Abstract
Many cases of GAD65 autoimmune encephalitis are described in adults, but this is an understudied topic in children. Not many pediatric cases of GAD65 encephalitis are reported in literature and the diagnostic criteria are not well defined. This is a retrospective observational study of 3 pediatric patients with GAD65 antibodies seen at Arkansas Children's Hospital between 2016 and 2022. GAD65 antibody is a biomarker of autoimmune central nervous system disorder and more commonly non-neurologic disorders like type 1 diabetes mellitus, autoimmune thyroid diseases and pernicious anemia. This case series illustrates the spectrum of symptoms associated with GAD65 in a pediatric population.
Keywords
GAD65 (glutamic acid decarboxylase 65) antibody is a biomarker of autoimmune central nervous system disorders including limbic encephalitis (LE), epilepsy, stiff person syndrome (SPS), cerebellar ataxia, myoclonus, nystagmus, and non-neurologic disorders such as type 1 diabetes mellitus (T1DM), pernicious anemia, and autoimmune thyroid disease.1,2 Patients often present with overlapping symptomatology, so they are denoted as GAD65 antibody spectrum disorders. In our literature review, there were only some small case series and a few isolated case reports on clinical presentations and investigations of GAD65 encephalitis in children. We report 3 such cases with positive anti-GAD65 antibodies. Two of our patients met the clinical criteria for definitive case of autoimmune encephalitis using the Cellucci et al 3 criteria, and one other patient had intractable seizures at the time of presentation. The 2 referenced patients also had a concurrent diagnosis of new onset T1DM. Our other patient had elevated antibodies consistent with antibodies found in Hashimoto thyroiditis, but she remained clinically euthyroid in regard to her symptoms. Here, we present extant literature and discuss our cases.
Methods
This is a retrospective observational study approved by the Institutional Review Board. GAD65-IgG in serum and CSF was measured using a radioimmunoassay (RIA), with results reported in nanomoles per liter. MOG antibodies were tested in all 3 patients and were negative.
Results and Interpretations
Case 1
A 16-year-old girl with a medical history of developmental delay, ventriculomegaly, and a remote history of seizures well-controlled with topiramate for 5 years was hospitalized for breakthrough seizures and altered mental status. Seizures consisted of versive head turning to the left and left side facial twitching with an electroencephalographic (EEG) correlate of right temporal electrographic onset. Patient needed multiple loading doses of antiseizure medications to stop the seizures. Magnetic resonance imaging (MRI) of the brain was unremarkable except for mild enlarged ventricles, with no parenchymal signal abnormalities and no contrast enhancement identified. During the course of hospitalization, patient labs were consistent with diabetic ketoacidosis. She was diagnosed with type 1 DM with ketoacidosis and transferred to the intensive care unit. She was started on insulin infusion and had rapid improvement to euglycemia. Despite this, she required airway protection as her clinical and electrographic seizures worsened.
Cerebrospinal fluid (CSF) analysis showed a normal cell count and glucose, and elevated protein at 133 mg/dL. CSF gram stain and culture was negative. Chest, abdomen, and pelvis computed tomographic scans were negative for any malignancy focus. A consequent comprehensive epilepsy panel was negative for any pathological gene. An autoimmune panel for her serum and CSF was eventually ordered for testing and showed increased levels of anti-GAD antibodies in serum (207 nmol/L) and CSF (24 nmol/L). All the other testable autoantibodies were negative, including MOG antibodies, resulting with the patient being diagnosed with GAD65 antibody-associated autoimmune encephalitis alongside GAD65 antibody-induced type 1 DM. She was then treated with intravenous immunoglobulins (1 /kg/d for 2 days) and intravenous methylprednisolone (1 g/d for 5 days), which was also repeated 1 more time with steroid taper. But because of inadequate clinical response, plasma exchange was started, and clinical improvement noted after 2 cycles. After 5 cycles of plasma exchange, the patient was communicating and ambulatory, with good seizure control. At 3-year follow-up postpresentation, the patient had improved tremendously, doing well in school and remaining seizure free and maintained on antiseizure medications.
Brain MRI at follow-up was normal on both occasions apart from the presence of ventriculomegaly, with no parenchymal signal abnormalities or contrast enhancement identified. Oligoclonal band analysis was not available for this patient.
Case 2
A 6-year-old boy with no significant medical history was admitted with new-onset seizures. Seizure consisting of lip-smacking and up-rolling of the eyes, followed by jerking of all 4 extremities and loss of consciousness. Continuous EEG showed left temporal slowing and brain MRI showed bilateral swelling of the insular cortex. The patient showed symptomatic recovery, so he was discharged home with a follow-up MRI scheduled in 2 weeks.
After discharge, the patient continued to have daily headaches, prompting a repeat MRI, which showed worsening T2 fluid-attenuated inversion recovery (FLAIR) hyperintensities in the hypothalamus, left thalamus, left cerebellar peduncle, and right insular cortex, without associated contrast enhancement. He was consequently admitted and lumbar puncture was performed, showing normal glucose and protein levels alongside a cell count of 7 cells/µL with lymphocytic predominance. CSF paraneoplastic panel, and NMDA antibodies were sent for testing, only for both to come back negative. The patient was then treated with intravenous steroids and was discharged after showing clinical improvement.
Because of recurrence of seizures, the patient was readmitted. Continuous EEG at that time showed numerous seizures from the left hemisphere, necessitating adding multiple antiseizure medications (topiramate, levetiracetam, and lacosamide). The patient had another MRI later that showed progression of insular cortex and hippocampal lesions. During this admission, he had high blood pressure and uncontrolled blood sugars, requiring administration of antihypertensives and insulin. Chest, abdomen and pelvis computed tomographic scans were normal. Cerebral angiography and vasculitis workup was negative. The patient was investigated for infectious, metabolic, mitochondrial, and genetic etiology; all these investigations resulted with negative findings.
Workup for autoimmunity showed a serum GAD65 assay of 0.28 nmol/L (reference value <0.02 nmol/L) and a negative CSF GAD65. This time, the patient was treated with intravenous immunoglobulin (1 g/d for 2 days). Because of minimal response, the patient was given another round of high-dose steroids (1 g/d for 5 days) with a prolonged taper. The patient was also maintained every 4 weeks of intravenous immunoglobulin (IVIG) therapy for 6 months.
Given that this patient was found to have borderline elevation of GAD65 antibodies in the setting of intractable seizures, abnormal MRI findings and extensive negative work up of other etiologies is suggestive of an underlying autoimmune etiology. New-onset type 1 DM also developing during this case supports the diagnosis for GAD65 autoimmune encephalitis. At 2-year follow-up, the patient remains seizure free.
Follow-up imaging demonstrated clear interval improvement, with complete resolution of previously noted T2/FLAIR hyperintensities involving the left amygdala and right insular cortex, and only minimal residual FLAIR signal changes persisting in the left insular region. Oligoclonal bands were positive on 2 separate occasions, with a pattern consistent with CSF-restricted bands (present in the CSF but not matched in the serum), indicating intrathecal IgG synthesis.
Case 3
A 21-year-old girl with a history of autism and headaches was admitted to the hospital for the evaluation of altered mental status. Initial examination showed that the patient was disoriented, agitated, and experiencing visual and auditory hallucinations. There were subacute changes in her speech, with slurring and difficulty finding words. MRI of brain and CSF analysis was within normal limits. The patient was subsequently diagnosed with probable autoimmune encephalitis. IVIG infusions were initiated and continued for two days, resulting in improvement of the mental state change including orientation and improvement of hallucinations. CSF and serum autoimmune panels were also negative at that time.
Four weeks later, the patient was presented with further deterioration of her mental status, behavior, and speech. During this admission, patient was treated with high-dose steroids. Workup this time included serum paraneoplastic antibody testing, showing positive GAD65 and anti-striated antibodies (GAD65 0.07 nmol/L, Str 1:240). Further investigation into anti-TPO and anti-TGO antibodies levels resulted with high titers of 178.9 and 108.99 IU/mL, respectively. Despite high antibody titers, her triiodothyronine (T3), thyroxine (T4), and thyroid-stimulating hormone (TSH) levels and thyroid ultrasonography were unremarkable. The patient was started on rituximab because of severe symptoms and showed significant improvement, almost returning to her baseline in 9 months. Patient anti-inflammatory therapy was held at this time.
A repeat lumbar puncture was performed during the subsequent admission; GAD65 IgG antibodies were detected only in serum, with no CSF positivity on either occasion. Oligoclonal band analysis was not available for this patient.
Because of recurrent episodes of worsened mental status and antibody levels, diagnosis of autoimmune encephalitis was made, although with low GAD65 antibody titer. Patient started monthly pulse steroids and rituximab maintenance. The patient improved clinically over time, with routine laboratory monitoring of her CD19 and CD20 amounts showing levels less than 1%.
Although only low titer serum GAD65 antibodies were detected in this patient, the classification as definite antibody-positive autoimmune encephalitis was supported by paraclinical markers of neuroinflammation on follow-up MRI. Follow-up imaging revealed subtle but definite T2/FLAIR hyperintensities involving the mesial temporal lobes, with involvement of the hippocampal amygdala complex, and mild left-sided predominance findings do not present on the initial scan. In accordance with the Cellucci et al3 criteria, these neuroimaging findings, in the context of compatible clinical features and seropositivity, were considered supportive evidence of CNS inflammation. It must be acknowledged, however, that the low GAD65 titer in this case, without CSF confirmation, places this classification at the boundaries of current diagnostic criteria. The existing literature does not clearly define a titer threshold for neurologic significance in pediatric patients, and the diagnostic weight of low-titer serum-only positivity remains insufficiently characterized in this age group. Future prospective studies with larger pediatric cohorts are needed to clarify the diagnostic significance of low-titer GAD65 seropositivity and to refine classification criteria for this evolving spectrum of disease.
Discussion and Review of Literature
Pathophysiology
Glutamic acid decarboxylase (GAD) is a pyridoxal enzyme that catalyzes the α-decarboxylation of l-glutamic acid, producing the inhibitory neurotransmitter γ-aminobutyric acid. 4 GAD exists in 2 forms according to their molecular sizes: GAD65 and GAD 67. GAD65 is encoded by gene GAD2 on chromosome 10(10p12.1), which is mainly responsible for the rapid synthesis of the GABA required for the synaptic transmission usually expressed in the postnatal stage. 5 GAD65 is a homodimer, and it is more antigenic than GAD 67. GAD65 is predominantly found in the nerve terminal and may be anchored to the vesicular membrane containing the neurotransmitter. In addition to the possible modulatory role during neurotransmission, it also may play a role as a neurodifferentiative agent during development. 6
Even though the pathomechanisms of GAD65-associated neurologic symptoms are debatable, interference with GABA synthesis and impairing the inhibitory GABAergic circuits that eventually result in a hyperexcitable state of the central nervous system is a commonly accepted mechanism so far. Proposed mechanisms of GAD autoimmunity include the following:
Antibodies against GAD interfere with the enzymatic action of GAD that eventually decreases the levels of inhibitory neurotransmitter GABA, resulting in increased excitatory activity. However, GAD is an intracellular enzyme, and it is not very clear how the antibodies gain access to this protein. Studies by Towns et al
7
and Jin et al
8
showed that palmitoylated GAD65 can associate with the plasma membrane and is able to surge to the extracellular space. Also, when the synaptic vesicle fuses with the plasma membrane during exocytosis, the intracellular portion of GAD65 might be transiently uncovered in the extracellular space.
5
Studies by Manto et al
9
described how in vitro administration of IgG from a patient with GAD antibodies into the lumbar para-spinal region showed effects that include continuous motor activity with repetitive discharges and abnormal exteroceptive reflexes, increased anterior horn cells excitability, and decreased corticomotor response to stimuli with intracerebellar administration of IgG. Such findings support the likely direct antigenic effect of GAD65 antibodies. Functional alterations in GABAergic transmission are also implicated in pathogenesis. According to Mitoma et al,
10
there are studies that have shown depressed GABA release in cerebellar slices after CSF IgGs obtained from patients with anti-GAD65 antibody–associated cerebellar ataxia are administered in vitro. Intracerebellar administration of these IgGs elicits deficits in cerebellar control of the motor cortex in vivo, likely because of the suppression of GABA release discovered from in vitro studies. Specifically, binding of anti-GAD65 antibody b78 to GAD65 interferes with the association of GAD65 with the cytosolic face of GABA-containing synaptic vesicles, which results in impairment of GABA packaging into the vesicles and shuttling of vesicles to the release site on the synaptic cleft. Role of T cell response mechanism: studies by Lancaster and Dalmau
2
showed disorders associated with intracellular antigens such as GAD65 and amphiphysin, both T cells and autoantibodies are implicated in the pathogenesis. Few in vitro and in vivo studies showed that GAD autoimmunity might be mostly cell mediated. Dade et al
5
found that inflammatory CD4+ T cells are found in the CSF of patients with stiff-person syndrome while inflammatory CD8+ T cells are found in patients with lethal encephalomyelitis along with pathologic evidence of the presence of inflammatory T cells present in the hippocampus with severe neuronal loss. Bien et al
1
demonstrated hippocampal specimens have T-cell infiltrates and that 's more consistent with the hypothesis that GAD antibodies are bystander biomarkers secondary to central nervous system inflammation. Epitope specificity of GAD65 antibody is also implicated in the heterogeneity of the neurologic symptoms of GAD65. Studies described by Li et al
11
have shown antibodies are closely clustered in the C terminal flexible region segregated into 2 clusters located on the opposite faces of the carboxy-terminal domain. Patients with neurologic disorders have been shown to specifically recognize the b78 epitope specificity, which is inhibitory, whereas T1DM patients recognize the b96.11 epitope, which is non–enzyme inhibitory. However, the specific epitope recognition was not always found among various symptoms, and there is no clear clinical significance so far. Anti-GAD65 antibodies may coexist with other additional cell surface antibodies such as anti-VGKC, anti-NMDA-R, anti-GABA A, or GABA B receptors.12,13 These antibodies can be directly pathogenic to the cell by direct receptor competition and internalization. In such a scenario, presence of GAD65 antibody may simply be a bystander effect generated by blood-brain barrier permeability.
Clinical Features
GAD65 Antibody-Associated Neurologic Syndromes
GAD65 antibodies are associated with stiff-person syndrome (SPS), cerebellar ataxia, limbic encephalitis (LE), epilepsy, and patients with overlapping features.
GAD65-Associated Encephalitis
For diagnostic purposes, we used the classification criteria defined by Cellucci et al 3 published in 2020. Cellucci et al classified pediatric autoimmune encephalitis (AE) into 3 subgroups including possible, definitive antibody-positive, and probable antibody-negative pediatric AE. GAD encephalitis is a subtype of AE, and it comprises a large spectrum with various neurologic and psychiatric symptoms.
Our first and third patients have been diagnosed as definitive antibody-positive AE. The diagnostic basis for each classification is detailed within the respective case descriptions, including the supporting neuroimaging and paraclinical evidence as per Cellucci et al criteria. Pediatric anti-GAD65 antibody encephalitis is rare and includes limbic and extralimbic encephalitis. 11 The most common clinical manifestations with GAD65-associated AE are seizures, cognitive impairment, executive dysfunction, and psychiatric disorders. According to studies described by Baizabal-Carvallo, 14 compared to patients with limbic encephalitis associated with other cell surface antibodies such as anti-VGKC or anti-NMDA-R, patients with anti-GAD65 antibodies were younger and presented with seizures rather than cognitive symptoms. The presence of oligoclonal bands in the CSF along with the intrathecal synthesis of anti-GAD65 antibodies shows the unremitting chronic disease course with persistent antibodies. In our series, case 2 demonstrated CSF-restricted oligoclonal bands on 2 separate occasions, consistent with intrathecal IgG synthesis; oligoclonal band data were unavailable for cases 1 and 3. Our second patient has positive oligoclonal bands, and the patient's seizures are not very well controlled.
Follow-up neuroimaging across our 3 cases showed distinct patterns: in case 1, brain MRI remained normal on both occasions apart from stable ventriculomegaly, with no parenchymal signal abnormalities or enhancement. In case 2, follow-up imaging showed clear interval improvement with complete resolution of the T2/FLAIR hyperintensities involving the left amygdala and right insular cortex, and only minimal residual FLAIR signal changes in the left insular region. In case 3, the initial MRI was normal; subsequent imaging revealed subtle but definite T2/FLAIR hyperintensities involving the mesial temporal lobes, with involvement of the hippocampal-amygdala complex and mild left-sided predominance—findings not identified on the initial study.
A literature review of patients with GAD65-associated limbic encephalitis and epilepsy by Gagnon et al 12 showed that 78% of patients had an abnormal MRI. A very small subset of patients in the study had normal MRIs initially. One of our patients presented with a normal MRI initially, and another of our patients had no new MRI abnormalities found on workup that correlated with presenting symptoms. The most common MRI findings in GAD65-associated autoimmune encephalitis are disproportionate parenchymal atrophy for age, as well as cortical and subcortical T2 hyperintensities. Few subsets of patients have hippocampal abnormalities. 15 The presence of refractory epilepsy with the above-mentioned findings should alert the clinician to test for GAD65 antibodies. One of our patients fits this scenario, for example, where T2 FLAIR hyperintensities were discovered 2 weeks after presenting with new-onset seizures, specifically in the hypothalamus, left thalamus, and left cerebellar peduncle. This patient even had progression of brain abnormalities after initial workup and treatment, showing progression of insular cortex and hippocampal lesions when seizures recurred. Fluorodeoxyglucose positron emission tomography (FDG-PET) may show hypermetabolism corresponding to the same region as an MRI lesion, and diffusion tensor imaging (DTI) tractography may show widespread white matter damage already at the early stage. In our patients, FDG-PET and DTI are not done. Initial MRI can be normal, and if clinical suspicion is present, repeat imaging is warranted, which may often reveal abnormalities.
Volumetric and shape analysis of the hippocampus in temporal lobe epilepsy presenting in patients with GAD65 antibodies by Blanco et al 16 showed that atrophy and volume loss can occur over the course of years. This is only demonstrated by vertex-wise shape analysis methodology. Such findings suggests that it is ideal to treat the patient in the window before atrophy is established in efforts to prevent irreversible damage. Discussed by Harmon et al 17 in a cohort study and literature review concerning pediatric anti-GAD65 autoimmune encephalitis, the linkage is acknowledged between permanent hippocampal injury and reduced improvement in memory following symptom onset in adults. In Harmon et al's cohort of 10 pediatric patients with GAD65 encephalitis, however, it is shown that significant improvement in memory and cognition can be attained, especially when treatment is initiated early in the disease process.
It is important to acknowledge the diagnostic overlap between GAD65 antibody-associated encephalitis and GAD65-associated epilepsy. In cases 1 and 2, the clinical presentation was predominantly seizure-predominant without clear or persistent encephalopathic features at onset. These findings highlight the recognized heterogeneity within the GAD65-associated autoimmune neurologic spectrum, where seizures may occur with or without overt encephalopathy. Systemic autoimmune features, variable neuroimaging evolution, and response to immunotherapy in our patients support an immune-mediated process, and we interpret these cases within the broader spectrum of GAD65-associated neurologic autoimmunity.
Seizures
GAD associated encephalitis is known to cause both acute/subacute seizures and chronic epilepsy. A review by Daif et al 18 describes this in detail, showing that anti-GAD65 antibodies were detected in 1.7% to 12.5% of unexplained adult-onset epilepsy and more than 6% of refractory childhood epilepsy, with adult-onset epilepsy frequently presenting with high antibody titers and temporal lobe onset. Other studies by Veri et al 19 and Tekturk et al 20 showed that high anti-GAD65 antibodies have been detected in 0.5% to 0.8% of pediatric epilepsy and about 2% of patients with cryptogenic epileptic encephalopathies.
In comparison to our literature review findings, both our first and second patients had abnormal findings on EEG in the presence of seizurelike activity, whereas our third patient had visual and auditory hallucinations with no specific findings on EEG. Our first patient showed right temporal electrographic onset initially, and multifocal epileptiform discharges. Our second patient also showed left temporal involvement on EEG (left temporal slowing with swelling of insular cortex on MRI) on initial presentation with new-onset seizures, but when seizures recurred, left hemisphere seizures were noted. This patient also had low GAD65 serum levels noted when refractory seizures presented. When acknowledging such a small percentage of patients with pediatric epilepsy also having high anti-GAD65 antibodies in Veri et al’s 19 and Tekturk et al’s 20 studies, higher levels of these antibodies may not necessarily have to be detected when diagnosing GAD-associated encephalitis in a child with a disease course similar to our patient.
The interpretation of serum GAD65 antibody titers requires careful clinical contextualization. Low or borderline serum GAD65 levels in isolation—particularly in the absence of CSF positivity—carry limited diagnostic specificity and may reflect non-neurologic autoimmune conditions or incidental seropositivity. Higher antibody titers are more consistently associated with neurologic syndromes, although even this relationship is not absolute. Two of our 3 patients had low or borderline serum titers without CSF positivity, underscoring the importance of integrating antibody results with the full clinical picture. CSF analysis is particularly valuable, as evidence of intrathecal antibody synthesis or neuroinflammation—such as CSF pleocytosis exceeding 5 cells/µL per Cellucci et al3 criteria—provides important supportive diagnostic data. Antibody levels should therefore never be interpreted in isolation, but in conjunction with clinical phenotype, neuroimaging, CSF findings, and treatment response.
In Daif et al's 18 review, two predominant clinical scenarios were observed, one of which consisted of patients with acute/subacute seizures (some being new onset refractory status epilepticus) either with or without psychiatric symptoms. These presentations showed association with MRI evidence of mesiotemporal inflammation consistent with limbic encephalitis. The other scenario is chronic epilepsy with or without MRI evidence of inflammation alongside some rare personality changes. Other seizure types observed in their review of patients with detected anti-GAD65 antibodies were myoclonic seizures, palatal myoclonus, and idiopathic generalized seizures.
In drug-resistant patients, a trial of immunotherapy is warranted. GAD65 antibody–mediated epilepsy is only moderately responsive to antiepileptics and steroids, and the patients often need IVIG, plasma exchange or immunosuppressants such as rituximab or cyclophosphamide. 18 This has proven to be a successful method of action in our patients; our two patients predominantly presenting with seizures have been seizure-free for 3 and 2 years post-immunotherapy treatment respectively.
Recent studies by Falip et al 21 have shown that music-induced reflex seizures can be a distinctive type of GAD65 epilepsy. It is described in adult patients who have temporal epileptogenic foci. Avoiding the provocative triggers is first-line treatment, but if the seizures are not controlled, antiepileptics are the second-line option. At that time, the patient also needs an autoimmune workup with anti-GAD antibody determination.
Pediatric Stiff Person Syndrome
Pediatric stiff person syndrome (SPS) is rare and often underdiagnosed. Clardy et al 22 showed that the median age of onset of symptoms was 11 years, and patients often have coexisting autoimmune disorders including T1DM, thyroiditis, and vitiligo. Patients usually have positive serum GAD65 antibodies and often respond well to symptomatic therapy as well as immunotherapy. The same study also showed that a subset of patients tested positive for antibodies against glycine receptors. Differential diagnoses including inherited hyperekplexia, inherited dystonia, and hereditary spastic paraplegia need to be considered when encountering patients with SPS phenotype.
Pediatric Anti-GAD65–Associated Cerebellar Ataxia
The retrospective study by Arino et al 23 showed the median age of onset was 58 years, and 82% of these patients were women. Patients typically present with episodes of brainstem and cerebellar dysfunction and may have other symptoms falling within the spectrum of SPS or LE. Autoimmune cerebellar ataxia due to GAD65 antibodies is rarely reported in children, but a few case reports are published. Testing for GAD65 antibody should be considered whenever patients present with cerebellar ataxia and in association with new-onset T1DM or other endocrine or autoimmune diseases. 24 No longitudinal studies are conducted in patients with pediatric SPS because of the rarity of the disease.
Coexisting Autoimmune Diseases and Antibodies
Patients with GAD65 disease often coexist with 1 or more systemic autoimmune diseases including type 1 diabetes mellitus, thyroiditis, immune deficiency syndromes, and several others. Two of our patients have a concomitant diagnosis of T1DM, presenting in active DKA. Patient 3 had elevated TPO and TGO antibodies with normal TSH and T4. Other antibodies including anti-gliadin, antiphospholipid antibodies, and antinuclear antibodies are noted in several patients with anti-GAD65 encephalitis with no overt clinical symptoms. Symptoms of other autoimmune diseases may start several years before the onset of neurologic symptoms. Cases of long latency between GAD detection and development of neurologic symptoms are reported in the literature. 25 Low titers of GAD65 antibodies have been detected in 80% of patients with T1DM and in 4% of normal subjects. Recent studies showed that 30% of GAD spectrum diseases have T1DM and only patients with typical GAD-SD neurologic syndromes exhibit high titers. Lin et al 26 reported that the presence of low-titer GAD antibodies was up to 35.3% in patients with encephalitis and childhood status epilepticus. Data on the diagnostic value of anti-GAD antibody titers and phenotype manifestation remain unclear in children. The non-neurologic manifestations in association with anti-GAD65 is wide, and further discussion is beyond the scope of this article.
Our third patient had elevated TPO and TGO antibodies with normal TSH and no evidence of thyroid inflammation in the ultrasound. There was a concern for possible Hashimoto encephalitis since the GAD65 titers were low and absent in the CSF. Diagnosis of Hashimoto encephalitis requires the presence of thyroid autoantibodies in serum as well as the exclusion of other possible etiologies of encephalopathy, and patients usually respond well to corticosteroid therapy. 27
The same patient did not respond well to the steroids and IVIG and only had a partial response to plasmapheresis. Because of recurrent episodes, patient was started on monthly pulse steroids and rituximab maintenance, which resulted in significant clinical improvement. Because of the initial presence of antibodies for both GAD65 autoimmunity and Hashimoto, it is difficult to precisely identify the etiology. This case raises the question of whether an association exists between anti-GAD65 antibodies (seen in autoimmune encephalitis) and CNS effects of TPO and TGO antibodies, independent of Hashimoto encephalopathy.
Paraneoplastic Significance
Anti-GAD65 antibody is not often associated with paraneoplastic forms. 18 In one of the large retrospective studies by Budhram et al, 28 patients with high titers of GAD65 antibodies and neurologic manifestations showed only 9% had cancer diagnosis within 5 years of symptoms onset. Cancers diagnosed includes thyroid cancer, lung cancer, breast cancer, and thymoma. In another large Norwegian study with 3168 adult patients who had GAD65 antibodies present, there was found to be no increase in the likelihood of cancer diagnosis in patients with GAD65 antibodies. 29 By multiple logistic regression analyses (adjusted for age and sex), there was no higher odds for cancer in any of the groups compared. According to Ariño et al, 30 adult studies showed the presence of GAD65, in addition to other neuronal cell surface antibodies, increases the risk of malignancy by 7-fold. Other studies reviewed by Ariño et al, paired with their own experience, had findings that overall indicate that paraneoplastic GAD autoimmunity is infrequent, but this possibility cannot be overlooked, especially since there are no large pediatric studies and data are unavailable. In our literature review, we could not find such studies in the pediatric population or any other diagnostic criteria established. In our own case series, investigation for cancer screening at the time of admission to the hospital was unremarkable.
Management
There are no clear guidelines for the management of pediatric GAD65 encephalitis and seizures because there are no prospective clinical trials. All the available data are from the case reports and case series. The overall goal is to reduce the immune response and enhance GABAergic activity. Most of the patients need a range of immunosuppressive and immune modulators in addition to antiepileptics. From the literature review and from our experience, GAD65-associated epilepsy is difficult to treat and often refractory, needing more aggressive immunotherapy. Even epilepsy surgeries are described in the adult patients with predominately hippocampal sclerosis, and in the recent retrospective multicentric study consisting of 8 patients, 2 patients had achieved Engel II outcome and 3/5 achieved Engel III outcome. 31 However, no such studies are done in the pediatric population. One case report did describe a patient with high-titer GAD65 antibodies (along with low-titer VGKC antibodies) undergoing temporal lobe epilepsy surgery before a diagnosis of limbic encephalitis was made, but the reported outcome was not very promising, resulting with the patient still having memory impairment and epilepsy. 32
In our experience, early initiation of plasmapheresis and immunosuppression with rituximab—if high-dose steroids (30 mg/kg methylprednisolone intravenously, up to 1 g daily for 5 days, followed by initiation of 1 g/kg—maximum 60 mg prednisolone and taper over 6 weeks) and/or IVIG (2 g/kg dose over 2 days, followed by same dose every 4 weeks for 6 months) response was suboptimal—helped to reduce the seizures, as well as the cognitive and psychiatric complications. Rituximab was administered using the standard neuroimmune induction protocol of 375 mg/m2 weekly for 4 consecutive weeks (maximum dose 1 g), repeated after 1 month and thereafter every 6 months for 2 years. Multicentric controlled prospective studies in the pediatric population would be more beneficial to formulate the regimen in this patient population.
Conclusion
This case series illustrates the spectrum of symptoms associated with GAD65 in pediatric and young adult population. Two of our patients met the criteria for definitive case of autoimmune encephalitis using the Cellucci et al 3 criteria, and our other patient had intractable seizure at the time of presentation. Two of our patients also had a concurrent diagnosis of new-onset type 1 diabetes mellitus whereas our other patient had elevated antibodies consistent with Hashimoto thyroiditis while remaining clinically euthyroid. The same third patient was positive for anti-striated antibodies but did not have symptoms of myasthenia gravis or thymoma.
A myriad of MRI, seizure, and cognitive findings were presented in our patients that both matched and differed from what was discovered in a literature review. Initial MRI findings were normal or stable in 2 of our patients, whereas 1 patient showed progression of MRI abnormalities over the course of disease. This patient showed T2 FLAIR hyperintensities and hippocampal abnormalities, similar to what was described in other studies, but abnormalities were also discovered in the hypothalamus, left thalamus, left cerebellar peduncle, and right insular cortex. Two of our patients had seizure activity involving temporal lobes, falling in line with what was described in the literature review studies. These patients eventually required immunotherapy to ultimately manage these seizures, as multiple antiepileptic medications failed to halt seizures or stop recurrence. All 3 patients required some form of immunotherapy involving steroids, IVIG infusions, rituximab, or plasma exchange to achieve clinical improvement and resolution of GAD65 encephalitis symptoms, including our 21-year-old patient presenting with episodes of altered mental status as well as auditory and visual hallucinations.
It remains to be investigated and conclusively proven whether GAD65 encephalitis along with another autoimmune disease represent the spectrum of autoimmunity rather than separate pathologies. These topics may be the subject of further research in the future.
Footnotes
Author Contributions
All authors contributed to the conception and design of the study, data acquisition, analysis and interpretation, drafting and critical revision of the manuscript, approved the final version, and agree to be accountable for all aspects of the work. All authors meet the International Committee of Medical Journal Editors (ICJME) criteria for authorship.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.