25% of neonatal encephalopathy is HIE (i.e. arises from birth asphyxia). In contrast, cooling studies from therapeutic referral centres, whose subject samples may differ from centre to centre, give estimates of the proportions of subjects with potentially asphyxial birth events, ranging from 40% to 60%.7–9 Since these trials were published, reports of cooling for the treatment of HIE, outside a clinical trial setting, indicate that between 22%10 and 51%11 of those treated experienced a potentially asphyxial birth event, suggesting substantial differences between different referral centre groups. These findings also suggest that, in the case of some infants being treated with therapeutic hypothermia, the timing and nature of injury are not obviously hypoxic–ischaemic. Study focus The purpose of epidemiological studies is to identify antecedents that may cause or predispose to, or be on a causal path to, neonatal encephalopathy or its subtype, HIE. The purpose of cooling studies is to identify neonates whose neurological illness results from intrapartum asphyxial injury (regardless of antecedents), which is the group of infants that therapeutic cooling was developed to treat, and to test cooling efficacy. These different goals lead to differences in exclusion and inclusion criteria. However, it is uncertain whether or not the inclusion and exclusion criteria of the cooling trials are suitable for specifically identifying asphyxial aetiologies, as discussed below. Cooling studies exclude infants with major birth defects identifiable in the first hours of life, and many exclude infants with severe fetal growth restriction (often defined by birthweight below an arbitrary value). A minority of cooling studies have excluded infants with inflammation, but have defined inflammation by highly insensitive measures, such as documented sepsis. Epidemiological studies
have shown that both later-discovered birth defects and fetal growth restriction are considerably more common in infants with neonatal encephalopathy than in neurologically asymptomatic neonates, and more common than potentially asphyxial events and markers of inflammation combined.12 Indeed, the chief evidence for the antepartum origin of neonatal encephalopathy and CP in many infants is the increased frequency of aberrant fetal growth and birth defects associated with these adverse neurological outcomes. Thus, the populations of infants included in population-based studies differ in important ways from those in cooling studies, the latter seldom being fully characterized in published reports. Furthermore, their analytical approaches also differ. A recent paper investigating antepartum and intrapartum factors associated with HIE13 excluded severe congenital malformations. Low birthweight for dates was included as an antepartum factor in multivariate analysis, along with hypertension and demographic characteristics. Intrapartum factors included abnormal electronic fetal monitoring, sentinel events, and thick meconium. This multivariate analysis indicated that antepartum factors were not significantly related to HIE when intrapartum factors were included in the model. Might this be misleading? The use of multivariate analysis, with entry depending on statistical significance, means that, if there are a variety of possible predisposing factors, and a small range of intrapartum events to which they predispose (especially when all sentinel events are combined into one variable), then there is greater statistical power to identify a ‘final common pathway’. However, the identity of specific predisposing factors will be lost, and only the intrapartum characteristics, which may result from a number of specific aetiological factors, will be identified as the sole causal characteristics. For example, abnormal cardiotocography may result from placental dysfunction caused by inflammation, autoimmune status, thrombotic tendency, etc. Inclusion of abnormal cardiotocography – the non-specific result of a number of prior causes – gives the impression of a single specific cause, and information will be lost on antecedents that contribute to cardiotocography abnormality and, consequently, information that may suggest early preventative strategies, will also be lost.
Criteria for birth asphyxia Differences in definitions of and criteria for birth asphyxia are important and often-acknowledged problems in perinatal medicine. Definitional issues contribute more than differences in study design to the variation in estimates of the proportion of infants with CP attributable to birth asphyxia.14 As already mentioned, some of these difficulties relate to the non-specific nature of many of the characteristics employed in the cooling literature to identify hypoxia–ischaemia, such as meconium in the amniotic fluid, abnormal fetal heart rate patterns, low Apgar scores, and neonatal seizures, all of which are characteristics that Does Aetiology of HIE Influence Outcome? Sarah McIntyre et al. 3
can result from a variety of causes often related to underlying abnormalities identifiable in the placenta. Moreover, diagnostic tools, such as those used to measure cord blood gas levels, are markedly underutilized in clinical practice. Hypoxia and ischaemia are possible endpoints of several different pathologies, and the effects of hypoxia and ischaemia are greatly dependent on the condition of the fetus when it meets an insult in the peripartum period. ‘If the role of birth asphyxia as an initiating factor . . . is to be correctly assessed, a surrogate must be used that is relatively specific to birth asphyxia and not itself an early symptom of the developing disorder’.14 Pin et al.,15 when reviewing the cooling literature, concluded that ‘researchers are using very loose diagnostic criteria of perinatal asphyxia and post-asphyxial neonatal encephalopathy, making the study samples heterogeneous’ and making it difficult to compare studies with one another. Rather than attempt to define birth asphyxia, some recent population studies have identified complications, such as major abruption of the placenta or uterine rupture, that are potentially capable of interrupting oxygen supply or blood flow to the fetal brain (overt ‘sentinel events’), and examined the frequency of their occurrence in children with adverse neurological outcomes compared with typically developing children. Such sentinel events are more frequent in neonates with a clinical diagnosis of HIE (with reports ranging from 22–60%) than in those without HIE, but still do not account for a large proportion of infants with HIE. In addition to hypoxia–ischaemia, recent evidence indicates that possible antecedents include inflammation (related to infection or immune processes), coagulation defects, maternal thyroid disorders, and other factors; however, much remains unknown.
How antecedents are identified There are important differences in how strenuously antecedents of neonatal encephalopathy or HIE are sought. Most population-based studies derive information on mother, family, pregnancy, and birth chiefly from medical records generated for clinical care. Such studies sometimes include statutorily collected data from registers of birth and death, or data from specialized registers, such as those for CP or birth defects. For uncommon outcomes, prospective studies are seldom performed because of their prohibitive cost. In retrospective use of records, investigators have no control over what information was collected or how it was recorded, and cannot impose study definitions, but must depend upon diagnostic impressions recorded by clinicians caring for patients. It is seldom possible to link these data to data from other sources, to check data accuracy (which is, for example, important for outliers), or to extend the range of variables. Despite these limitations, one of the major strengths of many population studies is the strenuous effort made to collect information relevant to antecedents prior to, or blinded to, the long-term outcomes. In many population studies, abstracters of medical records are specifically 4 HIE supplement 2015, 57 (Suppl. 3): 2–7
trained and random samples of records are checked for accuracy and completeness of data extraction. Statutorily maintained registers of births can provide an excellent sampling frame for selection of controls for total population research. The amount of detail available from record review is usually directly related to the effort spent in obtaining it, though quality of information is, of course, related to the care with which the records were created. In contrast, none of the cooling studies have described methods for the systematic and rigorous examination of medical records to identify antenatal factors, or what studies were undertaken to rule out aetiological factors other than hypoxia–ischaemia. Rigorous aetiological diagnosis is seldom complete before entry into cooling trials, which usually occurs within 6 hours of birth, since the necessary review of maternal and birth records, family history, placental examination, neonatal examination for internal birth defects, and possibly metabolic or genetic testing, cannot be completed within this time frame. Many infants who meet criteria for neonatal encephalopathy are found, on closer examination, to possess more than one associated risk factor.16 Many disorders that can cause neonatal encephalopathy, malformations of internal organs, or metabolic or other genetic disorders, commonly evade diagnosis in the neonatal period. None of these disorders would be expected to respond markedly to therapeutic hypothermia, so these disorders probably contribute to the observed majority of eligible neonates who do not benefit from cooling. Except for a few that include postnatal neuroimaging, no study has yet re-analysed the outcomes of therapeutic cooling, stratified by aetiologically pertinent information available post randomization; therefore, evidence supporting the presence or absence of significant antenatal factors, and the impact of those on outcome, is lacking. In a recent retrospective referral-based study whose specific objective was to determine the proportion of neonatal HIE caused by antepartum factors, alone or in combination with intrapartum factors, data on important antenatal variables (such as growth restriction, thyroid disorders, hypertension, and infections throughout pregnancy) were missing for 12% to 20% of infants with HIE, compared with only 3% to 4% of healthy infants.13 Does aetiology alter outcome? Given the lack of antepartum evidence, the literature on therapeutic hypothermia does not allow consideration of whether or not antecedents in selected infants alters their response to treatment, yet underlying aetiologies commonly do influence outcome: smallpox and chickenpox have different prognoses despite similarities at presentation, and different causes of prematurity are associated with differing outcomes of birth at the same gestational age.17,18 It has been observed that, among term-born infants selected for cooling, those with inflammatory placental histology19 or reduced fetal movements20 respond less well to therapeutic hypothermia. It seems likely that, as with other disease-associated outcomes, the antecedents of neonatal neurological abnormality will influence both long-term outcomes and response to therapies.
SOURCES OF INFORMATION The aetiological investigation of neonatal encephalopathy should include a full review of medical records of mother, pregnancy, birth, family history, and medical workup in the neonatal period and beyond. Other sources of information likely to play a more important role in the future are placental histological examination, multiplexed microassays, and genetic or genomic evaluation. The placenta In population-based studies, macroscopic examination of the placenta has identified that placental abnormalities, such as infarction, are more common in infants with neonatal encephalopathy or HIE than in healthy neonates.19,21 However, many important features of placental pathology cannot be identified by macroscopic examination. Abnormal placental histology has been reported in more than 75% of infants with neonatal encephalopathy (whether or not it is considered to be HIE) in clinical samples,11,22 with the most common placental lesions being vasculopathy, inflammation, or both. Placental histological examination is the definitive method for establishing the presence of inflammation, its locus, severity, and acuteness or chronicity. Importantly, histological examination is the only available means of differentiating two major subcategories of placental inflammation: chorioamnionitis or funisitis caused by infection, and chronic villitis, which is most commonly an immunological response related to the breakdown of maternal–fetal tolerance.23 The prevalence of chronic villitis increases with gestational age, and is associated with a marked increase in the risk of neonatal encephalopathy.24,25 Chronic villitis may coexist with, or be difficult to distinguish from, fetal thrombotic vasculopathy of the placenta,26 which is also associated with the risk of adverse neurological outcomes,24,25 and with stillbirth and cardiac malformations.26 Risk factors for both chronic villitis and adverse neurological outcomes include prior pregnancy loss, fetal growth restriction, abnormal fetal heart rate patterns, need for emergency surgical delivery, severe acidosis, and neuroimaging evidence of basal ganglia–thalamic injury.27 Some of these characteristics are sometimes assumed to be indicators of birth asphyxia but, since they are not specific to birth asphyxia, the possibility of misattribution of aetiology is high without a placental histological examination carried out by an experienced perinatal pathologist. Correct identification of chronic villitis is also important for its treatment implications: agents under consideration are steroids and intravenous immunoglobulins. Therapeutic hypothermia is not known to be an effective therapy in infants with placental villitis. Another important condition for which the placenta provides key evidence is fetal growth restriction, present in a significantly higher proportion of neonates with neonatal encephalopathy than those without, as indicated by population studies. Fetal growth restriction is also associated with
a range of other adverse neurological outcomes including stillbirth, epilepsy, intellectual disability, autism, and schizophrenia. Fetal growth restriction is itself a condition with many aetiologies, including aberrant expression of a variety of growth factors, genetic variants, and co-occurring birth defects. Fetal growth restriction contributes more to neonatal encephalopathy than sentinel events, and much more than sentinel events to total CP.12 Placental lesions indicative of disorders of maternal and fetal blood flow, and chronic villitis, are associated with fetal growth restriction. In future, studies of placental morphology and function are likely to become increasingly important. Such data will need to be linked with information on mother, pregnancy, birth, neonatal period, and outcome in longitudinal studies. Despite limitations in the quantity and quality of evidence, current reports indicate that placental abnormality is often associated with neonatal encephalopathy, perinatal stroke,28 low Apgar scores,29–31 and term stillbirth.32,33 However, the nature of the placental abnormality differs in several ways, supporting the aetiological diversity of these clinical syndromes. As there appears to be an important role for the placenta in the illnesses underlying the majority of neonatal encephalopathy, current animal models of asphyxial HIE may be inadequate for studying neonatal encephalopathy in human infants.
Multiplexed micro-assays In addition to inflammation and hypoxia–ischaemia, risk factors for neonatal encephalopathy include coagulation disorders, such as acquired thrombophilias, autoimmune conditions of the mother or fetus, and specific infections. Aetiological diagnosis is a matter of urgent concern for immediate medical care, but has been constrained by the small volumes of blood that can be drawn from a newborn infant. Technology is now available for miniaturized multiplexed assays that can provide answers to many questions from a tiny sample of neonatal blood or cerebrospinal fluid, and which could be assembled in one or more panels for immediate investigation of the neurologically symptomatic neonate. Among analytes that might be useful are markers of severity of neonatal illness and predictors of outcome, such as S100b, neuron-specific enolase, and ubiquitin; inflammatory indicators, including C-reactive protein, cytokines and chemokines; and immune factors including complement fractions and systematic lupus markers, coagulation indicators, hormones, and identifiers of specific infective organisms. It is hoped that clinicians and medical industries can cooperate to bring such diagnostic capabilities into practical use in the near future. Genetic and genomic studies Recent investigations of CP reveal copy number variants and mutations with possible relevance to propensity for inflammation, coagulation abnormalities, vulnerability to hypoxia and, at least in the case of preterm infants, to thyroid disorders.34 Longitudinal studies indicate that hypoxia Does Aetiology of HIE Influence Outcome? Sarah McIntyre et al. 5
and inflammation are related to CP via a pathway that includes neonatal encephalopathy,12,35 and maternal thyroid disorders are also associated with risk of neonatal encephalopathy.19,36 Vulnerability to neonatal encephalopathy is likely, in future, to be a fruitful subject for genetic and genomic investigation.
WHY IS KNOWLEDGE OF THE AETIOLOGY OF NEONATAL ENCEPHALOPATHY IMPORTANT? To improve understanding of the pathobiology underlying neonatal encephalopathy, and to develop more selective strategies for prevention and treatment, it is important that we understand the aetiology of neonatal encephalopathy. Knowledge of aetiology may suggest therapies, in addition or as an alternative to hypothermia. There may be subgroups of infants with neonatal encephalopathy, with certain aetiologies (anticipated to be those with HIE, perhaps best identified as those with sentinel events the timing of which is known), who are likely to respond relatively well to cooling, while those with different aetiologies may respond less well, or may even be harmed by this treatment. WHAT NEEDS TO BE DONE? A minimal dataset should be established, to be gathered post randomization in future neonatal neuroprotective trials, seeking to maximize information of possible aetiological relevance without excessively burdening the trials. Secondary analyses of cooling trials, with stratification that incorporates information gathered after randomization, should be undertaken. Significant collaboration will be required to obtain sample sizes sufficient to support such
stratification, along with adequate funding and careful design. As criteria for cooling expand, the pool of subjects for neonatal treatment is likely to become even more heterogeneous, and the relationships among antecedents, therapies, and outcomes even less clear. Therefore, it is hoped that the next generation of trials for neonatal neuroprotection will identify major aetiological subgroups in their samples, and provide a basis for secondary analysis of study results stratified by those subgroups, in order to maximize the interpretability of their results. Investigators involved in population-based studies, and those in neonatal care units providing therapeutic hypothermia to infants with neonatal encephalopathy, have different experiences and assumptions, but knowledge from both settings is needed for further improvements to the care of these infants. How much of the outcome of therapeutic hypothermia is related to the treatment, and how much to the pathologies that led to the treatment? Are there some encephalopathic newborn infants who need additional or alternative therapy? To date, the literature of therapeutic cooling does not fully address these questions, but it is important that future trials attempt to do so. A CK N O W L E D G E M E N T S Funding sources: the Castang Foundation supported SM to travel to the meeting; the Cerebral Palsy Research Foundation supported SM; Macquarie Group Foundation and Cerebral Palsy Research Foundation supported NB; NHMRC program grant No 353514 supported EB. The authors have stated that they had no interests that might be perceived as posing a conflict or bias.
REFERENCES 1. Leviton A, Nelson KB. Problems with definitions and
8. Jacobs SE, Morley CJ, Inder TE, et al. Whole-body
14. Ellenberg JH, Nelson KB. The association of cerebral
classifications of newborn encephalopathy. Pediatr Neurol
hypothermia for term and near-term newborns with
palsy with birth asphyxia: a definitional quagmire. Dev
1992; 8: 85–90.
hypoxic-ischemic encephalopathy: a randomized con-
2. Kurinczuk JJ, White-Koning M, Badawi N. Epidemiology of neonatal encephalopathy and hypoxic-ischaemic encephalopathy. Early Hum Dev 2010; 86: 329–38.
trolled trial. Arch Pediatr Adolesc Med 2011; 165: 692– 700. 9. Simbruner G, Mittal RA, Rohlmann F, Muche R. Sys-
3. Keogh JM, Badawi N, Kurinczuk JJ, Pemberton PJ,
temic hypothermia after neonatal encephalopathy: out-
Stanley FJ. Group B streptococcus infection, not birth
comes of neo.nEURO.network RCT. Pediatrics 2010;
asphyxia. Aust N Z J Obstet Gynaecol 1999; 39: 108–
126: e771–8.
10.
10. Hayes BC, McGarvey C, Mulvany S, et al. A case–con-
Med Child Neurol 2013; 55: 210–6. 15. Pin TW, Eldridge B, Galea MP. A review of developmental outcomes of term infants with post-asphyxia neonatal encephalopathy. Eur J Paediatr Neurol 2009; 13: 224–34. 16. Nelson KB, Grether JK. Selection of neonates for neuroprotective therapies: one set of criteria applied to a population. Arch Pediatr Adolesc Med 1999; 153: 393–8.
4. Grether JK, Nelson KB. Maternal infection and cerebral
trol study of hypoxic-ischemic encephalopathy in new-
17. Basso O, Wilcox A. Mortality risk among preterm
palsy in infants of normal birth weight. J Am Med Assoc
born infants at >36 weeks gestation. Am J Obstet Gynecol
babies: immaturity versus underlying pathology. Epidemi-
1997; 278: 207–11.
2013; 209: 29.e1–29.e19.
ology 2010; 21: 521–7.
11. Chang T, Reyes C, Teng J, Placette J, Massaro AN,
18. Gagliardi L, Rusconi F, Da Fre M, et al. Pregnancy dis-
for hypoxic-ischemic encephalopathy. Ann Neurol 2012;
Nelson KB. Neonatal encephalopathy, sentinel events
orders leading to very preterm birth influence neonatal
72: 156–66.
and the placenta. J Neonatal Perinatal Med 2012; 5: 1–8.
outcomes: results of the population-based ACTION
5. Volpe JJ. Neonatal encephalopathy: an inadequate term
cohort study. Pediatr Res 2013; 73: 794–801.
6. Jacobs SE, Berg M, Hunt R, Tarnow-Mordi WO, Inder
12. McIntyre S, Blair E, Badawi N, Keogh J, Nelson KB.
TE, Davis PG. Cooling for newborns with hypoxic is-
Antecedents of cerebral palsy and perinatal death in
19. Badawi N, Kurinczuk JJ, Keogh JM, et al. Antepartum
chaemic encephalopathy. Cochrane Database Syst Rev
term and late preterm singletons. Obstet Gynecol 2013;
risk factors for newborn encephalopathy: the Western
2013; 1: CD003311.
122: 869–77.
Australian case–control study. Br Med J 1998; 317:
7. Shankaran S, Laptook AR, Ehrenkranz RA, et al.
13. Martinez-Biarge M, Diez-Sebastian J, Wusthoff CJ,
Whole-body hypothermia for neonates with hypoxic-
Mercuri E, Cowan FM. Antepartum and intrapartum
20. Bonifacio SL, Glass HC, Vanderpluym J, et al. Perinatal
ischemic encephalopathy. N Engl J Med 2005; 353:
factors preceding neonatal hypoxic-ischemic encephalop-
events and early magnetic resonance imaging in thera-
1574–84.
athy. Pediatrics 2013; 132: e952–9.
peutic hypothermia. J Pediatr 2011; 158: 360–5.
6 HIE supplement 2015, 57 (Suppl. 3): 2–7
1549–53.
21. Blair E, de Groot J, Nelson KB. Placental infarction
27. Harteman JC, Nikkels PG, Benders MJ, Kwee A, Gro-
identified by macroscopic examination and risk of cere-
enendaal F, de Vries LS. Placental pathology in full-
bral palsy in infants at 35 weeks of gestational age and
term infants with hypoxic-ischemic neonatal encephalop-
33. Flenady V, Koopmans L, Middleton P, et al. Major risk
over. Am J Obstet Gynecol 2011; 205: e1–7.
athy and association with magnetic resonance imaging
factors for stillbirth in high-income countries: a system-
22. Wintermark P, Boyd T, Gregas MC, Labrecque M,
pattern of brain injury. J Pediatr 2013; 163: 968–95.
Hansen A. Placental pathology in asphyxiated newborns
28. Elbers J, Viero S, MacGregor D, DeVeber G, Moore
meeting the criteria for therapeutic hypothermia. Am J
AM. Placental pathology in neonatal stroke. Pediatrics
Obstet Gynecol 2010; 203: 579.e1–9.
2011; 127: e722–9.
ropathology and placental pathology. Pediatr Dev Pathol 2011; 14: 345–52.
atic review and meta-analysis. Lancet 2011; 377: 1331– 40. 34. Ho NT, Furge K, Fu W, et al. Gene expression in archived newborn blood spots distinguishes infants who
23. Tamblyn JA, Lissauer DM, Powell R, Cox P, Kilby
29. Naeye RL. Underlying disorders responsible for the
MD. The immunological basis of villitis of unknown ati-
neonatal deaths associated with low Apgar scores. Biol
ology – review. Placenta 2013; 34: 846–55.
Neonate 1979; 35: 150–5.
will later develop cerebral palsy from matched controls. Pediatr Res 2013; 73: 450–6. 35. Ellenberg JH, Nelson KB. Cluster of perinatal events
24. McDonald DGM, Kelehan P, McMenamin JB, et al.
30. Locatelli A, Incerti M, Ghidini A, Greco M, Villa E,
Placental fetal thrombotic vasculopathy is associated
Paterlini G. Factors associated with umbilical artery
with neonatal encephalopathy. Hum Pathol 2004; 35:
acidemia in term infants with low Apgar scores at
36. Ellis M, Manandhar N, Shrestha PS, Shrestha L, Man-
5 min. Eur J Obstet Gynecol Reprod Biol 2008; 139: 146–
andhar DS, Costello AM. Outcome at 1 year of neonatal
50.
encephalopathy in Kathmandu, Nepal. Dev Med Child
875–80. 25. Hayes BC, Cooley S, Donnelly J, et al. The placenta in infants >36 weeks gestation with neonatal encephalopa-
31. Wong L, MacLennan AH. Gathering the evidence: cord
thy: a case control study. Arch Dis Child Fetal Neonatal
gases and placental histology for births with low Apgar
Ed 2013; 98: F233–9. 26. Faye-Petersen OM. The placenta in preterm birth. J Clin Pathol 2008; 61: 1261–75.
identifying infants at high risk for death or disability. J Pediatr 1988; 113: 546–52.
Neurol 1999; 41: 689–95.
scores. Aust N Z J Obstet Gynaecol 2011; 51: 17–21. 32. Chang KT, Keating S, Costa S, Machin G, Kingdom J, Shannon P. Third-trimester stillbirths: correlative neu-
Does Aetiology of HIE Influence Outcome? Sarah McIntyre et al. 7
