Postasphyxial Hypoxic-Ischemic Encephalopathy in Neonates: Outcome Prediction Rule Within 4 Hours of Birth | Clinical Decision Support | JAMA Pediatrics | JAMA Network
[Skip to Navigation]
Sign In
July 2006

Postasphyxial Hypoxic-Ischemic Encephalopathy in Neonates: Outcome Prediction Rule Within 4 Hours of Birth

Author Affiliations

Author Affiliations: Department of Paediatrics, Mount Sinai Hospital (Drs Shah and Ohlsson), Departments of Paediatrics (Drs Shah and Ohlsson) and Public Health Sciences (Drs Beyene and To), University of Toronto, and Division of Neonatology, Department of Paediatrics (Drs Shah and Perlman) and Research Institute (Drs Beyene and To), Hospital for Sick Children, Toronto, Ontario.

Arch Pediatr Adolesc Med. 2006;160(7):729-736. doi:10.1001/archpedi.160.7.729

Objectives  To construct and validate a model and derive a simple rule that is usable in any birth location for the prediction of outcome of term infants with severe asphyxia.

Design  Retrospective cohort study.

Setting  Regional outborn neonatal intensive care unit.

Participants  Infants with postintrapartum asphyxial hypoxic-ischemic encephalopathy (n = 375).

Main Exposures  Clinical and laboratory predictors available at age 4 hours.

Main Outcome Measures  A logistic regression model was developed and internally validated (with random sampling and based on the year of birth) for severe adverse outcome, which was defined as death or severe disability (severe cerebral palsy, severe developmental delay, sensorineural deafness, or cortical blindness singly or in combination). A simple prediction rule was derived from 3 variables.

Results  Complete data were available for 302 (92%) of the 345 infants with known outcomes (204 infants with severe adverse outcome). Six independent predictors of outcomes were identified. Using the 3 most significant predictors (chest compressions, age at onset of respiration, and base deficit), severe adverse outcome rates were 46% (95% confidence interval, 33%-58%) with none of the 3 predictors, 64% (95% confidence interval, 54%-73%) with any 1 predictor, 76% (95% confidence interval, 66%-85%) with any 2 predictors, and 93% (95% confidence interval, 81%-99%) with all of the 3 predictors present. The internal validations revealed a robust model.

Conclusions  This predictive model for neonatal hypoxic-ischemic encephalopathy provides a sliding scale of probabilities that could be used for prognostication and to design eligibility criteria for decision making including neuroprotective therapy.

Approximately 23% of the 4 million annual global neonatal deaths are attributable to asphyxia.1 Postasphyxial hypoxic-ischemic encephalopathy (HIE) occurs in approximately 1 to 2 infants per 1000 live term births.2 Significant proportions of these infants die or survive with severe long-term morbidity. Among term infants, 6% to 23% of cases of cerebral palsy (CP) are attributable to intrapartum asphyxia.3,4 Predictions of long-term outcome in the immediate neonatal period are based on clinical,5-7 biochemical,8,9 electrophysiological,10,11 and imaging12 findings. The predictive accuracy in the very early postnatal period is moderate.5,13 Higher positive predictive values might be achieved with amplitude-integrated electroencephalography (EEG), but it is not available in many hospitals globally.

Certain strategies in experimental postasphyxial HIE prevent secondary neuronal injury when given within the therapeutic window of 5 to 6 hours from the insult.14 Trials of hypothermia in humans have used clinical and/or amplitude-integrated EEG criteria for inclusion.15-17 The choice of predictive inclusion criteria in research involves ethical and clinical considerations.18 Early evaluation of risk is essential for appropriate selection of patients for neuroprotective therapies.19 To decide on neuroprotective therapy within or outside of clinical trials, easily available, robust, and accurate prediction models are needed, preferably by age 4 hours to allow time for enrollment. Such a rule will also help to monitor the efficacy of interventions and for prognostication.

The objectives of this study were to develop and validate a prognostic model for term infants with postasphyxial HIE based on clinical and laboratory information available by age 4 hours and to develop a simple prediction rule based on the variables identified in the model.



We retrospectively reviewed the records of outborn infants with moderate to severe postasphyxial HIE who were admitted to the regional outborn neonatal intensive care unit at the Hospital for Sick Children, Toronto, Ontario (referral center for 55 000-60 000 births/y), between July 1, 1985, and December 31, 2000. The majority of infants were transported by our team and admitted between ages 3 and 8 hours.

Inclusion criteria

The following inclusion criteria were adapted from the American College of Obstetricians and Gynecologists20 and Society of Obstetricians and Gynecologists of Canada21 statements. Infants who had an acute perinatal event were eligible if they met the following 3 criteria: (1) Apgar score of less than 5 at 5 minutes, metabolic acidosis (base deficit >16 mmol/L) in cord arterial blood or the first postnatal blood gas sample within 4 hours of birth, or delayed onset of breathing for more than 5 minutes; (2) need for mechanical ventilation immediately after birth; and (3) moderate to severe encephalopathy including altered state of consciousness and/or seizures.22

All of the infants had multisystem involvement23 (encephalopathy and more than 1 other organ involved) identified as defined in Table 1. This was determined from the data beyond age 4 hours to ensure appropriate inclusion.

Table 1. 
Definitions of Organ System Involvement and Time Courses of Asphyxia
Definitions of Organ System Involvement and Time Courses of Asphyxia

Exclusion criteria

Infants with encephalopathies associated with preterm birth, congenital abnormalities including subtle dysmorphism of unknown significance, a major single-organ anomaly, inborn errors of metabolism, congenital viral infection, septic shock, or cranial birth trauma, patients with evidence of antenatal events causing prenatal injury such as prolonged loss of fetal movements, arthrogryposis, or lack of fetal heart rate variability on admission of the mother to the hospital, and patients with severe meconium aspiration syndrome with persistent and prolonged postnatal hypoxemia were excluded to avoid the confounding effects on brain outcomes.

Neonatal data

Data were collected on custom-made forms from the patients' records by fully trained neonatal transport team members.24,25 Data related to the maternal medical history, obstetric history, intrapartum details, and details of the resuscitation were collected. The time of onset of gasping, irregular breathing, regular breathing, or spontaneous breathing was recorded as available both in ventilated as well as unventilated infants. The age at onset of regular respiration was defined as the age at onset of spontaneous or regular breathing. The transport team members examined and recorded the neurological condition of the infant soon after arrival at the hospital of birth. The initial and subsequent neurological data included level of consciousness, presence of spontaneous movements and coma, altered muscle tone, the age at onset of seizures (subtle or tonic-clonic), and number and dosage of anticonvulsants received. This information along with a clinical description of the attending team was considered for determination of the degree of encephalopathy, which was classified according to the criteria by Sarnat and Sarnat.22 None of the patients received prophylactic anticonvulsants. Use of muscle relaxants was not routine. Where data were available, the time course of the asphyxial insult was categorized for each subject as prolonged, acute near-total, or combined.25 Two of us (P.S. and M.P.) assigned the time course of each subject independently, and discrepancies were resolved by consensus. When consensus was not reached or the time course could not be classified owing to insufficient data, the time course was categorized as indeterminate (Table 1). Blood gas analyses were performed at variable times during the first 4 hours after birth; therefore, each base deficit value was converted to the number of SDs from the mean normal values of full-term newborns for age at sampling26 (z score).

Outcome data

Severe adverse outcome was defined in the research protocol as death or severe neurodevelopmental disability in survivors that was attributable to intrapartum asphyxial HIE. Cerebral palsy was defined as mild, moderate, or severe as suggested by Capute et al.27 Developmental delay was defined as mild (developmental quotient, 81-90), moderate (developmental quotient, 71-80), or severe (developmental quotient, <70). Severe neurodevelopmental disability was defined as the following: (1) severe CP (probably nonambulatory) diagnosed by age 12 months by certified physical or occupational therapists and a pediatrician; (2) mild or moderate CP (likely ambulatory) but with severe cortical blindness or severe sensorineural deafness; or (3) moderate CP and suspected moderate or severe developmental delay at age 12 months confirmed by developmental testing between ages 18 and 36 months by certified occupational therapists or psychologists (Bayley scores >2 SD below the mean). The remaining children were classified as free of severe adverse outcome rather than good outcome, as some of these patients had minor disabilities.28 All of the children with suspected motor or cognitive impairments were assessed by audiologists and ophthalmologists. Data regarding persistence of seizures and microcephaly were collected.

The outcome data were ascertained from health records of the neonatal follow-up clinic, neurology clinics, and readmissions when available. The neonatal follow-up clinic routinely assesses infants who had neonatal HIE at ages 4, 8, 12, and 18 to 24 months. If developmental delay or motor deficit was suspected on clinical grounds or on the basis of screening tests, developmental tests were performed at ages 18 to 30 months. Children unable to perform the developmental assessment tests owing to physical or cognitive deficits were classified in the severe adverse outcome group. The procedure for obtaining data for patients with incomplete follow-up was described previously.24 The study was approved by the institutional research ethics board.

Statistical analyses and sample size

Data were analyzed using SPSS version 12 statistical software (SPSS, Inc, Chicago, Ill). Patients with missing outcome data were excluded from analyses. Available data from patients with missing data were evaluated for randomness. Owing to the finding of nonrandomness of missing data, values were not imputed. Potential predictors selected on the basis of P<.10 in univariate logistic regression analyses were entered into multivariable models, and the odds ratios and 95% confidence intervals were computed. Goodness of fit of the model was assessed by the Hosmer-Lemeshow statistic and receiver operating characteristic curve. Internal validation was performed using 2 methods: (1) on a randomly selected derivation cohort comprising 70% of the cases and a validation cohort of the remaining 30%; and (2) temporally based by year of birth (1985-1995 as the derivation cohort and 1996-2000 as the validation cohort). Eight variables selected by univariate analyses were entered into the multivariate model. The sample size was adequate for the logistic model, as the number of events per variable exceeded 12 for both outcome categories.29 An additional stepwise prediction rule was created using the 3 variables identified in multivariate logistic regression analysis to have the highest statistical significance based on the presence of none, any 1 or 2, or all of the 3 of these predictor variables, similar to the method used by Schmidt et al.30 Predicted probabilities were then computed and compared with observed outcome rates. Likelihood ratios were calculated.



A total of 1195 patient records with relevant diagnoses (birth asphyxia, asphyxia, HIE, birth hypoxia, perinatal depression, brain anoxia, hypoxic-ischemic encephalopathy, or hypoxic encephalopathy) were reviewed; 375 patients met the eligibility criteria. Most of the exclusions (n = 656) were owing to failure to meet the primary criteria of respiratory and/or neurological depression. These included most admissions with meconium aspiration syndrome. The remaining cases lacked evidence of moderate to severe encephalopathy (n = 90) or had hemorrhagic shock (n = 29), lack of fetal heart rate reactivity or absent fetal heart rate on admission (n = 18), cranial trauma (n = 17), congenital malformations (n = 9), and congenital viral infection (n = 1). The mean number of patients with moderate to severe postasphyxial HIE per year was 24 patients (range, 7-43 patients).


Outcome data were available for a total of 345 patients (327 from the hospital records, 4 from regional follow-up clinics, 13 from primary care physicians, and 1 from the mother). The mean age at follow-up assessment was 30 months (median age, 22 months; interquartile age range, 18-36 months). Of the 345 children with outcome data, 115 died; the median age at death was 48 hours (age range, 10 hours to 86 days), and the primary cause of death was HIE in all of the cases (due to withdrawal of life support in 105 infants based on high probability of adverse outcome; 10 infants died while receiving life-sustaining treatment). Of the survivors, 111 children had severe adverse outcome, and 119 children (34%) did not have this outcome. Thus, in total, 226 children (66%) had severe adverse outcome (Figure 1). A chronic seizure disorder or microcephaly was present at the time of assessment in 64 and 57 survivors, respectively.

Figure 1. 
Outcome distribution of the entire cohort. CP indicates cerebral palsy.

Outcome distribution of the entire cohort. CP indicates cerebral palsy.

Predictor variable analysis

Predictive data were not available for 43 children (21 children with severe adverse outcome and 22 free of severe adverse outcome). The missing data were type of asphyxia (n = 30), blood gas analysis results (n = 19), and time of onset of respiration (n = 1) (some children had missing data for >1 variable). Comparisons of the predictive variables indicate that the 73 children lacking data (30 children who were missing outcome data and 43 who were missing predictor variable data) had less severe neonatal illness than the 302 children with complete data (Table 2).

Table 2. 
Comparison of Patients With and Without Outcome Data
Comparison of Patients With and Without Outcome Data

For the remaining 302 children (204 children with severe adverse outcome and 98 free of severe adverse outcome), the results of univariate and multivariable logistic regression analyses with outcome as the dependent variable are respectively reported in Table 3 and Table 4. Six independent predictors of outcome were identified. For severe adverse outcome, the accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of the multivariable model were 73%, 89%, 41%, 76%, and 63%, respectively. The receiver operating characteristic curve (Figure 2) revealed an area under the curve of 0.74 (95% confidence interval, 0.69-0.80). The results of internal validation based on random case selection revealed accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of 74%, 89%, 39%, 77%, and 62%, respectively, for the derivation cohort and 72%, 91%, 41%, 72%, and 72%, respectively, for the validation cohort. Validation based on temporal division of the cohort revealed accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of 69%, 86%, 37%, 72%, and 58%, respectively, for the derivation cohort and 76%, 90%, 41%, 78%, and 60%, respectively, for the validation cohort.

Figure 2. 
Receiver operating characteristic curve for predicted probabilities of severe adverse outcome.

Receiver operating characteristic curve for predicted probabilities of severe adverse outcome.

Table 3. 
Univariate Logistic Analyses
Univariate Logistic Analyses
Table 4. 
Logistic Regression Model*
Logistic Regression Model*

Prediction rule

The results of the logistic regression model of only 3 predictors with high statistical significance (administration of chest compressions >1 minute, onset of breathing >30 minutes, and base deficit value >16 in any blood gas analysis within the first 4 hours from birth) and of the simple count of these variables are reported in Table 5. The observed and predicted rates of severe adverse outcome in subgroups with none, any 1, any 2, or all of the 3 of these predictors are reported in Table 6; a simple count of the variables individually and in various combinations correlated in a stepwise fashion with the observed (Figure 3) and predicted outcomes.

Figure 3. 
Stepwise prediction rule for severe adverse outcome. Error bars indicate 95% confidence intervals; horizontal line, observed adverse outcome rate of 68% for the entire cohort.

Stepwise prediction rule for severe adverse outcome. Error bars indicate 95% confidence intervals; horizontal line, observed adverse outcome rate of 68% for the entire cohort.

Table 5. 
Logistic Model Using 3 Variables and Their Counts
Logistic Model Using 3 Variables and Their Counts
Table 6. 
Simple Count of Presence of 3 Variables and Observed and Predicted Outcome Status
Simple Count of Presence of 3 Variables and Observed and Predicted Outcome Status


Using a combination of 6 markers in this cohort, we developed a model with improved sensitivity (89%) and positive predictive value (76%) for severe adverse outcome compared with our previous prediction model that was based on a smaller sample.5 Internal validation demonstrated the robustness of the model. The prediction rule derived by a simple count of 3 variables greatly improved predictive power. These 3 predictors are known to be associated with poor long-term outcome.31 Individual estimates of odds of severe adverse outcome associated with each of these 3 variables were similar to that of the entire cohort; however, absence of all of the 3 variables reduced the probability of severe adverse outcome by 20% from baseline, and there was an approximately 15% risk for each variable added to the count.

Therapeutic decisions are based on the prognosis of the infant, which varies with the severity of illness. Gluckman et al15 described their hesitations about enrolling babies with the most severe postasphyxial HIE (determined by amplitude-integrated EEG) to the head-cooling trial, and based on the results, they concluded that the therapy was probably not effective for patients with severe injury. Severe cases (eg, cases with 2 or all of the 3 predictor variables present) probably incur the risk of severe disability rather than death with neuroprotective therapy. In our opinion, prognostic models that offer a simple sliding scale of probability of severe adverse outcomes such as ours will be very useful in this regard for selection purposes. We agree with the speculation that clinical criteria are likely to complement amplitude-integrated EEG,32 even when the latter is widely available.

To decide whether to institute neuroprotective therapy, predictive models are needed within the therapeutic window of 5 to 6 hours,14 preferably by age 4 hours, to allow time for enrollment. Even a highly accurate predictive model or scoring system is not useful for experimental or clinical therapy beyond this therapeutic window.7,33,34 Major issues with existing models include inadequate sample size, selection bias, timing of the model, outcome assessment, and lastly, attrition.31,35 Perlman36 reviewed the measures of diagnosis and severity of asphyxia that were used during labor and in the delivery room and suggested that these markers are useful in combination to identify infants who will progress to develop HIE or seizures. Apart from an earlier study from our institution,5 few studies have built predictive models for the first 4 hours after birth. Perlman and Risser35 studied 96 infants with asphyxia and identified a 5-minute Apgar score of less than 5, intubation in the delivery room, and fetal acidemia as predictors of risk of seizures within 4 hours of birth in 5 patients. Carter et al37 developed a scoring system to predict multiorgan dysfunction based on fetal heart rate abnormalities, umbilical arterial acidosis, and 5-minute Apgar score. Other nonclinical markers such as serum or urine S100 levels,9 urinary lactate-creatinine ratio,8 and amplitude-integrated EEG results10,38 have shown promise as early markers of severity of HIE, but their relationships to long-term outcome have not been tested. The need for a special assay and expertise to interpret the results may limit usefulness. Imaging modalities such as magnetic resonance imaging, magnetic resonance spectroscopy, and diffusion-weighted imaging are recommended between ages 2 to 8 days,39 but their prognostic utility has not been established even on day 1.

To interpret our results, methodological issues need consideration. Despite a rigorous selection method, we cannot exclude the presence of antenatal events in some infants. It could be argued that clinical practice and outcomes changed during the study period. However, we are not aware of any substantive change in therapy of postasphyxial HIE between 1985 and 2000 or of reports of secular changes in outcomes. No change in outcome, including the ratio of death to severe disability, was observed over time. Other factors that may cause instability of predictive models are changes in the diagnostic criteria of postasphyxial HIE20,21,40-42 and the relatively low prevalence of intrapartum asphyxia as a cause of neonatal encephalopathy and CP.18,43 Patients lost to follow-up exhibited less severe neonatal illness and were therefore more likely to have relatively good outcomes; however, their outcomes remain uncertain. The great majority of severe adverse outcomes are diagnosed by age 12 months. Shankaran et al44 found that neurological examination at age 12 months can predict severe disability at age 5 years. We acknowledge that variability in developmental assessment may have introduced bias; however, the severe adverse outcomes defined in this study are far from subtle and have the ability to detect major differences between groups.

The referral patterns and the eligibility criteria for this study resulted in a cohort with characteristics including outcomes similar to those in infants enrolled in neuroprotective therapy trials15-17 and infants admitted to children's hospitals providing similar regional services.44,45 The inclusion of infants in our cohort with only moderate to severe encephalopathy is justified owing to the high likelihood of favorable outcomes in infants with mild encephalopathy.46 Combining patients with moderate and severe encephalopathy allowed us to develop a pragmatic model; however, we are aware that the outcome spectrum is different for these 2 categories. The adverse outcome rate in patients with moderate and severe encephalopathy (66%) was somewhat higher than in the control groups in 2 recent clinical trials (60%17 and 62%15), presumably owing to exclusion of infants in extremis.

The advantages of our model include selection of infants with moderate to severe HIE at high risk for adverse outcomes, a low attrition rate, simplicity of variables based on available clinical and biochemical criteria, and adequate power. Despite the selection bias against patients with most predictable outcomes (those with the most severe disease and the mildest disease), our results have high predictive values with narrow confidence limits and provide a sliding scale rather than categorical criteria.

Prognostic certainty is often unattainable. In patients referred to our center even with none of the 3 predictors present, there was a 46% possibility of poor outcome. Conversely, even with all of the 3 indicators present, there was a 7% possibility that the infant would remain free of severe adverse outcome. We caution that this prediction rule should be used only for the patients who meet the eligibility criteria outlined earlier. Improved models will depend on a better definition of asphyxia, precise data on predictors, and external validation.

Correspondence: Prakesh S. Shah, Department of Paediatrics, Mount Sinai Hospital, Room 775A, 600 University Ave, Toronto, Ontario, Canada M5G 1X5 (

Accepted for Publication: December 27, 2005.

Author Contributions:Study concept and design: Shah, Beyene, To, and Ohlsson. Acquisition of data: Shah. Drafting of the manuscript: Shah. Critical revision of the manuscript for important intellectual content: Shah, Beyene, To, and Ohlsson. Statistical analysis: Beyene and To. Administrative, technical, and material support: Shah, Ohlsson, and Perlman. Study supervision: Beyene, Ohlsson, and Perlman.

Lawn  JECousens  SZupan  J 4 million neonatal deaths: when? where? why?  Lancet 2005;365891- 900PubMedGoogle ScholarCrossref
Low  JALindsay  BGDerrick  EJ Threshold of metabolic acidosis associated with newborn complications  Am J Obstet Gynecol 1997;1771391- 1394PubMedGoogle ScholarCrossref
Nelson  KB The epidemiology of cerebral palsy in term infants  Ment Retard Dev Disabil Res Rev 2002;8146- 150PubMedGoogle ScholarCrossref
Volpe  JJVolpe  JJed Hypoxic-ischemic encephalopathy: intrauterine assessment  Neurology of the Newborn. Philadelphia, Pa WB Saunders Co2000;277- 295Google Scholar
Ekert  PPerlman  MSteinlin  MHao  Y Predicting the outcome of postasphyxial hypoxic-ischemic encephalopathy within 4 hours of birth  J Pediatr 1997;131613- 617PubMedGoogle ScholarCrossref
van de Riet  JEVandenbussche  FPLe Cessie  SKeirse  MJ Newborn assessment and long-term adverse outcome: a systematic review  Am J Obstet Gynecol 1999;1801024- 1029PubMedGoogle ScholarCrossref
Miller  SPLatal  BClark  H  et al.  Clinical signs predict 30-month neurodevelopmental outcome after neonatal encephalopathy  Am J Obstet Gynecol 2004;19093- 99PubMedGoogle ScholarCrossref
Huang  CCWang  STChang  YCLin  KPWu  PL Measurement of the urinary lactate:creatinine ratio for the early identification of newborn infants at risk for hypoxic-ischemic encephalopathy  N Engl J Med 1999;341328- 335PubMedGoogle ScholarCrossref
Nagdyman  NKomen  WKo  HKMuller  CObladen  M Early biochemical indicators of hypoxic-ischemic encephalopathy after birth asphyxia  Pediatr Res 2001;49502- 506PubMedGoogle ScholarCrossref
Hellstrom-Westas  LRosen  ISvenningsen  NW Predictive value of early continuous amplitude integrated EEG recordings on outcome after severe birth asphyxia in full term infants  Arch Dis Child Fetal Neonatal Ed 1995;72F34- F38PubMedGoogle ScholarCrossref
Taylor  MJMurphy  WJWhyte  HE Prognostic reliability of somatosensory and visual evoked potentials of asphyxiated term infants  Dev Med Child Neurol 1992;34507- 515PubMedGoogle ScholarCrossref
Rutherford  MAPennock  JMCounsell  SJ  et al.  Abnormal magnetic resonance signal in the internal capsule predicts poor neurodevelopmental outcome in infants with hypoxic-ischemic encephalopathy  Pediatrics 1998;102323- 328PubMedGoogle ScholarCrossref
Shalak  LFLaptook  ARVelaphi  SCPerlman  JM Amplitude-integrated electroencephalography coupled with an early neurologic examination enhances prediction of term infants at risk for persistent encephalopathy  Pediatrics 2003;111351- 357PubMedGoogle ScholarCrossref
Gunn  AJ Cerebral hypothermia for prevention of brain injury following perinatal asphyxia  Curr Opin Pediatr 2000;12111- 115PubMedGoogle ScholarCrossref
Gluckman  PDWyatt  JSAzzopardi  D  et al.  Selective head cooling with mild systemic hypothermia after neonatal encephalopathy: multicentre randomised trial  Lancet 2005;365663- 670PubMedGoogle ScholarCrossref
Inder  TEHunt  RWMorley  CJ  et al.  Randomized trial of systemic hypothermia selectively protects the cortex on MRI in term hypoxic-ischemic encephalopathy  J Pediatr 2004;145835- 837PubMedGoogle ScholarCrossref
Shankaran  SLaptook  AREhrenkranz  RA  et al. National Institute of Child Health and Human Development Neonatal Research Network, Whole-body hypothermia for neonates with hypoxic-ischemic encephalopathy  N Engl J Med 2005;3531574- 1584PubMedGoogle ScholarCrossref
Nelson  KBGrether  JK Selection of neonates for neuroprotective therapies: one set of criteria applied to a population  Arch Pediatr Adolesc Med 1999;153393- 398PubMedGoogle ScholarCrossref
Ferriero  DM Neonatal brain injury  N Engl J Med 2004;3511985- 1995PubMedGoogle ScholarCrossref
Committee on Obstetric Practice, American College of Obstetricians and Gynecologists, ACOG committee opinion: inappropriate uses of the terms fetal distress and birth asphyxia  Int J Gynaecol Obstet 1998;61309- 310PubMedGoogle ScholarCrossref
 Policy statement: task force on cerebral palsy and neonatal asphyxia (part 1)  J SOGC 1996;181267- 1279Google Scholar
Sarnat  HBSarnat  MS Neonatal encephalopathy following fetal distress: a clinical and electroencephalographic study  Arch Neurol 1976;33696- 705PubMedGoogle ScholarCrossref
Shah  PRiphagen  SBeyene  JPerlman  M Multiorgan dysfunction in infants with post-asphyxial hypoxic-ischaemic encephalopathy  Arch Dis Child Fetal Neonatal Ed 2004;89F152- F155PubMedGoogle ScholarCrossref
Klinger  GBeyene  JShah  PSPerlman  M Do hyperoxaemia or hypocapnia add to the risk of brain injury after intrapartum asphyxia?  Arch Dis Child Fetal Neonatal Ed 2005;90F49- F53PubMedGoogle ScholarCrossref
Shah  PSRaju  NVBeyene  JPerlman  M Recovery of metabolic acidosis in term infants with postasphyxial hypoxic-ischemic encephalopathy  Acta Paediatr 2003;92941- 947PubMedGoogle ScholarCrossref
Swanstrom  SBratteby  LE Metabolic effects of obstetric regional analgesia and of asphyxia in the newborn infant during the first two hours after birth, III: adjustment of arterial blood gases and acid-base balance  Acta Paediatr Scand 1981;70811- 818PubMedGoogle ScholarCrossref
Capute  AJShapiro  BKPalmer  FB Spectrum of developmental disabilities: continuum of motor dysfunction  Orthop Clin North Am 1981;123- 22PubMedGoogle Scholar
Robertson  CMFiner  NN Long-term follow-up of term neonates with perinatal asphyxia  Clin Perinatol 1993;20483- 500PubMedGoogle Scholar
Altman  DGRoyston  P What do we mean by validating a prognostic model?  Stat Med 2000;19453- 473PubMedGoogle ScholarCrossref
Schmidt  BAsztalos  EVRoberts  RSRobertson  CMSauve  RSWhitfield  MF Impact of bronchopulmonary dysplasia, brain injury, and severe retinopathy on the outcome of extremely low-birth-weight infants at 18 months: results from the trial of indomethacin prophylaxis in preterms  JAMA 2003;2891124- 1129PubMedGoogle ScholarCrossref
Peliowski  AFiner  NNSinclair  JBedBracken  MBed Birth asphyxia in the term infant  Effective Care of the Newborn Infant. Oxford, England Oxford University Press1992;249- 279Google Scholar
Gunn  AJGluckman  PDWyatt  JSThoresen  MEdwards  AD Author's reply: selective head cooling after neonatal encephalopathy  Lancet 2005;3651619- 1620Google ScholarCrossref
Caravale  BAllemand  FLibenson  MH Factors predictive of seizures and neurologic outcome in perinatal depression  Pediatr Neurol 2003;2918- 25PubMedGoogle ScholarCrossref
Kadri  MShu  SHolshouser  B  et al.  Proton magnetic resonance spectroscopy improves outcome prediction in perinatal CNS insults  J Perinatol 2003;23181- 185PubMedGoogle ScholarCrossref
Perlman  JMRisser  R Can asphyxiated infants at risk for neonatal seizures be rapidly identified by current high-risk markers?  Pediatrics 1996;97456- 462PubMedGoogle Scholar
Perlman  JM Intrapartum hypoxic-ischemic cerebral injury and subsequent cerebral palsy: medicolegal issues  Pediatrics 1997;99851- 859PubMedGoogle ScholarCrossref
Carter  BSMcNabb  FMerenstein  GB Prospective validation of a scoring system for predicting neonatal morbidity after acute perinatal asphyxia  J Pediatr 1998;132619- 623PubMedGoogle ScholarCrossref
al Naqeeb  NEdwards  ADCowan  FMAzzopardi  D Assessment of neonatal encephalopathy by amplitude-integrated electroencephalography  Pediatrics 1999;1031263- 1271PubMedGoogle ScholarCrossref
Ment  LRBada  HSBarnes  P  et al.  Practice parameter: neuroimaging of the neonate: report of the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society  Neurology 2002;581726- 1738PubMedGoogle ScholarCrossref
Australian and New Zealand Perinatal Societies, The origin of cerebral palsy: a consensus statement  Med J Aust 1995;16285- 90PubMedGoogle Scholar
MacLennan  A A template for defining a causal relation between acute intrapartum events and cerebral palsy: international consensus statement  BMJ 1999;3191054- 1059PubMedGoogle ScholarCrossref
American College of Obstetricians and Gynecologists' Task Force on Neonatal Encephalopathy and Cerebral Palsy, Neonatal Encephalopathy and Cerebral Palsy: Defining the Pathogenesis and Pathophysiology.  Washington, DC American College of Obstetricians and Gynecologists2003;1- 85
Badawi  NKurinczuk  JJKeogh  JM  et al.  Intrapartum risk factors for newborn encephalopathy: the Western Australian case-control study  BMJ 1998;3171554- 1558PubMedGoogle ScholarCrossref
Shankaran  SWoldt  EKoepke  TBedard  MPNandyal  R Acute neonatal morbidity and long-term central nervous system sequelae of perinatal asphyxia in term infants  Early Hum Dev 1991;25135- 148PubMedGoogle ScholarCrossref
Cowan  FRutherford  MGroenendaal  F  et al.  Origin and timing of brain lesions in term infants with neonatal encephalopathy  Lancet 2003;361736- 742PubMedGoogle ScholarCrossref
Robertson  CMFiner  NNGrace  MG School performance of survivors of neonatal encephalopathy associated with birth asphyxia at term  J Pediatr 1989;114753- 760PubMedGoogle ScholarCrossref