- Top of page
- CASE HISTORY
Summary: Purpose: To characterize seizure types and electroencephalographic features of glucose transporter type 1 deficiency syndrome (Glut-1 DS).
Methods: Twenty children with clinical and laboratory features of Glut-1 DS were evaluated. Age at seizure diagnosis, seizure classification, and response to treatment were determined by chart review. Thirty-two continuous 24-h EEG monitoring sessions and reports of 42 routine EEG studies were assessed.
Results: Age at seizure diagnosis was between 4 weeks and 18 months (mean, 5 months). Seizure types were generalized tonic or clonic (14), absence (10), partial (nine), myoclonic (six), or astatic (four). During 24-h EEGs, background activity showed generalized 2.5- to 4-Hz spike–wave discharges (41%), generalized slowing or attenuation (34%), no abnormalities (34%), focal epileptiform discharges (13%), or focal slowing or attenuation (9%). No seizures were captured during 69% of the studies; the remainder had absence (19%), myoclonic (9%), or partial seizures (3%). On evaluation of routine and 24-h EEG studies, focal epileptiform discharges (24%) and slowing (11%) were more frequent in ages 0–24 months. In older children (2–8 years), generalized epileptiform discharges (37.5%) and slowing (21%) were more common.
Conclusions: In all ages, a normal interictal EEG was the most common EEG finding. When abnormalities occurred, focal slowing or epileptiform discharges were more prevalent in the infant. In older children (2 years or older), a generalized 2.5- to 4-Hz spike–wave pattern emerged. Seizure types observed included, absence, myoclonic, partial, and astatic.
Glucose transporter type 1 deficiency syndrome (Glut-1 DS) was initially described in 1991 by De Vivo et al. (1) in two children with developmental delay, infantile seizures, and hypoglycorrhachia. Since that time, >80 children have been identified with this disorder. The condition results from a loss of functional glucose transporters, encoded by GLUT-1, that mediate glucose transport across the blood–brain barrier. This defect leads to a reduction in the CSF/blood glucose ratio to approximately one half of normal, and a reduction in CSF lactate (2). As d-glucose is the predominant fuel for brain metabolism, impaired cerebral function results.
Identification of children with this syndrome requires recognition of characteristic clinical features (developmental delay, ataxia, hypotonia, infantile seizures, acquired microcephaly) in association with a reduced CSF glucose concentration. A high level of suspicion should be maintained in individuals with epilepsy in the absence of the classic phenotype, as more subtle forms of Glut-1 DS have been reported (3). The laboratory hallmark is hypoglycorrhachia with an absolute CSF glucose <40 mg/dl. The diagnosis can be confirmed by blood testing, as the Glut-1 transporter protein also mediates glucose transport into erythrocytes (4). The GLUT-1 gene was initially mapped to the short arm of chromosome 1 (1p31.3-p35), and more recently to 1p34.2 (5). Molecular analysis of GLUT-1 in affected individuals has demonstrated a variety of mutations in this region (6,7). In addition, three families with autosomal dominant transmission of Glut-1 DS have been identified (3,8,9).
Although the occurrence of seizures often brings children with Glut-1 DS to medical attention, limited information is available regarding the seizure types or EEG features seen with this syndrome. In 1999, Boles et al. (10) described two affected children with atypical absence seizures and generalized paroxysmal 2- to 2.5-Hz spike–wave discharges on video-EEG monitoring. Von Moers et al. (11) recently studied two children with Glut-1 DS, comparing pre- and postprandial EEG recordings. The fasting background EEG in both children was described as mild to moderately slow, with multifocal and generalized high-amplitude irregular spikes and spike–waves. A significant reduction in epileptiform discharges was noted in the postprandial EEG recordings.
More than a decade after identification of Glut-1 DS, the clinical, EEG, and genetic features continue to be defined and revised. Although it is a rare disorder, increased awareness and testing for Glut-1 DS will likely increase the prevalence of this syndrome. Recognition is important for management of the seizures, as they are poorly controlled by antiepileptic medications (AEDs) (1). Initiation of the ketogenic diet provides an alternative cerebral energy source, leading to improvements in seizure control (2,11). Further characterization of the epilepsy syndrome will better enable the clinician to make an early diagnosis and initiate appropriate treatment.
- Top of page
- CASE HISTORY
This boy, now age 8 years, was recently diagnosed with Glut-1 DS. Pregnancy was term and uncomplicated, with a birth weight of 3,316 g. Family history did not reveal consanguinity or a history of neurologic conditions. Initial developmental milestones were normal, with global developmental delays emerging as he matured. His neurologic symptoms dated back to age 3 months, when his parents observed an exaggerated startle with stiffening of the arms and legs, and fine trembling accompanied by “eyes that would dance around.” These episodes lasted ≤30 s and occurred only when he was supine. They could not be provoked. These startles were initially infrequent, and then they began to occur in clusters several times per week.
At age 5 months, episodes of lateral head swaying developed, with wandering of the eyes “as if unable to focus.” The total duration was ∼5 min, and the boy would respond to his mother's voice by smiling during the event. At this time, he was hospitalized for evaluation. A routine EEG, a brain magnetic resonance imaging (MRI), and a cervical spine MRI were reported as normal. His diagnostic workup included screening for occult neuroblastoma [computed tomography (CT) of chest and abdomen, and 24-h urine collection for catecholamines], serum amino acids, ammonia, very long chain fatty acids, biotinidase, chromosomal analysis, lysosomal screening, urine oligosaccharides, urine organic acids, urine glycosaminoglycans, and a cardiology consultation (ECG, ECHO, and Holter monitor) because of an intermittent, faint systolic murmur. All of these tests were normal except for a mild elevation of ethylmalonate and methylmalonate in urine organic acids, and a mild elevation of branched-chain amino acids with serum amino acids. After an episode of hypoglycemia with an intercurrent gastroenteritis, a fasting glucose (40 mg/dl) and urine testing were performed with an increase in acylcarnitine esters and lethargy at the end of the fast. A skin biopsy was unremarkable.
At age 2 years, episodes of head drops and eyes jerking upward developed, lasting 1–2 s. The events were most prominent when he was hungry or fatigued, with improvement after eating. An EEG showed “bisynchronous 2- to 4-Hz spike–wave discharges,” and ethosuximide (ESM) was started. Topiramate (TPM) was later added because of daily seizures.
By age 7
years, he continued to have frequent episodes of head drops. A lumbar puncture revealed a low glucose of 30 mg/dl (serum glucose, 81 mg/dl) and a low-normal lactic acid of 1.1 mM compatible with a diagnosis of Glut-1 DS. At this time, his examination revealed a head circumference of 50.5 cm/<3% (prior measurement of 44.2 cm/∼25% at 7 months old); weight, 17 kg/<3%; and height, 111 cm/<3%. No abnormalities were found on general physical examination. Neurologically, significant delays were observed in cognition and behavior. He was nonverbal with poor oral coordination. No abnormalities of the cranial nerves or ocular fundi were noted. He had mild dystonic posturing of the limbs, particularly when held in a standing position, and assumed a decorticate limb posture when lying in bed. Stance and ambulation were abnormal and performed with assistance only. Truncal incoordination and impairment of rapid alternating movements were present. A rebound phenomenon was observed on cerebellar testing. Muscle bulk, tone, and strength were decreased. Tendon reflexes were increased (3+), and Babinski signs were present bilaterally.
A continuous EEG was performed at our institution before starting the ketogenic diet at age 8 years. The study showed mild diffuse background slowing, and infrequent generalized spike and polyspike–wave complexes activated by drowsiness and sleep. Multiple clinical seizures were observed with generalized myoclonic jerks and an ictal correlate of a 2- to 2.5-Hz frontal predominant spike–wave discharge without lateralizing features (Fig. 1). After initiation of the ketogenic diet, no further seizures were observed. His parents find that he is more alert, with mild improvements in neurologic symptoms.
Figure 1. Myoclonic seizure. The EEG shows a generalized spike-wave discharge with phase reversal at F3, F4 that is clinically associated with a myoclonic jerk of both arms.
Download figure to PowerPoint
- Top of page
- CASE HISTORY
The epilepsy syndrome associated with Glut-1 DS has not been well described, although seizures remain a prominent clinical feature. Analysis of the CSF/blood glucose ratio has been the primary tool for initial diagnosis of Glut-1 DS. An increased knowledge of the clinical spectrum of Glut-1 DS will assist clinicians in screening and identifying susceptible individuals. Unfortunately, attempts to better understand the seizure types and EEG features with Glut-1 DS are restricted by its low incidence and the relatively recent discovery of this syndrome. For these reasons, several limitations in our study included the small sample size, the reliance on routine EEG studies interpreted at other institutions, and the absence of pre- and postketogenic diet EEG studies in all individuals. Despite these limitations, several clinical and electrographic features were observed.
A normal interictal background remains the most common finding on both routine and prolonged EEG recordings in all age groups. In serial EEG studies in the same individuals, the presence or absence of epileptiform discharges varied. This is observed in individuals with known epilepsy and may reflect an artifact of sampling (12,13). Within our study group, a higher proportion of abnormalities is seen in the 24-h data compared with the routine EEGs. Because of the age difference between the two populations, a direct comparison cannot be made, although a sampling bias may contribute to this observation. In addition, variability of the EEG technical quality and interpretation between institutions may be present. As this syndrome is recognized earlier, it is our hope that more rigorous evaluation of the EEG and natural history in Glut-1 DS will be possible.
During childhood (2 years or older), the most common EEG abnormalities were 2.5- to 4-Hz generalized spike– or polyspike–wave discharges. As this EEG pattern is often a marker of idiopathic generalized epilepsy (IGE), this may lead to a delayed diagnosis of Glut-1 DS in a patient with “IGE” who is not responding to AED treatment. In our study population, both absence and myoclonic seizures were observed during 24-h continuous monitoring. Given this combination of EEG features and seizure types, Glut-1 DS should be considered in the differential diagnosis of childhood absence epilepsy or juvenile myoclonic epilepsy, particularly if atypical features are present.
In infancy (younger than 2 years), generalized spike–wave discharges were less frequent. When abnormalities were present, a trend toward focal slowing or epileptiform discharges was observed. This time point was chosen based on radiologic evidence that myelination approaches that of an adult by 18 months to 2 years (14). This age-related expression of epilepsy has been demonstrated in animal models. Kindling, changes in myelination, alterations in cell properties, and changes in neurotransmission have been proposed as possible explanations (15).
Although the age at seizure diagnosis in our patient group ranged from 4 weeks to 18 months, paroxysmal events, such as behavioral arrest and eye-movement abnormalities, were detected earlier and may have represented subtle seizures. In addition, episodes of transient hemiparesis, erratic eye movements, lethargy, and unsteadiness have been observed in older children. Although these events could be related to seizure activity or a postictal “Todd's phenomenon,” they also may represent distinct manifestations of Glut-1 transporter dysfunction. Similar nonepileptic events have been observed in paroxysmal neurologic disorders of childhood, such as movement disorders, hemiplegic migraines, periodic paralysis, and alternating hemiplegia of childhood. Each of these disorders has been suggested to share similar mechanisms with epilepsy (16–18).
The ketogenic diet remains the best treatment for the seizure disorder and encephalopathy. In addition to providing an alternative energy source for children with Glut-1 DS, the ketogenic diet is thought to have a secondary effect on cerebral glucose transport (19). In 1975, Gjedde and Crone (20) investigated glucose transport into the brains of rats and demonstrated a 5% increase in the brain-uptake index of glucose during starvation. The anticonvulsant effects of the ketogenic diet are well known and have been investigated in animal models (21). Appleton and De Vivo (22) compared electroconvulsive thresholds in rats on a high-fat diet with those on a high-carbohydrate diet and found that the stimulus needed to produce a convulsive response was significantly higher with the animals on the high-fat diet.
Some affected children continue to have seizures and/or developmental delay despite treatment with the ketogenic diet. As the Glut-1 transporter is multifunctional, allowing passage for galactose, water, glycopeptides, and dehydroascorbic acid, deficiencies of other substrates could contribute to the pathophysiology of this syndrome (23). Mechanisms to increase the number or function of the Glut-1 transporters may be a treatment option in the future. Thioctic acid, an antioxidant that translocates Glut-1 from intracellular pools to the plasma membrane in vitro, is being investigated as a therapeutic agent. Barbiturates, such as PB, and methylxanthines, such as caffeine and theophylline, should be avoided as they are known Glut-1 inhibitors (24,25). In the subset of Glut-1 DS patients followed up in our institution, 68% were treated initially with PB (D.C.D., personal observation). As more is learned about the solutes transported through Glut-1 and their effects on cerebral metabolism, additional therapeutic options may become available.
Early identification of children with Glut-1 DS remains important to prevent treatment with AEDs that may be ineffective or potentially detrimental, and to initiate an alternative energy source during a time of increased cerebral metabolism. Other metabolic and genetic etiologies of epilepsy and developmental delay should be included in the differential diagnosis, with the evaluation tailored to the presenting clinical features (26). We recommend that screening no longer be limited to those individuals with the classic phenotype, as milder cases have been described. A diagnostic lumbar puncture and serum glucose level should be considered in all individuals with idiopathic or cryptogenic epilepsy that is not easily controlled with AEDs.
Acknowledgment: We are grateful to the patients and their families for continued participation in these clinical research efforts. We also thank Sarah Jhung, B.S., Veronica Hinton, Ph.D., and Pamela Kranz-Eble, B.S., for assistance and critical comments. This study was supported by the Colleen Giblin Charitable Foundation, the Will Foundation, and USPHS grants NS37949 (D.C.D.) and RR00645 (D.C.D.).