Potentially pathogenic autoantibodies are found increasingly in adults with seizure disorders, including focal seizures and those of unknown cause. In this study, we investigated a cohort of children with new-onset seizures to see whether there were autoantibodies and the relationship to any specific seizure or epilepsy type.
We prospectively recruited 114 children (2 months to 16 years) with new-onset seizures presenting between September 2009 and November 2011, as well as 65 controls. Patients were clinically assessed and classified according to the new International League Against Epilepsy (ILAE) organization of seizures and epilepsies classification system. Sera were tested for autoantibodies to a range of antigens, blind to the clinical and classification details.
Eleven (9.7%) of 114 patients were positive for one or more autoantibodies compared to 3 of 65 controls (4.6%, p = ns). Patients had antibodies to the voltage-gated potassium channel (VGKC) complex (n = 4), contactin-associated protein-like 2 (CASPR2) (n = 3), N-methyl-d-aspartate receptors (NMDARs) (n = 2), or VGKC-complex and NMDAR (n = 2). None had antibodies to glutamic acid decarboxylase, contactin-2, or to glycine, 2-amino-3-(3-hydroxy-5-methyl-4-isoxazolyl) propionic acid (AMPA), or γ-aminobutyric acid B receptors. Ten of these 11 patients were classified as having epilepsy according to the new ILAE organization of seizures and epilepsy. Although, there were no significant differences in the demographic and clinical features between antibody-positive and antibody-negative patients, the classification of “unknown cause” was higher in the antibody positive (7/10; 70%) compared with the antibody negative subjects (23/86; 26.7%; p = 0.0095, Fisher's exact test). Furthermore, four of these seven patients with epilepsy (57.1%) were classified as having predominantly focal seizures compared with 12 of the 86 antibody-negative patients (13.9%; p = 0.015).
Because autoantibodies were more frequent in pediatric patients with new-onset epilepsy of “unknown cause,” often with focal epilepsy features, this group of children may benefit most from autoantibody screening and consideration of immune therapy.
A clear clinical diagnosis of patients who present with seizures and the subsequent accurate classification of their epilepsies is extremely important to the neurologist for treatment and prognosis and to the neuroscientist for facilitation of focused, robust research. To meet these dual needs, the revised International League Against Epilepsy (ILAE) system for the organization of seizures and epilepsies, published in 2010, presented a conceptualized framework for everyday clinical practice that also tried to reflect the advancement of basic epilepsy research (Berg et al., 2010). The most important and subsequent controversial recommendation, involved the replacement of the terms “idiopathic, symptomatic, and cryptogenic” with “genetic, structural/metabolic, and unknown cause” when describing etiology (Shinnar, 2010; Panayiotopoulos, 2011, 2012).
It is estimated that epilepsies of unknown cause account for approximately one third of all cases of epilepsy in adults (Berg et al., 2010) and 23–35% in children (van Campen et al., 2013). They are an important group for the identification of new etiologies, and an area in which the emerging identification of specific antibodies may be increasingly important. Antibodies (Abs) directed against neuronal surface proteins such as the voltage-gated potassium channel (VGKC) complex and its associated proteins, leucine-rich, glioma inactivated 1 (LGI1), contactin-associated protein-like 2 (CASPR2) and contactin-2, the N-methyl-d-aspartate receptor (NMDAR), γ-aminobutyric acid B receptor (GABABR), and 2-amino-3-(3-hydroxy-5-methyl-isoxazol-4-yl) propanoic acid receptor (AMPAR) and against intracellular proteins such as glutamic acid decarboxylase (GAD) have already been described. The association between these antibodies and seizures is manifest in patients with limbic encephalitis who often have temporal lobe seizures, and NMDAR-Ab encephalitis in which patients can have focal or generalized seizures. In addition, each of the antibodies has been identified in a proportion of patients with epilepsy but without other encephalitic clinical signs or symptoms, such as cognitive impairment or neuropsychiatric features (Peltola et al., 2000; McKnight et al., 2005; Majoie et al., 2006; Irani et al., 2008; Niehusmann et al., 2009; Barajas et al., 2010; Quek et al., 2012; Brenner et al., 2013), prompting the increased recognition of “autoimmune epilepsy” (Palace & Lang, 2000; Bien & Scheffer, 2011; Irani et al., 2011a; Vincent et al., 2011; Nabbout, 2012; Quek et al., 2012). The identification of a putative autoimmune etiology in some forms of epilepsies suggests that early identification of specific autoantibodies and subsequent treatment with immunotherapy could lead to reduced seizure frequency and may improve outcomes (Irani et al., 2011b; Quek et al., 2012).
In children, NMDAR-Ab encephalitis and VGKC-complex-Ab–associated encephalitis have also been recognized, with seizures being an important presenting symptom (Dale et al., 2009b; Florance et al., 2009; Haberlandt et al., 2011; Suleiman et al., 2011a). Fewer reports are available regarding the association of neuronal antibodies in defined pediatric epilepsies (e.g., Dhamija et al., 2011; Suleiman et al., 2011b, 2013). Here, we recruited a prospective cohort of children with new-onset seizures, recorded their clinical features, and classified them according to the revised ILAE organization of seizures and epilepsies. At the end of the study period, and blinded to the clinical and classification details, the sera taken at admission were tested for autoantibodies.
Recruitment criteria of patients with new-onset seizures
We prospectively recruited children aged 2 months to 16 years with new-onset seizures presenting to the Children's Hospital at Westmead Hospital, Sydney between September 2009 and November 2011. One hundred fourteen children were included in the study (58 females, mean age 4.39 years, median 2.55 years, range 0.13–15.25 years). All patients with a well-described paroxysmal episode compatible with a clinical seizure and who had serum collected within 6 months of seizure onset were included (in and outpatients). Neonates (0–60 days), and patients with clear etiologies for their seizures such as brain tumor, bacterial meningitis, and brain injury (traumatic or ischemic) were excluded from the study.
The control group for antibody testing consisted of hospital patients (n = 65, 27 female, mean age 9 years, median 9.14 years, range 1–16 years) who had serum collected as part of their routine investigations during 2007. The underlying medical problems were immunologic/inflammatory or allergic (n = 28); hematologic/oncologic (n = 20); or other medical condition such as respiratory, gastrointestinal, and endocrine (n = 17). Recruitment of the patients and control groups for this study was approved by the hospital ethics committee, and written consent was obtained from the patients or their families.
Patients were tested for antibodies to VGKC complex and NMDAR (n = 114); LGI1, CASPR2, and GAD (n = 113); glycine receptor (GlyR), AMPAR, contactin-2 (n = 112); and GABABR (n = 92). The antibodies to VGKC complex and GAD were tested by radioimmunoassays as described previously (McKnight et al., 2005). VGKC-complex-Abs and GAD levels >100 pM and 100 >U/ml, respectively, were considered positive. All other neuronal antibodies were tested by cell-based assays (CBAs) (for full descriptions see Irani et al., 2008, 2010a,b; Dale et al., 2009b). The CBAs were scored on a visual scale 0 (no binding), 1 (low but specific binding)-4 (strong binding to all transfected cells) by two independent observers. If positive, samples were titrated by serial dilution, and the final dilution at which the sample remained positive (score of 1) was given. The 65 controls were tested at the same time as the new-onset seizure patients for the following antibodies: LGI1, CASPR2, contactin-2, NMDAR, AMPAR, GABABR, GAD, and GlyR (n = 65) and VGKC complex (n = 62). Assays were performed and read by SW who was blinded to the clinical data, and re-scored by an additional reader. Incomplete testing was due to insufficient sample available. Cerebrospinal fluid testing could not be performed in this study owing to limited sample availability.
Classifications of seizure types, epilepsy, electroclinical syndrome, and etiology were performed by the primary investigator, a pediatric neurologist (JS), and were cross-checked by a pediatric neurologist and epileptologist (DG). Both were blinded to the results of the antibody testing at the time of performing the classification. The seizure type at presentation was determined for the 114 patients and classified where possible using the new ILAE organization of seizures and epilepsies (Engel, 2006; Berg et al., 2010) and terminology for ictal semiology (Blume et al., 2001). Classification of a patient's condition into an epilepsy syndrome was made by reviewing the child's age, neurologic and developmental state, seizure semiology (at onset and on follow up), electroencephalography (EEG), imaging, and other relevant investigations findings. The structure of the latest proposed ILAE organization of seizures and epilepsies (Berg et al., 2010) was used for this purpose in addition to previous ILAE classification when required (including Commission, 1989; Engel, 2001; Engel, 2006). Patients who had features compatible with one of the electroclinical syndromes described in the new ILAE organization of seizures and epilepsies were classified as such irrespective of the etiology of their epilepsy. For example, patients with epileptic spasms and hypsarrhythmic EEG were classified as West syndrome, despite some of these patients having a known structural or metabolic cause. Patients who did not fit into any electroclinical syndrome were classified (as per the ILAE 2010 recommendation) on the basis of presence of metabolic diseases or structural abnormalities. The term “epilepsy of unknown cause” was used to classify those patients who did not fit into an electroclinical syndrome and who had no structural or metabolic cause found for their epilepsy. Patients with seizures that are traditionally not diagnosed as a form of epilepsy including febrile seizures were kept in a separate group as per Berg et al. (2010). We also added patients with acute symptomatic (provoked) seizures (Beghi et al., 2010; Beleza, 2012) and with single unprovoked seizures to this group.
Statistics and data analysis
Fisher's exact test was used to compare categorical data and the Mann-WhitneyU test to compare continuous data variables.
Demographic, clinical features, investigations, and treatment of total cohort
One hundred one of the 114 patients in the study were admitted to the hospital during their initial presentation or during the first 6 months after seizure onset (median length of stay 4.5 days, range 1–360 days). Eighteen patients required admission to intensive care. Thirty-six patients had one or more preexisting developmental, motor, or psychiatric abnormalities including 33 with developmental delay and/or learning difficulty, 14 with motor deficits, and 4 with behavioral or psychiatric disturbance. There was a positive family history of seizures or epilepsy (including febrile seizures) in the first-degree relatives of 19 patients (16.7%).
Nearly half of the patients (n = 56) presented with focal seizures (Table S1) including 12 with focal dyscognitive seizures. Twenty-one patients presented in status epilepticus and 77 patients had an early recurrence of seizures (occurring within 48 h of first seizure onset). At the time of presentation, 32 patients had intercurrent infections including 21 with fever (defined as 37.8°C or higher [axillary temperature]), 7 of whom fulfilled the definition of febrile seizures (Berg et al., 1992). Seizures were associated with other features in 48 patients including encephalopathy (n = 33), behavioral or psychiatric alteration (n = 27), motor impairment (n = 17), cognitive alteration (n = 12), and movement disorder (n = 5).
EEG findings were abnormal in 86 (76.1%) of 113 patients, and contributed significantly to the epilepsy diagnosis and classification in 22 patients (19.5%). Brain magnetic resonance imaging (MRI) scans were abnormal in 52 of 105 patients and were diagnostic or suggestive of a cause of the epilepsy in 30 patients (28.6%). CSF analysis for pleocytosis, protein, neopterin, or infections was performed in 66 (57.9%) of 114 patients and contributed to epilepsy diagnosis in 13 patients (19.7%). Other investigations were performed as clinically indicated including metabolic, genetic, infectious, and immunologic investigations, and contributed to the epilepsy diagnosis in six patients.
Forty-six patients received acute seizure treatment and 85 patients received long-term (7 days or longer) antiepileptic drugs (AEDs). Twenty-three patients received variable regimens of immunotherapy; including steroids alone (oral prednisolone, intramuscular synthetic adrenocorticotropic hormone [ACTH] and intravenous methyl prednisolone) in 18 and steroids and intravenous immunoglobulins (IVIGs) in 5. In all cases the decision to treat with immunotherapy was made on the basis of a clinical presentation suggestive of an immune-mediated cause (e.g., Rasmussen encephalitis); a recognized immunotherapy responsive seizure syndrome (e.g., West syndrome); or as adjunctive treatment in intractable epilepsy, prior to and independent of autoantibody test results. A positive clinical response, seen in 15 of 23, was defined as clinical improvement in seizures and/or encephalopathy as judged by the treating clinician. Thirteen of 18 patients given steroids alone showed improvement, including West syndrome (n = 5), epilepsy of unknown cause (n = 4), Lennox-Gastaut syndrome (n = 2), epilepsy attributed to malformation of cortical development (n = 1), and epilepsy attributed to perinatal insult, encephalomalacia (n = 1). Two of five patients who received steroids in combination with IVIG also showed a positive clinical response (Rasmussen encephalitis, n = 1; and basal ganglia encephalitis, n = 1).
Ten (8.8%) of the 114 patients had no follow-up information available; for the remaining 104 patients, the mean length of follow-up (from seizure onset) was 11.7 months (median 11 months, range 1–36 months) at the time of assessment. Eighty-four patients had ongoing epilepsy, including 23 with drug-resistant (refractory) epilepsy (Kwan et al., 2010). Thirty-two patients had new long-term deficits (other than epilepsy), including one or more of the following: developmental delay or cognitive impairment (n = 24), behavioral/psychiatric impairment (n = 18), and motor deficits (n = 13).
Epilepsy subgroups separated by classification
The 114 patients with new-onset seizures were classified according to the structure of the latest proposed ILAE organization of seizures and epilepsies (Berg et al., 2010), Eighteen patients had seizures not traditionally diagnosed as a form of epilepsy; the remaining 96 patients had epilepsy: 30 with electroclinical syndromes, 33 with epilepsy attributed to structural-metabolic causes, and 33 with epilepsy of unknown cause (Table 1).
Table 1. Epilepsy classification as for the new ILAE organization of seizures and epilepsies (Berg et al., 2010)
Conditions with epileptic seizures that are traditionally not diagnosed as a form of epilepsy (n = 18)
These patients had a systemic infection that was presumed to have provoked the seizure, but no evidence of CNS infection or inflammation, and either had no fever or were outside the age definition for febrile seizures. The precipitating infections were gastroenteritis (n = 3), upper respiratory tract infection (n = 2), H1N1 influenza (n = 1), fever alone (n = 1).
“The arrangement of electroclinical syndromes does not reflect etiology” (Berg 2010).
Antibodies and clinical features of positive patients and controls
Eleven (9.7%) of 114 patients with new-onset seizures were positive for one or more of the tested antibodies: VGKC-complex (n = 4), CASPR2 (n = 3), NMDAR (n = 2), and VGKC-complex and NMDAR (n = 2) (Fig. 1A). None of the patients or controls were positive for GAD, GlyR, AMPAR, GABABR, LGI1, or contactin-2 antibodies.
All antibody-positive patients had been admitted to hospital on presentation, compared to 87.4% of the antibody-negative group. The mean duration of hospital stay for the positive patients was 3.72 days (median 3 days, range 2–11 days), but none required admission to the intensive care unit. Two of the antibody-positive patients had preexisting developmental delay, which was severe in one. There was a positive family history of epilepsy in two patients. One of the patients (case 8) had status epilepticus of 45 min duration at presentation. Early seizure recurrence in the first 48 h of presentation occurred in 7 of the 11 patients. Features associated with seizures on presentation are presented in Table 2. Overall there were no significant differences in the demographic and clinical features between antibody-positive and antibody-negative patients (Table 3). The timing of the samples from symptom onset for the positive patients was less but not significantly different from that of the antibody-negative patients (mean 17.5; median 4, range 1–75 days vs mean 40; median 20 and range 1–180 days, respectively). The mean age was 4.39 years for both positive and negative patients; median was 3.4 years for antibody-positive patients and 2 years for antibody-negative patients.
Table 2. Demographic, clinical, electrographic, and imaging features of positive patients (n = 11)
EEG abnormal findings are described in the following order when present: slowing and location, epileptic activity and location, seizure and onset.
CBA were scored on a visual scale; 0 (no binding), 1 (low but specific binding)-4 (strong binding to all transfected cells) by two independent observers. Titration values are highest dilution at which the sample remains positive compared to controls.
Patient defined as nonepileptic with acute symptomatic (provoked) seizure.
Epileptic: left frontotemporal. Seizure: left temporal
Table 3. Comparison of features of positive and negative cases by ILAE organization of seizures and epilepsies (A) and by clinical characteristics (B)
Presence of characteristic in Ab-positive patients (%) n = 11
Presence of characteristic in Ab-negative patients (%) n = 103
ns: not significant, corrected for multiple comparisons.
(A) Epileptic seizures not diagnosed as a form of epilepsy (n = 18)
Patients with epilepsy
n = 10
n = 86
Electroclinical syndrome (n = 30)
Epilepsy attributed to structural-metabolic causes (n = 33)
Epilepsy of unknown cause (n = 33)
Epilepsy of unknown cause (focal) (n = 16)
ns (0. 059)
Epilepsy of unknown cause (generalized) (n = 10)
Epilepsy of unknown cause (spasms) (n = 4)
Epilepsy of unknown cause (undetermined) (n = 3)
(B) Clinical characteristic
Previously normal (n = 78)
Early seizure recurrence
Hospitalization (n = 101)
Intensive care (n = 18)
EEG and MRI brain were performed in 10 of 11 antibody-positive patients. The EEG results were abnormal in 8 and MRI abnormalities were recorded in 4 (Table 2). CSF was tested in only 4 of the 11, and it was normal apart from CSF neopterin, which was elevated in 1 of 3 tested at 68 nm (normal <30; case 1), suggestive of central nervous system (CNS) inflammation (Dale et al., 2009a).
Nine of the positive patients received long-term antiepileptic drugs, but none of the antibody-positive patients received immunotherapy during the period of disease course studied. Two antibody-positive cases were lost to follow-up (cases 8 and nonepileptic case 6). The mean length of follow-up for the remaining nine cases was 10.33 months (median 11 months, range 6–11 months), and was not different from the remaining antibody-negative cases. All of the nine patients with follow-up had ongoing epilepsy, and one had drug-resistant epilepsy (case 9). Five patients were on one AED, three were on two AEDs, and one was on four AEDs. Two patients had a new neurologic or developmental deficit (other than epilepsy) including speech impairment and hyperactivity (case 3), and motor deficit and behavioral alteration (case 5).
Three of the 65 controls (4.6%) were positive for antibodies. Control 1 (VGKC-complex-Abs 497 pm) was a 3-year-old boy with type 1 diabetes mellitus. Control 2 (VGKC-complex-Abs 200 pm) was a 2-year-old boy with vomiting illness and poor weight gain. Control 3 (NMDAR score 2, end point dilution 1 in 100) was an 8-year-old boy with an acute lymphocytic leukemia relapse treated with bone marrow transplantation, complicated by chronic graft-versus-host disease. No neurologic abnormality was noted in any of these controls, although no natural follow-up was recorded in control 2.
Patients with antibody-positive new-onset epilepsy according to the new ILAE organization of seizures and epilepsies
One patient with low levels of CASPR2 antibodies had nonepileptic (provoked) seizures (patient 6, Table 2). The remaining 10 antibody-positive patients with epilepsy had the following ILAE epilepsy classifications: electroclinical syndrome (n = 2), epilepsy attributed to structural metabolic causes (n = 1), and epilepsy of unknown cause (n = 7). In the total epilepsy cohort, after exclusion of nonepileptic patients (n = 96), only 2 of 30 patients with electroclinical syndromes (6.7%) had positive antibodies (cases 9 and 11). Case 9 had Lennox-Gastaut syndrome (LGS) and antibodies to NMDAR, and case 11 had febrile seizures plus with antibodies to both the NMDAR and VGKC-complex. Similarly, only one of 33 patients with epilepsy attributed to a structural-metabolic cause (3%) had positive antibodies (case 7), again not significantly different from the controls; this case had mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS) and was positive for CASPR2 antibodies.
In comparison, 7 of 33 patients from the epilepsy cohort with epilepsy of unknown cause (21.2%) had positive antibodies, which was different from the controls (4.6%; p = 0.03, Fisher's exact test) (Fig. 1). Overall, the proportion of antibody-positive patients classified with epilepsy of unknown cause (7/10; 70%) was significantly higher than the remaining proportion of antibody-negative epilepsy subjects (23/86; 26.7%; p = 0.0095, Fisher's exact test) (Fig. 1B). Furthermore, 4 (57%) of these 7 patients had focal epilepsy compared with 12 of the 86 antibody-negative patients (13.9%; p = 0.059, Table 3).
Of these seven antibody-positive patients with epilepsy of unknown cause, four had VGKC-complex-Abs only (cases 1–4), three of whom presented with focal seizures, and three had early seizure recurrence (cases 1, 3, and 4) (Table 2). Case 5 had CASPR2 antibodies, presented with focal tonic–clonic seizures and had intercurrent infections. Case 8 was positive for NMDAR-Abs and had focal dyscognitive seizures and focal status epilepticus on presentation. Case 10 was double positive for NMDAR and VGKC-complex and had a number of comorbidities. She presented with seizures following an infliximab infusion given for her treatment-resistant Crohn's disease. Infliximab is a monoclonal antibody used in the treatment of some autoimmune diseases and previously described to be associated with seizures (Brigo et al., 2011).
The finding of pathologically relevant autoantibodies to neuronal proteins in a small but significant minority of patients with epilepsy (9–13%, McKnight et al., 2005; Majoie et al., 2006; Brenner et al., 2013) is becoming increasingly recognized, but there have been limited studies in children with epilepsy. Herein we studied a large group of children presenting with new-onset seizures to a tertiary children's hospital and related the results to the recently revised ILAE organization of seizure and epilepsy. The cohort included a heterogeneous group of patients with a wide spectrum of seizure severity and etiology. Overall antibody positivity was detected in 9.6% of the cohort, and surprisingly in 4.6% of the controls. Nevertheless, in patients with epilepsy of “unknown cause,” often with focal seizures, antibodies were more frequent than in the other categories of seizures.
Many patients in the cohort were young and complicated, required intensive acute treatment, or had ongoing refractory epilepsy, as they represented an inpatient sample rather than a community outpatient sample.
The cohort description and classification represented a challenging exercise, particularly at a time when the ILAE classification and terminology is still evolving. It is notable that the patients were classified with investigators blinded to the result of antibody testing. Nearly one third of the patients were determined to have an “unknown cause” for their epilepsy, as also noted in a recent study comparing the new and old systems (van Campen et al., 2013). Of note, 7 of the 10 antibody-positive results were in this epilepsy category. The most frequently found antibody was to the VGKC-complex (4/10). These four patients, however, were negative for the VGKC-complex proteins LGI1, contactin-2, and CASPR2. The lack of an identified complex protein has been reported previously in children (Dhamija et al., 2011; Ilingworth et al., 2011; Suleiman et al., 2011b) and contrasts with adults in whom antibodies to LGI1, and less frequently to CASPR2, are often found (Irani et al., 2008, 2010a,b; Lai et al., 2010), although not inevitably (Patterson et al., 2013). It is likely that the VGKC-complex antibodies bind to other antigenic targets that are yet to be identified in children (Hacohen et al., 2013). On the other hand, three patients were positive for CASPR2 antibodies, which have not been identified before in children, but were negative for VGKC-complex antibodies in the radioimmunoassay, so would have been missed with standard VGKC-complex-Ab screening.
None of the VGKC-complex or CASPR2 antibody-positive patients had a phenotype typical of limbic encephalitis, although most of them presented with focal seizures, suggesting that neuronal antibodies can be present beyond the spectrum of this recognized syndrome, especially in children. Two patients with positive NMDAR antibodies had epilepsy in the absence of classic autoimmune NMDAR-Ab encephalitis: one presented with focal dyscognitive seizures and one had mixed seizure types. This is similar to adult cases described with new-onset epilepsy associated with NMDAR Abs (Niehusmann et al., 2009; Brenner et al., 2013). In addition, two patients in our cohort, both presenting with generalized seizures, were positive for both NMDAR and VGKC-complex-Abs. The finding of two or more neuronal Abs is beginning to be recognized (Irani et al., 2010a; Pellkofer et al., 2010; Haberlandt et al., 2011) and may reflect a wider activation of the immune system, perhaps by infections or possibly as a secondary response to neuronal damage. NMDAR-Abs have recently been reported in one patient with MELAS (Finke et al., 2012), and secondary activation may be a possible explanation for the finding of two antibody-positive patients with known structural or metabolic cause for their epilepsy (cases 7 and 9).
Nine of the 11 positive patients with recorded follow-up had on going epilepsy and/or seizures, but unfortunately none of the 11 positive patients received immunotherapy. These patients did not develop long-term cognitive or behavioral impairment, although the follow-up was short and detailed neuropsychology was not done. Because this was a prospective observational cohort study, we can only hypothesize that immunotherapy, if given, might have improved the epilepsy outcome in these patients.
The control group consisted of children with various immunologic, oncologic, and medical conditions, and three controls were positive for one of the tested antibodies, although they did not have apparent neurologic symptoms. This level of positivity appears to be higher than in adult control cohorts (e.g., McKnight et al., 2005; Brenner et al., 2013). However, few other reports have used childhood controls; indeed all children in this cohort had been hospitalized, and approximately one third of the patients had infectious/inflammatory conditions. The ability to develop antibodies in response to circulating pathogens or environmental factors may be higher in children, particularly in those who are already unwell. However, if antibody levels are sustained, it is possible that these children could be at risk of future neurologic dysfunction.
This study is the first to describe a relatively large prospective cohort of pediatric patients with new-onset seizures, apply the new ILAE system for the organization of seizures and epilepsies (as for Berg et al., 2010), and test them for neuronal antibodies within 6 months of seizure onset. We found that specific neuronal antibodies were more common, and these may play an etiologic role in children with epilepsy of unknown cause, indicating that the new ILAE system may “facilitate the identification of nongenetic determinants of epilepsy” (Berg 2010). We suggest that further antibody testing could be performed in pediatric epilepsy patients classified in this “unknown” category, particularly those with focal epilepsy; it has already been suggested that increasing the number of investigations should help to identify the etiologies of this group (van Campen et al., 2013). Moreover, because the new ILAE organization does not incorporate autoimmune forms in the epilepsy or etiology classification, the possibility of including “immune-mediated epilepsy” as a separate category (Brenner et al., 2013) needs to be considered. However, it is also possible that the presence of neuronal antibodies are, in some circumstances, an epiphenomenon or secondary to structural damage or generalized immune activation. Further studies should concentrate on identifying antibodies in children with epilepsy of unknown cause closer to disease onset, with systematic testing of the effects of immune therapies, and applying the recent guidelines to determine whether these seizure-related antibodies define forms of “autoimmune epilepsy.”
The study was supported by research awards from the National Health and Medical Research Council (NHMRC), Australia (RD, JS); The Wellcome Trust (Clinical Research Training Fellowship to SW) Epilepsy Research United Kingdom (BL), and the NIHR Oxford Biomedical Research Centre (PW, AV). In addition, we would like to thank Philippa Pettingill, Rosie Pettingill, Mark Woodhall, Leslie Jacobson, and Sian Peach (Oxford) for technical support.
AV, PW, BL, and the Nuffield Department of Clinical Neurosciences in Oxford receive royalties and payments for antibody assays. JS, SW, DG, FB, KP, PP, AN, and RD do not report any conflict of interest with respect to this study. We confirm that the authors have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.