Epilepsy or seizures can be a feature of a number of autoimmune or inflammatory disorders. In these conditions, the seizures may be a direct result of the primary disease pathology or could be secondary to the proinflammatory processes that are involved. There could be specific pathogenic antibodies (Abs) in a subpopulation of the patients that react with neuronal antigens and cause the central nervous system (CNS) disorder. But it is also possible that their coexistence may be purely coincidental or reflect some genetic predisposition that is common to the primary disease and the cause of the epilepsy. The study of these conditions, therefore, could help in the understanding of the role of different auto-Abs and/or inflammation, including proinflammatory cytokines, in epilepsy and epileptogenesis.
In this article, we review the incidence and significance of seizures in well-established autoimmune disorders, including multiple sclerosis (MS), diabetes mellitus, celiac disease, thyroid disease, and systemic lupus erythematosus (SLE). The five following presentations discuss the incidence and possible pathogenesis of epilepsies that are found in these well-known autoimmune conditions. There is a large body of evidence describing the clinical presentation of seizures with MS and SLE, and showing that refractory epilepsy can complicate these already challenging disorders. However, the mechanisms involved are complex and generally not well understood. Neurologic syndromes, including seizure disorders, can also be a feature of celiac disease (CD) or subclinical CD, sometimes associated with cerebral calcification. The association between type-1 diabetes mellitus (T1DM) and epilepsy is unclear and requires more definitive epidemiologic analysis, despite the fact that antibodies to glutamic acid decarboxylase may provide a link between the two conditions. The association between thyroid disorders and encephalopathies, often termed Hashimoto’s encephalopathy, is well known but the pathogenic significance of antithyroid antibodies in this condition is still debated. In general, the relationships between autoimmune mechanisms and seizures in these conditions are unclear; the seizures are likely to be caused by a variety of mechanisms, including ischemia, neuronal damage, and specific and nonspecific immunity.
Multiple Sclerosis (P. Striano, Genova, Italy)
Multiple sclerosis (MS) is a highly prevalent demyelinating disease that affects individuals of all ages and can result in significant neurologic disability. It is among the most common neurologic diseases affecting younger (<30 years) individuals, often female, but is also found in male individuals, generally at an older age. Approximately 200 new cases are diagnosed each week in the United States. Diffuse clinical manifestations include fatigue, decreased energy levels, and depression, whereas focal neurologic deficits such as weakness, sensory loss, bladder and bowel dysfunction, and visual loss cause the more profound long-term sequelae. Although the pathogenesis of MS has not been fully defined, there is solid evidence that the development of demyelinating lesions in the white matter reflects an autoimmune process. Analysis of the lesions seen on magnetic resonance imaging (MRI) that are evident pathologically as MS plaques, reveals inflammatory lymphocytic infiltration consistent with active inflammation, both humoral and cellular.
Seizures were first described in MS shortly after the original definition of the disease in the late 19th century, and are a well-described feature of the disease. Cohorts of MS patients with seizures and epilepsy have been identified worldwide (Engelsen & Grønning, 1997). A recent large meta-analysis of >19,000 patients revealed that seizures appear in approximately 2–5% (range 0.5–7.8%; Koch et al., 2008) of MS patients and will often require medical therapy. Seizures may occur throughout the disease course and may be complicated by either epilepsia partialis continua or generalized tonic–clonic status epilepticus. The duration of MS symptoms prior to first seizure is generally several years.
Tonic–clonic (primary or secondary generalized) seizures account for approximately two thirds of all seizures in MS patients. Among the partial seizures, simple partial seizures are about twice as common as complex partial seizures in MS, whereas in the general population, complex partial seizures occur more frequently than simple partial seizures (Sander et al., 1990; Spatt et al., 2001; Striano et al., 2003). Unusual seizure semiologies, such as aphasic status epilepticus and musicogenic epilepsy, have been sometimes described.
Most typically, seizures are seen in patients with lesions that abut the gray–white matter junction, and these lesions may exhibit gadolinium enhancement on MRI, further suggestive of active inflammation. However, in some patients, no correlation with structural lesion is found. Indeed, the variable timing of MRI relative to seizure(s) in the majority of MRI studies of MS patients offers few data from which to draw definitive conclusions. Electroencephalography (EEG) studies performed in MS patients often exhibit focal slowing, spikes, and sharp waves, but can also be normal.
Occasionally seizures are the first manifestation of the disease. One series reported five patients with temporal lobe epilepsy and two cases of epilepsia partialis continua as the first clinical manifestation. Although some authors believe that seizures are more likely to occur during relapses (Büttner et al., 1989; Sokic et al., 2001), others could not confirm this notion (Kinnunen & Wikström, 1986; Ghezzi et al., 1990; Martínez-Juárez et al., 2009). Indeed, seizures have been observed in relapsing–remitting as well as in secondary or primary progressive MS. Given the rarity of seizures in MS, it is difficult to investigate whether seizures are a risk factor for a quicker accumulation of disability or a quicker entry into the secondary progressive phase.
In the setting of an acute relapse, seizures are benign and self-limiting and do not necessarily require treatment, whereas recurrent seizures unrelated to relapse should be treated. Antiepileptic drugs (AEDs) may aggravate some symptoms of MS, such as fatigue, vertigo, ataxia, diplopia, and cognitive slowing. AEDs with a potential harmful impact on motor and cognitive functions should be avoided. There are no clinical trials for treatment of epilepsy in MS patients and, therefore, no clear recommendations can be given at this time, but a systematic review of treatments is available (Koch et al., 2009).
In terms of seizure pathogenesis, it has been suggested that the edema associated with active lesions may be responsible for seizures in MS patients. In addition, there is clear evidence for microglial activation and enhanced expression of tumor necrosis factor (TNF)-α-, interferon (IFN)-β, and specific interleukins (ILs) in MS lesions. How these proinflammatory cytokines contribute to epileptogenesis in MS remains inconclusive, but their roles in seizures are the subject of many current studies (see Friedman & Dingledine, this supplement).
Systemic Lupus Erythematosus (F. Cendes, Campinas, Brazil)
Systemic lupus erythematosus (SLE) is an autoimmune disease that may affect any part of the peripheral nervous system or CNS. There is predominance in women, and overall neuropsychiatric involvement may occur in up to 75% of patients. CNS involvement may vary from subtle signs to severe, life-threatening conditions, and the definition of neuropsychiatric involvement is difficult due to the diffuse CNS involvement in many cases, and the lack of a satisfactory gold standard (Afeltra et al., 2003; Appenzeller et al., 2006a).
SLE has a clear and often vexing relationship with seizures and epilepsy. In one series, of 518 consecutive patients with SLE (follow-up 4–6.8 years), 88 SLE patients (17%) had epileptic seizures, 60 (11.6%) were considered as primary manifestation of SLE, 23 (4.4%) were secondary acute metabolic causes, and 5 (1%) had epilepsy prior to the diagnosis of SLE (Appenzeller et al., 2004). The occurrence of stroke, nephritis, and the presence of immunoglobulin (Ig)G phospholipid antibodies covaried with seizures in these patients.
EEG may show focal slowing, spikes, or sharp waves. EEG findings showed interictal epileptic activity in all patients with recurrent epileptic seizures, although the majority of patients with single epileptic seizures had normal interictal EEG. Indeed, current recommendations are that patients with antiphospholipid antibodies and single epileptic seizure should be followed carefully, because of the greater risk of recurrent seizures. However, most patients with SLE who present with first epileptic seizure will not need to be treated with AEDs, as only 1.3% had recurrent, unprovoked epileptic seizures.
MRI demonstrates cerebral cortical atrophy and variable injury to the subcortical white matter seen on fluid-attenuated inversion recovery (FLAIR) sequences. Changes in white matter are not necessarily dependent on presence of active disease but are markers of remote vasculopathy and do seem to progress over time. Interestingly, there is evidence that these effects may result in axonal dysfunction (not always seen on structural MRI) in patients with active SLE, independent of overt CNS manifestations, that may be reversible during periods of inactivity of disease. These structural changes can be associated with progressive decline and the development of focal neurologic deficits. Death was observed in 58 patients in this cohort of 519 patients during the follow-up period. Hippocampal atrophy can predict cognitive impairment (Appenzeller et al., 2006b) and status epilepticus was the primary cause of death in two patients, with no other evidence of major organ involvement at the time of death.
On one hand, there is evidence that the seizures occur in the setting of brain lesions and damage to subcortical white matter following vasculitic, inflammatory, or ischemic brain injury (Appenzeller et al., 2008). On the other hand, the underlying systemic inflammatory cascades that are activated may be sufficient to cause seizures (see Friedman & Dingledine, this supplement), or specific antibodies to neuronal antigens may play a role in some patients, although the evidence is inconclusive (see Colasanti et al., 2009 for review). Irrespective of the cause, seizures are part of the American Rheumatism Association diagnostic criteria for SLE (ACR, 1999). The development and progression of a small vessel, inflammatory vasculopathy is a hallmark of so-called “lupus cerebritis” or “lupus vasculopathy.” As blood vessels become more inflamed, there is diminution of blood flow to target areas and the development of ischemia or hemorrhage. A slow often inexorable accumulation of cortical and subcortical ischemic injury can occur and may predispose to recurrent seizures. In addition, cytokine effects, autoantibody-mediated lesions, choroid plexus dysfunction, and abnormal hypothalamic-pituitary axis response have all been implicated in various studies, but the evidence is inconclusive.
Overall, seizures in SLE are diverse in semiology and likely to be equally variable in their cause, but better biomarkers that help to distinguish between different forms with possible implications for treatment should be sought.
Gluten and Occipital Calcification (G. Gobbi, Bologna, Italy)
Celiac disease (CD) is an autoimmune condition characterized by chronic inflammation in the wall of the small intestine with villous atrophy due to intolerance toward gluten, a constituent of wheat flour. The immune reaction is triggered by ingestion of gluten, a heterogeneous mixture of glutenin and gliadin that induces an abnormal immune response in predisposed subjects. The condition is associated with the HLA class II gene HLA-DQ2 in 90% of CD patients, and HLA-DQ8 in 10% of cases, strongly suggesting that CD is, as widely thought, a primary autoimmune disease. Immune responses against gliadin include the activation of specific T cells, with secretion of aggressive proinflammatory cytokines (INF-γ, TNF-α, IL-1α, IL-1β, IL-2, IL-6) that are involved in the intestinal pathology. In addition, there are both humoral and cellular immune responses against tissue transglutaminase (tTG), one form of which, tTG6, is largely distributed in the CNS (amygdala, hippocampus, cerebellum, cerebral cortex). It is possible that a primary immune response targeting different tTG isoenzymes might explain the different manifestations of CD.
Gastrointestinal symptoms are the predominant presenting complaint in patients with CD, but it has been increasingly recognized that neurologic disorders, particularly cerebellar ataxia, peripheral neuropathy, dementia, myopathy, myelopathy, and epilepsy, are not uncommon (Cooke & Thomas-Smith, 1966; Gobbi et al., 1992). The prevalence of these neurologic disorders in well-established CD is from 10–22.5%, but there are also many reports of neurologic manifestations in patients who do not have gastrointestinal symptoms, but in whom the presence of antibodies to gliadin or tTG, and jejunal biopsy are consistent with silent or latent CD (reviewed in Hadjivassiliou et al., 2010). In some patients, epilepsy with or without cerebral calcifications has been proposed to be one such indication of silent or latent CD. Because CD may present with malnutrition ranging from severe malabsorption to near health, nutritional deficiencies could be responsible for the development of epilepsy and other neurologic manifestations, and cerebral calcifications could result from chronic folic acid deficiency due to malabsorption. Alternatively, this deficiency could be due to some of the AEDs, but AED-induced folate deficiency is rare (e.g., by phenytoin), and in some cases AED therapy was started after the discovery of the calcifications.
In the case of gluten ataxia and peripheral neuropathies, detailed postmortem reports have shown an inflammatory process primarily involving the cerebellum and other regions of the CNS and peripheral nerves, suggesting an immune-mediated pathogenesis. Circulating activated T cells may cross the intact blood–brain barrier (BBB) and could damage myelin or myelin-producing cells. Moreover, inflammatory cytokines or IgA deposition in blood vessels of the brain, and the perivascular inflammatory infiltrate, could compromise the BBB, thereby enabling circulating antibodies to enter the CNS and to trigger CNS involvement. In patients with epilepsy and cerebral calcifications, it was shown that 123I-IL-2 is taken up both in the bowel and in the occipital lobe where there were calcifications. This could be due to local inflammation or to binding to glial cells, which can express IL-2 receptors. Cerebral calcification could also be due to chronic immune-complex–related endothelial inflammation (see Signore et al., 1997). Because circulating cytokines may have proconvulsant actions and worsen seizures (see Friedman & Dingledine, this supplement), any or all of these mechanisms could contribute to the seizures that accompany this pathology.
The role of immunity toward the nervous system in CD, and its involvement in causing epilepsy, are areas that need clarification, both to address the cause of the neurologic disorders such as cerebellar ataxia, that are relatively commonly associated, and the cerebral calcification and the epilepsy that occurs. The possibility that antibodies to tTG2 and tTG6 are specifically involved (Hadjivassiliou et al., 2010) needs to be further investigated, and serum and CSF levels of these or other antibodies, and cytokines, need to be studied.
Diabetes Mellitus (A. Marson, Liverpool, United Kingdom)
Type-1 diabetes mellitus (T1DM) is a common condition caused by autoimmune destruction of the pancreatic islet cells. The T cell–mediated process is accompanied by auto-Abs to islet cell antigens such as glutamic acid decarboxylase (GAD) and islet cell antigen 512. Interestingly, the presence of GAD-Abs in T1DM was first demonstrated in patients with stiff-person syndrome (SPS), a neurologic disorder that is a rare but well known complication of diabetes, and sometimes associated with seizures. GAD is expressed in the CNS where it is responsible for the synthesis of γ-aminobutyric acid (GABA), the main inhibitory neurotransmitter. However, because the enzyme GAD is intracellular, there is no reason, a priori, to suspect antibody-mediated dysfunction of the inhibitory neurons in the etiology of either diabetes or of SPS.
In endocrine disorders, one has to consider that seizures could be the result of metabolic disturbance or of associated neuroinflammation/autoimmunity. Indeed, genetic causes such as mutations in JCNJ11, or POLG1, or MELAS, can be associated with both diabetes and generalized epilepsy; conversely, both hypoglycemia and hyperglycemia can precipitate seizures. In addition, AEDs that cause weight gain may lead to type-2 diabetes. Here the discussion is confined to the role of immunity in T1DM.
Seven of 518 patients with T1DM had either juvenile myoclonic epilepsy or generalized tonic–clonic seizures, which was higher than the predicted incidence of epilepsy (1.6 cases per 1,000) from a control population (McCorry et al., 2006). One other large study, found a prevalence of 9 per 1,000 active epilepsy patients within a similar pediatric diabetic cohort, but, in this case, a similar prevalence in the control cohort (O’Connell et al., 2008). A higher prevalence (2.4%) was concluded from a smaller group of patients seen in a specialist pediatric center in Italy (Mancardi et al., 2010) but with no control group for comparison. Generally speaking it seems unlikely that there is a substantial increase in idiopathic generalized epilepsy, and no evidence has been provided for an increase in focal epilepsy. Unfortunately, GAD-Abs, at the least a useful sign of increased autoimmunity, were not measured in any of these epidemiologic studies.
There is increasing evidence that the presence of high titers of GAD-Abs (>1,000 U/ml) is a marker for an immune-mediated disorder. The conditions associated with GAD-Abs include not only SPS, but also cerebellar ataxia and a form of limbic encephalitis in which temporal lobe epilepsy is common, and a female predominance (see Bien and Scheffer, this supplement). In addition, individual case reports include a variety of other focal epilepsies. The question that needs to be addressed is how often these Abs are found in patients with typical cryptogenic or idiopathic epilepsy—whether focal or generalized. Increased incidence of GAD-Abs has been reported in a few cohorts of patients, either including those with an autoimmune predisposition or with drug-resistant epilepsy (Peltola et al., 2000; McKnight et al., 2005). These aspects are more fully discussed by Bien and Scheffer, this supplement.
In conclusion, epilepsy is not a very common accompaniment of diabetes in young patients, although there might be a modest increase in incidence. Autoimmunity to GAD in epilepsy is beginning to be well recognized, mainly so far in adult cases, and this needs to be subjected to more extensive analysis. Although GAD-Abs themselves are unlikely to be directly pathogenic, GAD-Abs may be a marker for the presence of other, more pathogenic Abs. Future studies need to address more carefully both the existence of GAD-Abs in epilepsy and diabetes, the presence of Abs against extracellular determinants, and their role.
Thyroid Immunity and Hashimoto’s Encephalopathy (S. Rüegg, Basel, Switzerland)
Thyroid disease is another area in which the association with epilepsy is not very clear. Thyroid hormones acting via their transporters and specific nuclear receptors influence both the nervous system and the immune system. Triiodothyronine (T3) influences oligodendrocyte differentiation, and hormone deprivation during development causes cretinism (Schweizer et al., 2008; Williams, 2008). The effects of thyroid hormones on the human immune system are less well known (Bittencourt et al., 2007) but include altered T-cell subsets and increased secretion of Abs by thyroid-stimulating hormone (TSH) and thyrotropin-releasing hormone (TRH). However, the main relationship of interest is between the presence of antithyroid Abs and a severe but reversible encephalopathy.
Hashimoto’s thyroiditis (HT) is associated with auto-Abs directed against thyroid peroxidase (TPO) or thyroglobulin (TG) associated with destruction of the thyroid follicular cells and induces hypothyroidism. The prevalence of subclinical hypothyroidism is 4–8% and may reach up to 20% in women >60 years; however, in only very rare cases (about 200 cases reported in the literature; Chong et al., 2003; Chaudhuri & Behan, 2003; Castillo et al., 2006), a severe acute/subacute encephalopathy develops and is usually referred to as Hashimoto’s encephalopathy (HE).
HE is broadly defined by signs and symptoms of an encephalopathy, exclusion of other causes, and the presence of at least elevated anti-TPO-Abs and/or other thyroid antibodies, usually anti-TG-Abs. A recent study (Castillo et al., 2006) suggests a more rigorous definition involving seven criteria: (1) encephalopathy as indicated by cognitive impairment, neuropsychiatric features, myoclonus, generalized tonic–clonic or partial seizures or focal neurologic deficits; (2) serum antithyroid Abs as above; (3) a euthyroid or mildly hypothyroid state (with appropriately raised TSH levels); (4) no evidence of infectious, toxic, metabolic, or neoplastic process; (5) no evidence of specific antineuronal Abs that have been implicated in immune-mediated encephalopathies (see Bien and Scheffer, this supplement); (6) no clear findings on neuroimaging; and (7) complete or near complete clinical response to steroids. It is unlikely that all of the previously reported cases would fulfill these criteria, and many clinical laboratories do not have access to the antibody testing required to fulfill criterion no. 5.
The patients who have been diagnosed in the past have diverse presentations, and defining clinical or imagining features is clearly difficult. There may be stroke-like episodes, movement disorders, migraine, or psychosis, and in some cases with white matter lesions on MRI, the features are more suggestive of “recurrent” acute demyelinating encephalomyelitis (ADEM; Chaudhuri & Behan, 2003). The MRI may show signs of inflammation, white matter changes, or vasculitis, but is normal in up to half the cases; the CSF often shows mild pleocytosis and elevated protein, but oligoclonal bands are not always present. The EEG is almost always markedly altered, and clinically seizures are present in 70–80% of patients, with status epilepticus in 10–20% (Schäuble et al., 2003), but the EEG patterns and seizure-types are diverse. Histopathologic evaluation reveals four patterns of changes in the brain: gliosis, demyelination, spongiform transformation, or marked vasculitis of venules only.
Despite this diversity, the patients can respond dramatically to high-dose steroids within days, and complete recovery is usually observed after 2 months. The disorder may relapse, especially after tapering steroids, but most patients definitively remit after a few relapses. More intensive and combined immunomodulatory treatment with intravenous immunoglobulins, plasma exchange, cyclophosphamide, or rituximab may be necessary in some cases. The role of nonsteroid drugs with immunosuppressant actions like azathioprine and methotrexate during long-term immunosuppression is not clear.
The compulsory presence of the Abs and the rapid treatment response are the main features that distinguish this condition from the other autoimmune forms of encephalitis (see Bien & Scheffer, 2011), but anti-thyroid Abs, and hypothyroidism, are very common in the population, particularly in older individuals. It is very likely that in the past this diagnosis has been given to patients with other autoimmune forms of encephalopathy who happen to have antithyroid Abs and are treatment responsive. The treatment responses observed are not different from those found in other antibody-mediated conditions that have recently become recognized (see Bien & Scheffer, this supplement).
The variability in EEG findings, the lack of distinctive MRI changes, and the variable pathologic features—some of which, such as spongiform transformation, are more likely to represent Creutzfeldt-Jakob disease (Chong & Rowland, 2006)—suggest considerable heterogeneity of the disease as previously defined and some findings are difficult to reconcile with a reversible condition. For these reasons, and because the Abs do not necessarily parallel the disease course, many authors now prefer to talk about a steroid-responsive encephalopathy associated with autoimmune thyroiditis (SREAT; Schäuble et al., 2003).
The role of the antithyroid Abs in HE is uncertain (Schiess & Pardo, 2008). These Abs might cause direct immune-mediated damage to the brain, but it seems equally, if not more likely, that they could be an innocent bystander or a marker for the existence of other antibodies that are directly pathogenic. Further studies need to define more rigorously HE or rely on SREAT, and to look directly with current techniques for the possible pathogenic effect of the antithyroid Abs or of the presence of other Abs that are more likely to be involved in reversible neuronal damage.
There is clear evidence that both systemic and brain specific inflammatory disorders are associated with epilepsy. Although in some disorders seizures are associated with actual structural brain injury, in others only the inflammatory response may trigger seizures. In some patients, specific auto-Abs may play a role. Important issues that remain to be elucidated include defining particular inflammatory cascades that may predispose to seizures in these conditions, identifying reliable serum biomarkers, including specific Abs that could be used to predict seizure risk, and discovering specific therapies that may target the inflammatory response and act as antiepileptogenic treatments.
Fernando Cendes, Campinas, Brazil; Guiseppe Gobbi, Antonella Ciriaco, and Eliana Parente, Bologna, Italy; Anthony Marson, Liverpool, United Kingdom; Stephan Rüegg, Basel, Switzerland; Pasquale Striano, Genova, Italy.
AV serves as a paid consultant for Athena Diagnostics and Bayhill Therapeutics; she and the University of Oxford receive royalties and payments for antibody tests. PBC delcares no conflicts of interest. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.