BACKGROUND. Symptomatic neonatal hypoglycemia may be associated with later neurodevelopmental impairment. Brain injury patterns identified on early MRI scans and their relationships to the nature of the hypoglycemic insult and neurodevelopmental outcomes are poorly defined.
METHODS. We studied 35 term infants with early brain MRI scans after symptomatic neonatal hypoglycemia (median glucose level: 1 mmol/L) without evidence of hypoxic-ischemic encephalopathy. Perinatal data were compared with equivalent data from 229 term, neurologically normal infants (control subjects), to identify risk factors for hypoglycemia. Neurodevelopmental outcomes were assessed at a minimum of 18 months.
RESULTS. White matter abnormalities occurred in 94% of infants with hypoglycemia, being severe in 43%, with a predominantly posterior pattern in 29% of cases. Cortical abnormalities occurred in 51% of infants; 30% had white matter hemorrhage, 40% basal ganglia/thalamic lesions, and 11% an abnormal posterior limb of the internal capsule. Three infants had middle cerebral artery territory infarctions. Twenty-three infants (65%) demonstrated impairments at 18 months, which were related to the severity of white matter injury and involvement of the posterior limb of the internal capsule. Fourteen infants demonstrated growth restriction, 1 had macrosomia, and 2 had mothers with diabetes mellitus. Pregnancy-induced hypertension, a family history of seizures, emergency cesarean section, and the need for resuscitation were more common among case subjects than control subjects.
CONCLUSIONS. Patterns of injury associated with symptomatic neonatal hypoglycemia were more varied than described previously. White matter injury was not confined to the posterior regions; hemorrhage, middle cerebral artery infarction, and basal ganglia/thalamic abnormalities were seen, and cortical involvement was common. Early MRI findings were more instructive than the severity or duration of hypoglycemia for predicting neurodevelopmental outcomes.
Transient low blood/plasma glucose levels are common during the period of metabolic transition to the extrauterine environment among infants born at term.1,2 In a significant minority, hypoglycemia is associated with acute neurologic dysfunction, and it has been associated with long-term neurodevelopmental impairment.3,4 The neuroanatomic substrate of injury and its relationships with the severity and duration of hypoglycemia in humans are unclear; consequently, long-term outcomes are difficult to predict.
There is no universally accepted, “safe” blood glucose level for newborns, partly because individual susceptibility to brain injury varies, on the basis of factors such as gestational age (GA), type and volume of early milk feedings, presence of comorbid conditions, and ability of the infant to produce and to use alternative cerebral fuels. In many centers, this has led to the development of guidelines designed to detect infants at high risk and the implementation of operational thresholds for intervention, such as those proposed by Cornblath et al.5
Animal studies and human postmortem studies suggest that severe prolonged hypoglycemia gives rise to distinct patterns of brain injury, with a propensity for occipital and parietal cortex and subcortical white matter (WM) involvement.6–8 This regional susceptibility to occipital and parietal WM involvement in hypoglycemic brain injury also has been reported for infants with hypoglycemia,9–15 but studies were limited to small patient groups, including some infants with hypoxic-ischemic encephalopathy (HIE), which confounded image analysis. No study has defined, with a large number of infants with symptomatic hypoglycemia but without HIE, the range of patterns seen on early brain MRI scans, relating the patterns to clinical presentation and neurodevelopmental outcomes. The aims of this study were to document, in term infants with symptomatic neonatal hypoglycemia, (1) patterns of brain injury on early MRI scans, (2) any relationship between brain injury and the documented severity and duration of hypoglycemia, (3) prenatal or perinatal differences between neonates with hypoglycemia and neurologically normal, low-risk, term infants, and (4) any relationship between brain MRI patterns and neurodevelopmental outcomes.
Ethical approval was obtained from the Hammersmith Hospital Research Ethics Committee.
Infants who were inborn or were referred to the Hammersmith Hospital/Queen Charlotte's and Chelsea Hospital and the Robert Steiner MRI unit at the Hammersmith Hospital for MRI between 1992 and 2006, after ≥1 episode of hypoglycemia, were identified. Study entry criteria were (1) GA of >36 completed weeks at birth, (2) ≥1 documented episode of hypoglycemia (blood or plasma glucose concentration of ≤2.6 mmol/L) associated with acute neurologic dysfunction during the first postnatal week, and (3) MRI at postnatal age of <6 weeks. Exclusion criteria were congenital infections, major brain or other malformations, multiple dysmorphic features, chromosomal abnormalities, and evidence of HIE.
Hypoglycemia was defined as stated in criterion 2, and the following factors were recorded: lowest glucose level, apparent age and mode of presentation, and apparent duration and nature of any symptoms. Symptoms included poor feeding, hypothermia, jitteriness, hypotonia, irritability, lethargy, seizures, cyanosis, and apnea. Hypoglycemia was classified as brief if it responded to the first intervention and as prolonged/recurrent if it did not. Hypoglycemia was classified as severe if the lowest documented blood/plasma glucose level was ≤1.5 mmol/L.
Prenatal and Perinatal Data
Data were extracted from the medical notes for all infants with hypoglycemia. Equivalent data for comparison were available from a well-studied cohort of 229 low-risk term infants recruited from the postnatal ward at Queen Charlotte's and Chelsea Hospital in 1996–1997. Those infants were considered neurologically normal at birth and later16,17 and hereafter are referred to as control subjects. Prenatal factors recorded were maternal age, race, parity, obstetric history (infertility or multiple pregnancy), medical history of systemic disease, any pregnancy-related illness, hypertension (essential or pregnancy induced), abdominal pain in the third trimester, bleeding in any trimester, maternal infection, and family history of seizures or other neurologic disorders.
Perinatal factors recorded were acute intrapartum events, maternal intrapartum fever (>38°C), epidural or general anesthesia, induction or augmentation of labor, intrapartum complications (shoulder dystocia, cord around the neck, or breech presentation), meconium-stained liquor, mode of delivery, Apgar scores at 1 and 5 minutes, and resuscitation measures needed. GA, gender, birth weight, and head circumference (HC) were recorded. The GA and gender were used to determine the birth weight and HC percentiles from the British growth reference tables. Minor resuscitation was defined as requiring oxygen or intermittent positive pressure ventilation, and major resuscitation was defined as requiring intubation (no infant required cardiac compressions or drugs).
MRI and Analysis
Case subjects underwent MRI as part of the investigation of their clinical symptoms. They were sedated by using orally administered chloral hydrate (30–50 mg/kg). Parental consent for MRI was obtained. Infants wore ear protection and were monitored by using pulse oximetry and electrocardiography during scanning. An experienced, MRI-trained neonatologist was in attendance throughout the procedure. MRI was conducted by using a 1.0-, 1.5-, or 3-T Philips system (Philips Medical Systems, Best, Netherlands), depending on the system being used in our unit at the time of presentation. Minimal image acquisition included a T1-weighted sequence in the transverse and sagittal planes and a T2-weighted sequence in the transverse plane.
All MRI scans were assessed by an experienced neuroradiologist (Dr Rutherford) for anatomic features and abnormal signal intensity (SI). Cortex was described as appropriate cortical configuration and SI for GA, loss of cortical markings, or abnormally high SI on T1-weighted MRI scans (known as cortical highlighting). WM was graded as follows: 0, normal; 1, mild, that is, mildly increased T1- or T2-weighted MRI SI in the periventricular or deep WM; 2, moderate, that is, more-marked increase in T1- or T2-weighted MRI SI in the periventricular, deep, or subcortical WM and/or small punctate hemorrhage; 3, severe, that is, unilateral or bilateral overt infarction or a severe focal hemorrhagic lesion. The nature and location of WM damage also were documented.
Basal ganglia/thalami (BGT) was graded as follows: 0, normal/mild; 1, moderate/severe (grouped together because both are associated with the development of cerebral palsy [CP]). Mild indicates focal, possibly transient, SI abnormalities, with apparently normal myelination in the posterior limb of the internal capsule (PLIC). Moderate indicates multifocal BGT abnormalities, usually with abnormal or equivocal PLIC findings. Severe indicates abnormalities of the entire BGT and PLIC.
Outcomes at a minimum of 18 months were determined by using a standardized neurologic assessment16 and Griffiths’ Mental Developmental Scales,18 from which a development quotient (DQ) was calculated. CP was defined according to published criteria.19 Although we had longer-term information for many children, outcomes at 18 to 24 months were analyzed for uniformity. Head growth, occurrence of seizures, and specific visual or speech/language difficulties at any age (range: 2–10 years) were noted. For children who could not be evaluated by us, specific information was obtained from local pediatric and child developmental services. Outcomes were classified as follows: normal, DQ of >85, normal neurologic examination results and head growth, and absence of seizures; mild, DQ of >85 but with mild motor impairment (including mild hemiplegia with independent finger movements), suboptimal head growth, specific visual or language problems, or seizures or DQ of <85 but >70 without other complications; moderate, DQ of <70, severe hemiplegia, or mild quadriplegia, with or without seizures; severe, unassessable with the Griffiths’ scales, suboptimal head growth, severe spastic quadriplegia, or ongoing seizures.
Data were analyzed by using StatsDirect 2.5.6 (StatsDirect Ltd, Altrincham, Cheshire, United Kingdom). Where appropriate, unpaired Student's t tests or Mann-Whitney U tests for continuous data and χ2 tests or Fisher's exact tests for categorical variables were used. Data were assessed for normality by using the Sharipo-Wilk test, and the level of significance was set at P < .05.
Eighty-four infants with ≥1 documented episode of early postnatal hypoglycemia and early MRI were identified. Thirty-nine infants were excluded because of evidence of HIE and 10 because their MRI was performed at >6 postnatal weeks, leaving 35 infants who fulfilled all study criteria.
Overall Case and Control Infant Characteristics
The case and control infants were comparable with respect to GA (case subjects: mean: 39.47 weeks; range: 37–42 weeks; control subjects: mean: 39.56 weeks; range: 37–42 weeks; P = .92), birth HC (case subjects: mean: 34.28 cm; range: 31–47.6 cm; control subjects: mean: 34.55 cm; range: 31–38; P = .34), and multiple births. More case infants demonstrated growth restriction, defined as birth weight of <10th percentile (case subjects: n = 14, 39%; control subjects: n = 17, 7%; P = <.0001), and lower birth weight (case subjects: mean: 3098 g; control subjects: mean: 3395 g; P = .008). There were more boys among the infants with hypoglycemia (P = .0019).
Infants with hypoglycemia were more likely to be born to mothers who developed pregnancy-induced hypertension (P = .04) or had a family history of seizures or neurologic disease (P = .0008). Six of the case subjects had a family history of seizures. In 1 case, these occurred in the neonatal period in a second-degree relative, although the cause was unknown. In 5 cases, seizures began outside the neonatal period, in first-degree relatives in 2 cases and in second-degree relatives in 3 cases.
There was a significantly lower rate of spontaneous vaginal delivery among case infants (P = .01), and more case infants were delivered through emergency cesarean section (P < .0001) (Table 1). Case infants were more likely to require minor or major resuscitation at birth (P = .04).
The majority (n = 30; 86%) had severe hypoglycemia (<1.5 mmol/L) on ≥1 occasion (Table 2). Twenty-two infants had transient hypoglycemia that resolved promptly with glucose administration. Eight of those infants had intrauterine growth restriction, and 1 had hypocortisolemia. Thirteen infants had prolonged/recurrent hypoglycemia. Six had intrauterine growth restriction, and 3 had an underlying metabolic or endocrine diagnosis (1 each of glucose-6-phosphate dehydrogenase deficiency, ketotic hypoglycemia, and nonketotic hyperinsulinemic hypoglycemia). Twenty of the 35 infants presented on day 1.
Twenty-two infants experienced symptoms including seizures, 8 had seizures alone, and 5 did not develop seizures. Ten infants were detected after routine screening in the neonatal unit, whereas the remainder were identified only after exhibiting symptoms.
MRI scans were acquired at a median age of 9 days (range: 1–42 days).
Patterns of Injury
Two case subjects with transient hypoglycemia (lowest glucose levels: 1.3 mmol/L and 1.2 mmol/L) on day 1 had normal MRI findings (Table 3).
Almost all infants (n = 33; 94%) had some evidence of WM abnormalities, with the majority (80%) being classified as either moderate (n = 13) or severe (n = 15). Six infants with moderate and 1 with severe WM injury also had focal small punctate lesions (Fig 1) consistent with hemorrhagic injury. Two infants with severe WM changes had larger areas of focal hemorrhage, 3 had unilateral middle cerebral artery (MCA) territory infarctions (Fig 2), and 10 had more-diffuse WM infarction, most with a predominantly posterior parasagittal distribution (Fig 3).
Among all groups with WM injury, 13 infants (39%) had global involvement (Fig 4), 4 with the most severe changes in the posterior WM. Six infants (18%) had posterior changes alone (Fig 3), and 12 (36%) had injury confined to the periventricular regions. Only 2 infants (6%) had solely unilateral lesions, with the remainder having bilateral lesions (1 infant with MCA infarction had additional contralateral changes) (Fig 2). Of the infants with bilateral injury, 15 had symmetrical changes.
BGT and PLIC
Fourteen infants (40%) had some involvement of the BGT. Four cases were moderate or severe and associated with abnormal PLIC findings. One of those 4 patients had extensive unilateral MCA territory infarction (Fig 2), 1 had multiple hemorrhagic lesions throughout the BGT, 1 had complete unilateral loss of myelin associated with abnormal SI in the lentiform nuclei on the same side and multiple areas of hemorrhage in the WM, and 1 had only mildly abnormal SI within the PLIC associated with posterior parasagittal infarction. The 10 milder BGT lesions predominantly involved abnormal SI in either the thalamus or lentiform nuclei; 3 infants had changes in the globus pallidi (Fig 5), and all had a normal-appearing PLIC. The 2 other infants with focal arterial infarction had no involvement of the PLIC.
Eighteen (51%) infants had cortical abnormality. In 34% this was cortical highlighting and in the majority of cases, this was widespread and in a parasagittal distribution. Twenty-six percent of infants had loss of cortical markings, a finding associated with early cortical infarction and often seen adjacent to WM injury, giving rise to a loss of gray matter/WM differentiation. A small number of infants (n = 3) had both abnormalities.
Injury in other sites was rare. One infant had abnormal SI in the brainstem, 1 had a small hemorrhagic lesion and 1 had abnormal signal intensity in the cerebellum. Five infants had extracerebral hemorrhage, and 3 had germinal layer-intraventricular hemorrhage.
Patterns of Brain Injury According to Severity of Hypoglycemia
The median value of the lowest recorded blood glucose reading was 1.0 mmol/L (range: 0–2.5 mmol/L). Thirty infants had severe hypoglycemia (≤1.5 mmol/L), and 5 had mild hypoglycemia (<2.6 to >1.5 mmol/L). The severity of hypoglycemia was not associated with specific patterns of injury.
Patterns of Brain Injury According to Duration of Hypoglycemia
Twenty-two infants responded rapidly to treatment, without subsequent recurrence. In 13 infants, the hypoglycemia was either prolonged or recurrent. No striking distinguishing MRI features were seen for infants with transient versus prolonged/recurrent hypoglycemia.
Patterns of Brain Injury According to the Presence of Seizures
Nine children had seizures over a period of several days; all had moderate or severe WM injury and 7 had cortical involvement. Five children did not develop seizures; 1 had normal MRI findings, 1 had mildly abnormal WM, 3 had moderately abnormal WM, and none had cortical involvement. For the remaining 21 children who had transient seizures on day 1, the MRI findings ranged from normal findings to a severe pattern of injury.
Information was available for 34 of 35 children at a minimal age of 18 months. Eight children had normal outcomes, 15 had mild impairments, 8 had moderate impairments, and 3 had severe impairments. The child for whom no outcome data were available had normal scan results and early transient hypoglycemia.
Twenty-five children demonstrated normal findings, 6 had CP (3 spastic quadriplegia and 3 hemiplegia), and 3 had mild motor delays. Twenty-nine children (85%) were walking by 2 years of age. Four children were not yet 2 years of age when outcome data were obtained; however, all except 1 were already walking by the time of assessment. The exception was a child who had developed severe spastic quadriplegia.
Nineteen children were functioning in the reference range, 3 had mild delays, 5 had moderate delays, and 5 could not be assessed with standard testing methods. Two children thought to have normal cognitive abilities had some speech and language delays.
Twelve children developed later seizures. Three had infantile spasms, 2 had generalized seizures, 1 had focal seizures, and 1 had seizures associated with later recurrent hypoglycemia. The seizure types were not specified for 2 infants, and 3 infants had febrile seizures. All seizures developed before the age of 2 years.
Nine of 28 children with HC measurements at 2 years had suboptimal head growth (decrease of >2 SD across the percentiles, compared with HC at birth).
The following abnormalities were noted: squint (n = 5), field defect (n = 2), cortical visual impairment (n = 2), immature visual attention and tracking (n = 1), and visuospatial difficulties (n = 1).
Relationship Between Hypoglycemia and Outcomes
All except 1 of the infants with early transient hypoglycemia and all 5 infants who did not have seizures had outcomes within the reference range or mildly abnormal outcomes. The infant from this group with a moderately abnormal outcome had an extensive MCA infarction, with severe hemiplegia, infantile spasms, a visual field defect, and delayed speech and language development. Outcomes were more varied for infants with prolonged or severe hypoglycemia.
MRI Findings and Outcomes
Outcomes correlated well with MRI findings (Table 4). Infants with CP either had a unilateral focal infarction or posterior parasagittal infarction involving the PLIC, giving them hemiplegia (n = 2), or had severe bilateral WM injury, giving them spastic quadriplegia (n = 3); 1 infant with hemiplegia had bilateral WM injury with asymmetrical involvement of the thalami. The other 2 infants who had PLIC changes had delayed fine motor skills at the follow-up evaluations. Nine of the 14 children who had severe global WM changes had very low DQ values or could not be assessed; the other 5 children had normal or mildly abnormal outcomes, but all had either unilateral or asymmetrical WM injury with a normal BGT. Eight children had WM infarction (2 asymmetrical and 6 symmetrical) in the posterior parasagittal region; asymmetrical injury was associated with milder outcomes than symmetrical injury.
This is the first study of a large cohort of term symptomatic infants with hypoglycemia that assesses MRI injury patterns in relation to clinical presentations and neurodevelopmental outcomes. We show that the patterns of brain injury detected on early MRI scans are more varied than those described in the literature. Previous smaller studies with heterogeneous cohorts reported that the most severe injury is localized to the parietal and occipital cortex and subcortical WM.9–15,20 In our study, which was confined to term infants and excluded infants with HIE, only 29% of subjects showed a primarily posterior pattern of injury.
Why the parietal and occipital lobes should be most severely affected after neonatal hypoglycemia is unclear. There are several putative mechanisms for hypoglycemia-induced cellular injury, including excitatory neurotoxins active at N-methyl-d-aspartate receptors,21,22 increased mitochondrial free radical generation and initiation of apoptosis,23 and altered cerebral energetic characteristics.24 The parietal/occipital lobes are not known to be particularly vulnerable to these effects, but some occipital regional vulnerability has been demonstrated as increased regional cerebral blood flow during hypoglycemia, with significant reduction in regional glucose uptake in the occipital WM.25 This might make this region more vulnerable, although our data suggest that the posterior WM and cortex may not be as selectively vulnerable as suggested previously.
WM injury was the most common abnormality among the group, and 80% of cases were classified as moderate or severe. The vulnerability of WM to hypoglycemia-induced injury raises the possibility of an exaggerated pathophysiological response, compared with other cerebral tissue types. Abnormal cerebral blood flow occurs in preterm human neonates and in animal models of neonatal hypoglycemia,26–28 which might account for the periventricular distribution of injury seen in some of our infants. Given the maturation-dependent vulnerability of the WM of preterm infants to various injurious processes,29 it is possible that infants who are vulnerable to the adverse cerebral effects of hypoglycemia have less-mature WM, although we did not see evidence of this in conventional analyses.
Approximately 30% of our cohort had WM lesions consistent with hemorrhage; 20% were focal, and 80% were multiple and punctate, mainly in the periventricular WM. It is unclear whether these lesions definitely represent hemorrhage, because we have no postmortem histologic data; it is possible that they are ischemic or ischemic with secondary hemorrhage, and they may be of arterial or venous origin. Hemorrhage has not been reported previously in association with hypoglycemia. Hypoglycemia and polycythemia may cooccur, and this raises the possibility of cerebral venous thrombosis, although no definite evidence for this was seen on the MRI scans. However, the mean postnatal age at the time of MRI was 9 days, which is late for reliable detection of venous thrombosis, and optimal MRI sequences were not used for all infants. Focal arterial territory infarction occurred in 3 infants. An association between this form of stroke, which may have a thrombotic or embolic pathogenesis, and hypoglycemia was reported by Benders et al30 for preterm infants, although no clear explanation was elucidated.
Severity of WM injury was a good predictor of outcomes at a minimal age of 18 months; 80% of infants with moderate or severe outcomes had severe WM injury, whereas no infant with mild WM injury had more than a mildly abnormal outcome. Infants who had severe WM changes with normal/mildly abnormal outcomes all had either unilateral or asymmetrical injuries, which suggests that the relatively preserved hemisphere was able to compensate functionally for the severely damaged side.
Although glucose is the primary fuel for cerebral oxidative metabolism, the normal metabolic adaptation to postnatal life involves the use of lactate31,32 and ketone bodies2 as alternative substrates for oxidative metabolism. Fourteen (40%) of the case subjects had growth restriction and thus were at risk because of reduced substrate stores and less ability to generate and/or to use alternative fuel sources in the event of hypoglycemia. However, the MRI findings were not typical of those seen among infants with growth restriction who do not have acute neurologic dysfunction. Six other subjects (17%) had risk factors associated with either poor feeding, hyperinsulinism, or suboptimal hormone/enzyme responses to hypoglycemia, but 15 subjects (43%) had no known risk factors. We did not find other prenatal, intrapartum, or postpartum factors that could account for the patterns of injury or neurologic symptoms seen among the case subjects. Although there were higher rates of delivery through emergency cesarean section and a tendency to need more resuscitation among the case subjects (although none required cardiac compressions or drugs), the cord pH values, Apgar scores at 1 and 5 minutes, and clinical assessment findings were not consistent with HIE. It is possible that adverse peripartum events rendered the case infants more susceptible to brain injury in the context of hypoglycemia, but there was no evidence that other causal factors accounted for the patterns of injury or the neurologic dysfunction observed.
A limitation of our study was that we did not have detailed early feeding histories or knowledge of maternal medication use, which might have enabled us to identify other case infants who were at risk of metabolic maladaptation. Transient hyperinsulinism has been reported for 3 of 4 infants who had symptomatic hypoglycemia and abnormal MRI findings.33 The suppressed production of alternative cerebral fuels that accompanies hyperinsulinism could render infants with this abnormality more susceptible to injury, and the profiles of hormones and enzymes involved in early glucose regulation in this group of infants warrant further investigation.
There was no relationship between the severity or duration of hypoglycemia and neurodevelopmental outcomes. The data that suggested such associations were from specific groups, such as preterm infants34 and hypoglycemic infants of mothers with diabetes mellitus,35 and from severe, protracted, insulin-induced hypoglycemia in monkeys.36 Our study was not designed to test this relationship; to do so would require larger numbers and more-precise measures of the duration of hypoglycemia.
A significant number of infants had adverse sequelae. Cognitive impairment was more likely than motor impairment to be severe, because most children were walking by 2 years of age; this correlates with the predominant pattern of WM injury, with sparing of the BGT and PLIC. One of the 3 infants with an MCA infarction developed hemiplegic CP, which could be predicted from the site of the infarction,37 whereas those who developed spastic quadriplegia all had severe bilateral WM changes. The mild, PLIC-sparing BGT lesions we found were different from those seen in HIE and did not seem to affect early motor outcomes adversely; however, longer follow-up monitoring is needed to confirm both this finding and the prevalence of subtle cognitive deficits. Although we did not present the data, cognitive impairments and seizures have persisted for children for whom we have later outcome data. Head growth was suboptimal by >2 SD for one third of infants, and only 25% had HC values of >50th percentile in follow-up evaluations, consistent with the high rates of WM injury.38
Visual impairments and epilepsy were common, and both are reported outcomes associated with neonatal hypoglycemia.20,39 Thirty-one percent of our cohort had visual abnormalities at 2 years and this might have been underestimated, because formal detailed visual assessment was not performed for all children. The majority of our infants with abnormal vision did have involvement of the posterior cortex and WM; however, some with severe occipital injury had apparently normal visual development. Twenty-seven percent of our infants developed later seizures. Infants with symptomatic hypoglycemia were more likely to have a family history of seizures. In one case these were neonatal seizures of unknown aetiology in a second degree relative, and in the remaining 5 cases the onset of seizures was in childhood or adulthood in a first degree relative (2 cases) or second degree relative (3 cases).
The patterns of injury associated with symptomatic neonatal hypoglycemia are more diverse than reported previously. Valuable information regarding the neurodevelopmental outcomes of infants with hypoglycemia can be ascertained from brain MRI in the neonatal period, and MRI findings are more instructive than the severity or duration of hypoglycemia as prognostic indicators of later outcomes. Poor cognition and seizures were relatively common in our patients at 2 years, and there is clearly a need for longer-term follow-up monitoring to assess the functional impact of these problems at school age.
This work was supported by Philips Medical Systems (Best, Netherlands), the Medical Research Council, the Academy of Medical Sciences, and the Health Foundation.
We are grateful to the nursing, medical, and radiographic staff members who were involved in scanning, the parents who consented to take part, and the consultant colleagues who referred infants.
- Accepted November 1, 2007.
- Address correspondence to Frances M. Cowan, MRCPCH, PhD, Department of Paediatrics, Hammersmith Hospital, 5th Floor, Ham House, Du Cane Rd, London, England W12 0HS. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject
Symptomatic hypoglycemia in the newborn period is associated with long-term neurodevelopmental impairment.
What This Study Adds
Patterns of injury associated with symptomatic neonatal hypoglycemia were more varied than described previously. Early MRI findings were more instructive than severity or duration of hypoglycemia in predicting neurodevelopmental outcomes.
- ↵Hawdon JM, Ward Platt MP, Aynsley-Green A. Patterns of metabolic adaptation for preterm and term infants in the first neonatal week. Arch Dis Child.1992;67 (4 Spec No):357– 365
- ↵Cornblath M, Hawdon JM, Williams AF, et al. Controversies regarding definition of neonatal hypoglycemia: suggested operational thresholds. Pediatrics.2000;105 (5):1141– 1145
- ↵Anderson JM, Milner RD, Strich SJ. Effects of neonatal hypoglycaemia on the nervous system: a pathological study. J Neurol Neurosurg Psychiatry.1967;30 (4):295– 310
- ↵Larroche JC. Developmental Pathology of the Neonate. New York, NY: Excerpta Medica;1977
- ↵Barkovich AJ, Ali FA, Rowley HA, Bass N. Imaging patterns of neonatal hypoglycemia. AJNR Am J Neuroradiol.1998;19 (3):523– 528
- Spar JA, Lewine JD, Orrison WW Jr. Neonatal hypoglycemia: CT and MR findings. AJNR Am J Neuroradiol.1994;15 (8):1477– 1478
- Alkalay AL, Flores-Sarnat L, Sarnat HB, Moser FG, Simmons CF. Brain imaging findings in neonatal hypoglycemia: case report and review of 23 cases. Clin Pediatr (Phila).2005;44 (9):783– 790
- ↵Mercuri E, Dubowitz L, Brown SP, Cowan F. Incidence of cranial ultrasound abnormalities in apparently well neonates on a postnatal ward: correlation with antenatal and perinatal factors and neurological status. Arch Dis Child Fetal Neonatal Ed.1998;79 (3):F185– F189
- ↵Griffiths R. The Abilities of Young Children: A Comprehensive System of Mental Measurement for the First Eight Years of Life. London, England: Child Development Research Centre;1970
- ↵Wieloch T. Hypoglycemia-induced neuronal damage prevented by an N-methyl-d-aspartate antagonist. Science.1985;230 (4726):681– 683
- ↵Pryds O, Christensen NJ, Friis-Hansen B. Increased cerebral blood flow and plasma epinephrine in hypoglycemic, preterm neonates. Pediatrics.1990;85 (2):172– 176
- ↵Benders MJ, Groenendaal F, Uiterwaal CS, et al. Maternal and infant characteristics associated with perinatal arterial stroke in the preterm infant. Stroke.2007;38 (6):1759– 1765
- ↵Lucas A, Morley R, Cole TJ. Adverse neurodevelopmental outcome of moderate neonatal hypoglycaemia. BMJ.1988;297 (6659):1304– 1308
- ↵Stenninger E, Flink R, Eriksson B, Sahlen C. Long-term neurological dysfunction and neonatal hypoglycaemia after diabetic pregnancy. Arch Dis Child Fetal Neonatal Ed.1998;79 (3):F174– F179
- ↵Brierley JBM, ed. Brain Hypoxia. Philadelphia, PA: Lippincott;1971
- ↵Mercuri E, Ricci D, Cowan FM, et al. Head growth in infants with hypoxic-ischemic encephalopathy: correlation with neonatal magnetic resonance imaging. Pediatrics.2000;106 (2):235– 243
- ↵Traill Z, Squier M, Anslow P. Brain imaging in neonatal hypoglycaemia. Arch Dis Child Fetal Neonatal Ed.1998;79 (2):F145– F147
- Copyright © 2008 by the American Academy of Pediatrics