Introduction
Glucose Transporter-1 (GLUT1) Deficiency Syndrome (GLUT1DS) is caused by defective transport of glucose across the blood–brain barrier into brain cells, resulting in hypoglycorrhachia. Reference Almuqbil, Go, Nagy, Pai, Mamak and Mercimek-Mahmutoglu1 The GLUT1 protein is encoded by SLC2A1 (Gene/Locus MIM#138140). Heterozygous pathogenic or likely pathogenic variants in SLC2A1 result in GLUT1DS. In the majority of patients, GLUT1DS is the result of a de novo variant in SLC2A1. Reference Seidner, Alvarez and Yeh2 The phenotypes range from infantile-onset epileptic encephalopathy to adult-onset paroxysmal exercise-induced dyskinesia. The infantile-onset epileptic encephalopathy phenotype is the most commonly reported phenotype (∼90%) and is characterized by global developmental delay (GDD), infantile-onset (1–6 months) drug-resistant epileptic encephalopathy, progressive microcephaly, and movement disorder (e.g. ataxia, spasticity, dystonia, chorea, tremor, myoclonus, and paroxysmal dyskinesia) (MIM#606777). Reference Koch and Weber3,Reference Pons, Collins, Rotstein, Engelstad and De Vivo4 Mild or atypical presentations (∼10%) include paroxysmal movement disorders, cognitive dysfunction, alternating hemiplegia, and epilepsy and those symptoms may be triggered by exercise, fasting, or stress (MIM#612126; MIM#601042).
Low cerebrospinal fluid (CSF) to blood (collected within 30 minutes prior to lumbar puncture) glucose ratio or low CSF glucose levels suggest GLUT1DS. In a systematic review, CSF glucose levels were reported between 0.9 and 2.8 mmol/L and CSF to blood glucose ratio was between 0.19 and 0.59 in patients with GLUT1DS. Reference Leen, Wevers, Kamsteeg, Scheffer, Verbeek and Willemsen5 The ketogenic diet is the gold standard treatment of GLUT1DS. Early diagnosis and initiation of the ketogenic diet will likely improve seizures, neurodevelopmental disorder, and movement disorder. Reference Klepper, Akman and Armeno6
The objective of the present study is to report the phenotypic spectrum of pediatric presentations in GLUT1DS as well as their diagnostic methods from a single center in Canada.
Methods
The Research Ethics Board at the Hospital for Sick Children approved this retrospective cohort study (Approval#1000064305). We included all patients with a molecular genetically confirmed GLUT1DS followed in a single neurometabolic clinic. We reviewed patient charts for clinical features, biochemical investigations, molecular genetic investigations, neuroimaging, treatment modalities, and treatment outcomes.
Results
There were 13 patients. Their phenotypic description and investigations are presented in Table 1. Patients 1, 2, 4, 8, 10 were previously published. Reference Almuqbil, Go, Nagy, Pai, Mamak and Mercimek-Mahmutoglu1,Reference Cordeiro, Bullivant, Siriwardena, Evans, Kobayashi, Cohn and Mercimek-Andrews7–Reference Mercimek-Mahmutoglu, Patel and Cordeiro9 The mean age of symptom onset was 1.6 years (range 1 month–4 years). The mean age of diagnosis was 8 years (range 2–16 years). The mean time between symptom onset and the diagnostic confirmation was 6.4 years (range 0 months–11.5 years). The most common initial symptom was seizure (n = 9). Seven patients had no history of GDD or cognitive dysfunction at the time of symptom onset or initial presentation.
Table 1: Clinical features, EEG, brain MRI, genotypes, and treatments of patients with GLUT1DS
Abbreviations: AbsS = absence seizures;AD(H)D = attention deficit (hyperactivity) disorder; AED = antiepileptic drugs;AS = atonic seizures;ASD = autism spectrum disorder; CA = cerebral atrophy; CLB = clobazam; CMPZ = carbamazepine;CSF = cerebrospinal fluid;ETS = ethosuximide; F = female; FS = focal seizures; FTT = failure to thrive; GTCS = generalized tonic-clonic seizures; GDD = global developmental delay; GS = gelastic seizures; GTS = generalized tonic seizures; HC = head circumference; Htz = heterozygous; ID = intellectual disability; IS = infantile spasms; KD = ketogenic diet; LAC = lacosamide; LGI = low glycemic index; LMT = lamotrigine; LVT = levetiracetam; M = male; m = months; MCDK = multicystic dysplastic kidney; MCT = medium chain triglyceride; MRS: magnetic resonance spectroscopy; MS = myoclonic seizures; N = normal; NA = not available; ND = not done; NP = not performed; OCD = obsessive compulsive disorder; P5P = pyridoxal-5’-phosphate; PHB = phenobarbital; PrevReport = previously reported; PYR = pyridoxine; SNHL = sensorineural hearing loss; Spk = spike(s); TPM = topiramate; VGB = vigabatrin; VPA = valproic acid; VUS = variant of unknown significance; WES = whole exome sequencing; WM = white matter; yrs = years.
* Twin brothers.
Reference ranges: CSF glucose reference range as per our chemistry laboratory = 2.1–3.6 mmol/L equals to 37.8–64.8 mg/dl; CSF glucose reference range in most diagnostic laboratories = ≥3.3 mmol/L equals to 60 mg/dl; Blood glucose reference range 2.8–6.1 mmol/L.
All patients demonstrated various degrees of developmental delay, intellectual disability, or cognitive dysfunction during their follow-up. The most common seizure type was absence seizures (n = 11) (presented between the ages of 6 months and 8 years). Ten patients (77%) presented with movement disorders in their first decade of life. Dystonia and ataxia were the most common movement disorders (each n = 5). Nearly half (46%) of the patients had more than one type of movement disorder. Microcephaly (head circumference <2nd centile) was present in 3 patients (23%). None of the patients had eye–head movement abnormalities during their follow-up in the neurology clinics or in the neurometabolic clinic.
Only two patients had CSF glucose measurement, and both patients had low CSF glucose levels (2.1 mmol/L equals to 37.8 mg/dl), which were below the lower limit of normal CSF glucose of 3.3 mmol/L equals to 60 mg/dl in the post-absorptive state for glucose. None of the patients had blood glucose measurements prior or after the lumbar punctures.
All patients had a pathogenic or likely pathogenic variant in SLC2A1 identified by direct Sanger sequencing and deletion duplication analysis of SLC2A1 (n = 3), next-generation sequencing panel for epilepsy (n = 4), next-generation sequencing panel for hemiplegic migraine (n = 1), positive family history with familial variant testing (n = 1), and whole-exome sequencing (n = 4). Missense variants were the most common variant type (62%). In four patients with asymptomatic parents, the parental segregation was not performed due to the unavailability of the parents.
Twelve patients were started on the ketogenic diet. Detailed treatments and outcomes are summarized in Table 1 and Supplementary Table 1. One patient was not a suitable candidate for the ketogenic diet therapy due to severe vomiting and gastroesophageal reflux disease despite gastro-jejunal tube feeding and the high risk of aspiration. Patient 11 was initially treated with phenobarbital and her seizures were progressed to infantile spasms, which was switched to vigabatrin. Phenobarbital did not improve seizures in Patient 12, which was switched to levetiracetam and resulted in better seizure control.
Discussion
We present 13 patients with GLUT1DS followed in a single neurometabolic clinic at our institution. In our study cohort, absence seizures were the most common seizure type (85%). Movement disorders were present in 77% of the patients in the first decade of life in our study cohort. Interestingly, nine patients had their symptom onset at or before the age of 2 years, but the age of diagnosis was at or less than 3 years of age in only three patients. It seems that despite the application of next-generation sequencing panel for epilepsy or whole-exome sequencing, there are still delays in the diagnosis of GLUT1DS in our study cohort. We report variable phenotypes in presumed monozygotic twins including alternating hemiplegia in one and GDD in another one. Generalized tonic–clonic seizures are the most common seizure type in GLUT1DS Reference Leary, Wang, Nordli, Engelstad and De Vivo10,Reference Pong, Geary, Engelstad, Natarajan, Yang and De Vivo11 , whereas in our study, absence seizures were the most common seizure type.
International consensus guidelines recommend characteristic clinical features (movement disorders; paroxysmal events; early-onset absence epilepsy, mycolonic-atonic epilepsy), AND hypoglycorrhachia AND a pathogenic variant in SLC2A1 for the definite diagnosis of GLUT1DS. Reference Klepper, Akman and Armeno6 If the characteristic clinical features were present, pathogenic variants in SLC2A1 is sufficient for the diagnosis of GLUT1DS. All of our study patients had characteristic clinical features and pathogenic variants in SLC2A1 confirming the diagnosis of GLUT1DS. Despite hypoglycorrhachia is the metabolic hallmark of the GLUT1DS, only two of our patients had low CSF glucose levels to fulfill all three required features. In the remaining 11 patients, the diagnosis was confirmed by characteristic clinical features and pathogenic variants in SLC2A1.
The prevalence of GLUT1DS ranged from 0.9 to 2.4% in our previous studies. Reference Costain, Cordeiro, Matviychuk and Mercimek-Andrews8,Reference Mercimek-Mahmutoglu, Patel and Cordeiro9,Reference Hewson, Brunga and Ojeda12 Prior to 2015, there were two patients diagnosed with GLUT1DS, known to us at our institution. Since the availability of targeted next-generation sequencing panels and whole-exome sequencing in clinical settings, we have confirmed the diagnosis of GLUT1DS in more than 10 patients at our institution. Next-generation sequencing technologies have also expanded the phenotypic spectrum of the GLUT1DS in our clinical practice. In about 10–15% of the patients, targeted next-generation sequencing and whole-exome sequencing will likely not identify variants in SLC2A1, and deletion/duplication analysis should be performed. Reference Klepper, Akman and Armeno6 One of the characteristic clinical features is good response to the ketogenic diet with decreased seizure frequency. Ten patients had the ketogenic diet with improved seizure control.
Ataxia, dystonia, chorea, and paroxysmal dyskinesia may occur individually or in combination. Movement disorders can be permanent or intermittent, often precipitated by infections or prolonged fasting. Reference Pons, Collins, Rotstein, Engelstad and De Vivo4,Reference Wang, Pascual, De Vivo, Adam, Ardinger, Pagon, Wallace, Bean, Stephens and Amemiya13 The spectrum of movement disorders has been reported in 57 patients with GLUT1DS previously. Reference Pons, Collins, Rotstein, Engelstad and De Vivo4 There is wide variability in the age of onset ranging from 3 months to 42 years. Dystonia was the most common movement disorder in 86% of the patients. Gait disturbances, in the form of ataxia or in combination with spasticity, were observed in 70% of the patients. Chorea, tremor, and myoclonus were also reported. Only 28% of the patients had paroxysmal movement disorders; unfortunately, the age of onset for the paroxysmal events was not reported in that study. Reference Pons, Collins, Rotstein, Engelstad and De Vivo4 In our study, we only included pediatric patients (up to 18 years of age) at our center. The frequency of movement disorders was 77%, and ataxia (50% of patients) and dystonia (50% of patients) were the most common movement disorders in our study cohort. Alternating hemiplegia has been reported as one of the clinical features of nonclassical GLUT1DS with a milder phenotype. Reference Wang, Pascual, De Vivo, Adam, Ardinger, Pagon, Wallace, Bean, Stephens and Amemiya13–Reference Weller, Leen and Neville15 Weller and colleagues sequenced SLC2A1 in 42 patients with hemiplegic migraine. Reference Weller, Leen and Neville15 They identified one patient (2.4%) with an overlapping phenotype of atypical hemiplegic migraine and atypical alternating hemiplegia of childhood due to GLUT1DS. Reference Weller, Leen and Neville15 We identified alternating paroxysmal hemiplegia in two patients in our study population; in one of the patients, this was the presenting symptom at the age of 1.5 years. It seems that our study cohort has a higher prevalence and earlier onset of movement disorders.
We present the phenotypic and genotypic spectrum of patients with childhood-onset GLUT1DS from a single neurometabolic clinic. We report absence seizures as the most common seizure type and early onset movement disorders in this pediatric patient population. We also report on the increased diagnostic rate of GLUT1DS since the application of next-generation sequencing in the clinical setting. We propose that GLUT1DS should be considered as an underlying cause for a broad variety of neurological phenotypes and should be included in targeted next-generation sequencing panels for epilepsy, movement disorders, developmental delay, intellectual disability, and atypical migraines. We think that more patients with rarer phenotypic presentations will be diagnosed at an earlier age in the near future, which will likely improve neurodevelopmental outcomes due to early initiation of the ketogenic diet.
Acknowledgments
The authors would like to thank parents and all staff physicians, nurse practitioners, and dieticians including general pediatrics, neurology, and clinical genetics at the Hospital for Sick Children who referred their patients for diagnostic workup or management of GLUT1DS to our clinic. The authors would like to thank Khadine Wiltshire for formatting this manuscript according to the journal’s requirements.
Disclosures
The authors disclose no conflicts of interest.
Statement of Authorship
GAMN, DKB: Reviewed charts, generated the database, drafted the manuscript, read, and approved the final manuscript.
DC: Reviewed charts, generated the database, read, and approved the final manuscript.
JK: Diagnosed some of the patients and applied the ketogenic diet, read, and approved the final manuscript.
SM-A: Planned, designed the study, supervised GAMN, DKB, and DC and revised and finalized the manuscript.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/cjn.2021.3.
