Clinical manifestations

Presentation and course

Although the etiology is diverse, nonparaneoplastic immune-mediated cerebellar ataxias share some common clinical phenotypes. The onset of cerebellar ataxias is usually acute/subacute, but it can be chronic or insidious. The presence of other autoimmune disorders in the patient or first-degree relatives is common. The main ataxic feature is gait ataxia, which hinders gait, standing, and steady walking. This is due to the preferential involvement of the vermis. Other manifestations include variable degrees of limb incoordination, dysarthria, and nystagmus. MRI can be normal or show atrophy mainly in the vermis depending on the duration of the disease. CSF pleocytosis or oligoclonal bands are sometimes observed (48; 52; 98).

There is no consensus on the classification of immune-mediated cerebellar ataxias. We have proposed a classification based on (1) whether the cerebellum is the main target of autoimmunity or not, and (2) whether autoimmunity is triggered by other conditions or not (92; 98). In cerebellar ataxias associated with connective tissue diseases, the cerebellum is one of many targets of the autoimmune attack, and cerebellar ataxias are usually associated with other extracerebellar symptoms. On the other hand, when the cerebellum is the sole target of autoimmunity, the immune-mediated cerebellar ataxias are categorized into two groups: those where autoimmunity is triggered by other conditions, such as infection (eg, postinfectious cerebellitis) and gluten sensitivity (gluten ataxia), and those where autoimmunity has no obvious triggers (eg, anti-GAD ataxia). Another clinical entity, primary autoimmune cerebellar ataxia (PACA), was proposed by Hadjivassiliou and colleagues (42). Cerebellar ataxias are diagnosed as primary autoimmune cerebellar ataxia when immune-mediated mechanisms are strongly suspected, but no underlying trigger or pathogenic autoantibodies are identified. Thus, this group is likely to be heterogeneous and probably includes several autoimmune etiologies that await characterization.

Prognosis and complications

Based on the response to treatment, the clinical course can be classified into six patterns (93; 98). In patterns (a)-(c), immunotherapy can stop the progression, whereas patterns (e)-(f) show no therapeutic benefits. The difference among these groups reflects the effects of immunotherapy on the different autoimmune processes. Notably, although immunotherapy can stop the progression (patterns (a)-(c)), the subsequent prognosis is mixed, ranging from different patterns of recovery (a) (full recovery) and (b) (partial recovery) to no recovery in pattern (c) (stabilization). Thus, the scheme indicates that the outcome of any therapy depends on two factors: the responsiveness to immunotherapy and the presence of cerebellar reserve. The latter is determined by the degree of damage, suggesting the existence of a threshold. Above such threshold, ataxia can recover if immunotherapy can stop the disease process. In contrast, below this threshold, although the disease process can be stopped, cerebellar ataxias remain unchanged, and the patient remains disabled. Based on this recoverability, we introduced the concept of restorable (presence of cerebellar reserve) and nonrestorable stage, which can be viewed as a cerebellar circuitry, which can be recovered or not in terms of functionality (no cerebellar reserve) (96; 100; 98).

Clinical vignette

A 58- year-old woman developed progressive unsteadiness of gait, vertigo, and oscillopsia related to postural head changes during the past year. Her medical history was significant for type 1 diabetes mellitus diagnosed at age 45 years that required insulin treatment at 6 months of diagnosis and psoriasis at age 38 years. General examination was normal except for psoriasis. The neurologic examination disclosed downbeat vertical nystagmus and ataxic gait. The remainder of her neurologic examination was normal. MRI showed type 1 Arnold-Chiari malformation. Routine laboratory analysis including thyroid function tests were normal. Over the ensuing 2 years she noticed progressive dysarthria, and her gait became more unsteady. Clinical progression was initially believed to be related to the Chiari malformation, and the patient underwent posterior suboccipital decompression. After surgery, her dysarthria and downbeat nystagmus improved but not her gait. Over the ensuing 3 years she developed clumsiness of the right hand and increasing unsteadiness. Neurologic examination revealed mild dysarthria, upward vertical nystagmus, limb dysmetria more marked on the right, and severe ataxic gait that prevented her to walk alone more than a few meters. MRI demonstrated resolution of cerebellar tonsillar descent and mild atrophy of the vermis. The following organ-specific autoantibodies were detected: pancreatic islet-cell, gastric parietal cell, and thyroid microsomal antibodies; adrenal glands and thyroglobulin antibodies were negative. Whole-body CT scan was normal. CSF examination was normal except for the presence of oligoclonal IgG bands. High levels of GAD antibodies were confirmed in the serum and CSF with a ratio compatible with intrathecal synthesis of GAD antibodies. The patient was treated with multiple courses of intravenous immunoglobulins, azathioprine, and rituximab without improvement. At last visit, 15 years after the onset of the ataxia, she was wheelchair bound with a severe pancerebellar syndrome.

Comment. This patient showed poor response to immunotherapies. One should underline that the disorder was evolving for at least 2 years. If diagnosis was delayed (partly because the Chiari malformation may have been the cause), neuronal loss occurred inevitably, with negative consequences on the cerebellar reserve. This case suggests the utmost importance of early diagnosis and therapy (time is cerebellum).

Biological basis

Etiology and pathogenesis

The mechanisms underlying any autoimmune diseases include: "pathogenic roles of effector T cells (Th1/17 cells and CD8 T cells) or autoantibodies," "autoimmune triggers by deficits in immune tolerance or molecular mimicry," "pathological permeability of blood-brain or blood-nerve barrier," and "exacerbation by local neural inflammation" (97). However, little is known about the contribution of these processes in the pathogenesis of non-paraneoplastic immune-mediated cerebellar ataxias.

Many autoantibodies are associated with nonparaneoplastic immune-mediated cerebellar ataxias. It has been a focus of debate whether some autoantibodies have a pathological role or they are mere markers of an autoimmune response. Autoantibodies to ion channels or ion channel-related proteins (Ca or K channel), glutamate receptors, or GABA synthesis enzyme (GAD) have been considered to be pathogenic in the development of cerebellar ataxias, some of which (anti-GAD and anti-mGluR1 antibodies) have been confirmed both in vitro and in vivo (97). The same is true for gluten ataxia where antigliadin and transglutaminase antibodies (TG2 and TG6) have been shown to cross react with cerebellar cells and are capable of inducing ataxia in mice (09).

On the other hand, the role of cell-mediated autoimmunity is still unclear. Studies on experimental autoimmune encephalomyelitis have suggested possible T cell-mediated autoimmune mechanisms, specifically involving CD4+ T cells. First, naïve CD4+ T cells (Th 0 cells) are activated by antigens presented by major histocompatibility complex class II molecules on antigen-presenting cells and differentiate into Th1 and Th17 cells in the periphery. Subsequently, they infiltrate the CNS, where they start to orchestrate an inflammatory cascade by secreting cytokines (eg, IFN-γ by Th1 and IL-17 by Th17), leading to demyelination or cell death (114).

In addition, regulatory T cells (Treg) are a population of differentiated CD4+ T cells that maintain self-tolerance, inhibit autoimmunity, and act as critical negative modulators of inflammation in various autoimmune diseases (22). Thus, dysfunction of Treg is also one of mechanisms that precedes and leads to cell-mediated autoimmunity (22). Further studies are necessary to elucidate these cell-mediated mechanisms in nonparaneoplastic immune-mediated cerebellar ataxias.

Epidemiology

The study by Hadjivassiliou and colleagues is the only large-scale investigation of the prevalence of immune-mediated cerebellar ataxias (51). Based on a study performed at the Sheffield Ataxia Centre, United Kingdom of 1500 patients with progressive ataxia, the authors reported that 33% of the patients had genetic disorders, although some did not show evident family history, and that 11% of the patients had multiple system atrophy. Apart from the above, 30% had definite immune-mediated cerebellar ataxias, 25% had gluten ataxia, and 3% had primary autoimmune cerebellar ataxia, whereas 2% had anti-GAD ataxia (Table 1), 1% had postinfectious cerebellitis, and less than 1% had opsoclonus myoclonus syndrome. Interestingly, 24% of the patients were classified to have idiopathic sporadic ataxia. This category probably included a large number of patients with primary autoimmune cerebellar ataxia (48). This study was published prior to the publication of the diagnostic criteria for primary autoimmune cerebellar ataxia.

Gebus and colleagues reported a smaller number of patients with isolated vermian pathology (two with paraneoplastic cerebellar degeneration and one patient with postinfectious cerebellitis amongst 80 patients) (33). In a systematic study of 684 South Korean patients with progressive ataxia, only 21 had isolated vermian pathology; 14 patients had paraneoplastic cerebellar degeneration, three patients had postinfectious cerebellitis, and four patients had other immune causes (74).

Table 1. Clinical Profiles of Common Etiologies of Nonparaneoplastic Immune-Mediated Cerebellar Ataxias

Prevention

It is uncertain how to prevent the risk of autoimmune damage towards the cerebellum. If it can be done it is best to intervene during the restorable stage using combinations of immunotherapies. In cases of gluten ataxia, there is evidence to suggest to gastroenterologists that patients with Coeliac disease already have evidence of cerebellar dysfunction even if they do not necessarily complain of ataxia. In this situation, the introduction of strict gluten-free diet will help prevent the development of overt ataxia (43). Screening for the presence of gluten sensitivity–related antibodies in healthy individuals may prove to be one way of preventing the future development of gluten ataxia.

Differential diagnosis

Immune-mediated cerebellar ataxias have diverse etiologies. Here we review detailed clinical features and discuss differential diagnosis (Table 1).

Common etiologies in nonparaneoplastic immune-mediated cerebellar ataxias

Gluten ataxia. Gluten ataxia is defined as sporadic cerebellar ataxia associated with gluten sensitivity (46). This is by far the most common immune-mediated cerebellar ataxia and one of the few with a known antigenic stimulus (gluten proteins) (52).

Gluten ataxia affects mainly women (55%) with mean age of 52 years and exhibits either chronic or insidious onset (38; 48; 52). Gluten-sensitive enteropathy or gastrointestinal manifestations are seen in about half (47%) of the patients. The HLA type DQ2 is detected in 70% of the patients, and association with other autoimmune diseases, such as thyroiditis, type 1 diabetes mellitus, and pernicious anemia, is common. Gluten ataxia is sometimes associated with sensorimotor axonal neuropathy and is less common with focal myoclonus and palatal tremor. Patients present with gait ataxia and variable degrees of limb ataxia, scanning speech, and ocular ataxia; MRI and MR spectroscopy consistently show vermian involvement (48). Up to 50% of patients with gluten ataxia have CSF abnormalities (38; 48; 52).

Gluten sensitivity is assessed using autoantibody assays. Because anti-transglutaminase 2 antibodies, found in patients with celiac disease, are negative in 53% of patients with gluten ataxia without enteropathy (54), this antibody is not sufficient to diagnose gluten ataxia. Anti-gliadin antibodies (IgG and IgA) are reliable for the diagnosis (52); however, the cutoff level and utility of different assays should be noted (39). In patients with gluten ataxia without enteropathy, the main autoimmune responses occur within the CNS, resulting in low levels of serum anti-gliadin antibodies (52). The calibration process used in certain commercially available anti-gliadin antibody kits is based on the use of serum from patients with celiac disease; therefore, the cutoff level is too high for the detection of gluten ataxia (52). Anti-transglutaminase 6 antibody, a brain expressed transglutaminase is positive in 72% of patients with gluten ataxia as defined by positivity for anti-gliadin antibodies (40). Because anti-transglutaminase 6 is primarily expressed in the CNS, anti-transglutaminase 6 antibodies could be an important specific biomarker (52).

In some cases, the neurologic deterioration can be rapid and devastating, mimicking postinfectious cerebellitis and paraneoplastic cerebellar degeneration (107). This atypical subtype requires prompt diagnosis and rapid intervention (using immunosuppression as well as gluten free diet) in order to avert severe and permanent neurologic disability.

Gluten-free diet is considered an effective therapy for gluten ataxia, based on avoidance of antigens that can trigger immune-mediated mechanisms, similar to the strategy used in celiac disease (38; 48; 52). In a study involving 43 patients with gluten ataxia, significant improvements in cerebellar ataxias and a decrease in anti-gliadin antibodies were noted in patients who adhered to gluten-free diet compared with those who refused gluten-free diet (44). Other reports described the effectiveness of intravenous immunoglobulins in patients with resistance to gluten-free diet (126). It is now considered that resistance to gluten-free diet is due to either poor adherence to gluten-free diet or hypersensitivity to even small amounts of gluten present in some commercially available gluten-free food or due to cross-contamination (52). Because persistently high levels of anti-gliadin antibodies are present in the above two groups of patients (47; 55), monitoring anti-gliadin antibodies can be useful. MR spectroscopy can also be used to detect the effects of gluten-free diet (48). Patients on strict gluten-free diet show an increase in the relative N-acetylaspartate/creatine (NAA/Cr) area in the cerebellar vermis, whereas no such increase is noted in patients on gluten-free diet with persistently positive antibodies and those who do not adhere to gluten-free diet. Taken together, in patients with persistently high anti-gliadin antibody titers or no changes in MR spectroscopy, strict adherence to gluten-free diet should be considered before switching to immunotherapy. When cerebellar ataxias cannot be controlled with gluten-free diet, maintenance therapy with immunosuppressants (eg, mycophenolate mofetil and rituximab) is recommended (38; 48; 39).

Cerebellar ataxias associated with cortical myoclonus is a rare subtype of gluten sensitivity-related neurologic disorder (119). Although the myoclonus is of cortical origin, hyperexcitability of the cerebral cortex is elicited by the cerebellar pathology (119). Characteristically, this subtype shows resistance to gluten-free diet. The neurologic refractoriness is associated with residual enteropathy, which is detected by repeat duodenal biopsies (119). Some of these patients have refractory celiac disease type 2 and are at high risk of development of enteropathy-associated lymphoma (119). Based on these features, this condition is termed "neurologically refractory" celiac disease (52). All such patients require gluten-free diet plus immunosuppression, usually with mycophenolate and, in some instances cladribine, in addition to anti-epileptic drugs (levetiracetam, perampanel, and clonazepam) for the control of the myoclonus (119). The prognosis remains poor (119).

Postinfectious cerebellitis. Infection-induced cerebellar ataxias are classified into two categories: (1) inflammation caused by direct invasion of viral or bacterial microorganisms, and (2) inflammation induced by immune mechanisms triggered by the infection (07; 52; 98). Some groups termed the former type "acute cerebellitis" (125) and the latter as post- or para-infectious cerebellitis, postinfectious cerebellar ataxia, or acute cerebellar ataxia (120; 07). The latent interval of several days to weeks is reminiscent of a delayed antigen-antibody reaction (87). Postinfectious cerebellitis affects mostly young children after an episode of infection, usually viral infection, most commonly varicella) (19; 07; 125). Other etiologies include viral infections, such as Epstein-Barr virus, Coxsackie virus, influenza A and B virus, parainfluenza virus, measles, mumps, and rubella, and bacterial infections, such as diphtheria, pertussis, typhoid, Legionnaires disease, leptospirosis, Borrelia and mycoplasma (87; 07; 125). Based on a systemic survey of 73 patients (19), 60% of the patients were between 2 and 4 years of age. Furthermore, 25% of the children had varicella, 52% had other viral infections, and 3% developed postinfectious cerebellitis after immunization. The mean latency between infection and the onset of cerebellar ataxias was 9.9 ± 7.9 days, although it should be noted that 19% of the patients did not exhibit any preceding infection.

The main clinical feature of acute-onset cerebellar ataxias is gait ataxia (19). Patients are usually afebrile and do not show meningeal signs, high intracranial pressure, or extracerebellar manifestations, such as temporary clouding of consciousness, seizures, extremely altered mental status (eg, extreme irritability), or cerebral focal signs (19). Rather, the presence of these clinical signs is suggestive of inflammation induced directly by infection rather than inflammation induced by autoimmunity (125). However, mild behavioral changes, such as mild irritability, hyperactivity, moodiness, and whining are sometimes noted by parents; the behaviors correlate with the severity of cerebellar ataxias, suggesting cerebellar involvement in cognitive and emotional disorders (121). CSF examination shows pleocytosis (roughly equal numbers of granulocytes and lymphocytes in about 75% of the patients and sometimes lymphocyte predominance in the remaining cases) and high CSF/serum IgG index in half of the children. Oligoclonal bands are sometimes present (19). MRI usually shows no atrophy or areas of abnormal intensity.

Because postinfectious cerebellitis is self-limiting, close monitoring of the patient and follow-up form the basis of clinical management (39; 98). Only when the cerebellar ataxia persists or progresses, a combination of immunotherapies should be considered (07; 39). However, one has to bear in mind that some other immune-mediated ataxias may present very similar to acute cerebellitis and, therefore, the absence of improvement and evidence of progression should raise the possibility of another immune-mediated ataxia. One large scale-study of 60 pediatric patients showed full recovery of the gait ataxia in 72% of patients, within about 2 months in most cases (19). Despite favorable prognosis in the majority of cases, some patients develop permanent sequelae, especially in the elderly. The possibility of swelling of the cerebellum with hydrocephalus should be kept in mind.

Opsoclonus myoclonus syndrome. Opsoclonus myoclonus syndrome shows opsoclonus (repetitive, involuntary, random, and rapid eye movements in both horizontal and vertical directions) and action myoclonus, with subacute onset (05; 76; 03). Because cerebellar ataxias are sometimes associated with opsoclonus myoclonus syndrome, the latter is also known as opsoclonus myoclonus ataxia syndrome (52). Opsoclonus myoclonus syndrome affects primarily children; adult opsoclonus myoclonus syndrome is rare (03). Opsoclonus myoclonus syndrome is classified into three types: paraneoplastic, postinfectious, and idiopathic (52; 98). Opsoclonus myoclonus syndrome is due to neuroblastoma in the majority of pediatric cases. In a retrospective study of 24 patients by Bataller and colleagues, 10 patients had idiopathic opsoclonus myoclonus syndrome (05). The age of onset in the idiopathic group was 40, and there was no evidence of the trigger factors. All patients within the idiopathic group had truncal ataxia.

In the case of paraneoplastic opsoclonus myoclonus syndrome, the neoplasm should be excised immediately (if possible) after confirmation of the diagnosis and should be followed by immunotherapies (eg, monotherapy or a combination of corticosteroids, intravenous immunoglobulins, plasmapheresis, immunosuppressants, and rituximab) (05; 76; 03; 39; 98). Postinfectious opsoclonus myoclonus syndrome is self-limiting whereas idiopathic opsoclonus myoclonus syndrome shows spontaneous recovery in some patients (05; 76; 03; 39; 98). Thus, similar immunotherapies should be used in patients with postinfectious or idiopathic opsoclonus myoclonus syndrome, especially those who show persistent or progressive symptoms (05; 76; 03; 52; 98). Taken together, the treatment strategy should include the exclusion of neoplastic lesions using positron emission tomography and onconeuronal antibodies (52). Several studies investigated the differences in prognosis between paraneoplastic and idiopathic opsoclonus myoclonus syndrome, which demonstrated good response to therapy, defined by a modified Rankin Score of 2 or less, in 39% of patients with paraneoplastic opsoclonus myoclonus syndrome and 84% of those with idiopathic opsoclonus myoclonus syndrome and relapse in 24% of patients with paraneoplastic opsoclonus myoclonus syndrome compared with 7% of patients with idiopathic opsoclonus myoclonus syndrome (03). A similar response to therapy was observed in other studies (05; 76).

Anti-GAD ataxia. GAD is an enzyme that catalyzes the conversion of glutamate to GABA (97; 99; 36). GAD exists in two isoforms: GAD65 and GAD67. The autoantibodies against the smaller isozyme are associated with cerebellar ataxias (97; 99). Serum and CSF anti-GAD65 Ab titers are higher in anti-GAD ataxia (usually more than 10,000 U/mL, or 10 to 100-fold higher) compared to those in patients with type 1 diabetes mellitus (T1DM) (92; 97; 99; 36). Importantly, cerebellar ataxias associated with low anti-GAD antibodies titers are not categorized into this group (99). The radioimmunoassays or enzyme-linked immunosorbent assays (ELISA), which are utilized in routine commercial tests, are adequate to detect low titer of anti-GAD65 in type 1 diabetes mellitus (36). Low titer of anti-GAD antibodies is observed in serum of 1% to 8% of healthy subjects (36). Although autoimmunity toward GAD65 does not affect only the cerebellum but rather involves the entire central nervous system, the cerebellum is one of the most vulnerable areas. The causes that trigger the autoimmunity have not yet been identified. Anti-GAD ataxia can be associated with other types of immune mediated cerebellar ataxias, such as paraneoplastic cerebellar degeneration and gluten ataxia (92; 36).

The condition mostly affects women (82%) in their 60s, with subacute, chronic, or insidious time course (02; 97; 99; 36). Autoimmune diseases, such as type 1 diabetes mellitus, autoimmune thyroid diseases, and pernicious anemia, are associated with anti-GAD cerebellar ataxias in at least some cases. Cerebellar ataxias are sometimes associated with extracerebellar symptoms, including temporal lobe epilepsy, limbic encephalitis, ophthalmoplegia, opsoclonus, and stiff-person syndrome. The overlap syndromes are observed during follow-up and over the years in 14% to 36% of the patients with cerebellar ataxia (21). Some patients show oligoclonal bands in the CSF, whereas MRI shows normal or atrophic changes depending on the duration of illness (02; 97; 99; 21).

The significance of anti-GAD65 antibodies has been a matter of debate, with some researchers arguing that these antibodies have no pathogenic roles for the following three reasons (99; 86; 21; 36). First, GAD65 is an intracellular antigen. Second, anti-GAD65 antibodies are associated with type 1 diabetes mellitus and other neurologic diseases, such as epilepsy, limbic encephalitis, ophthalmoplegia, and stiff-person syndrome (29). Finally, passive transfer studies failed to mimic neurologic symptoms. For example, cerebroventricular or intrathecal administration of IgG from stiff-person syndrome patients have not consistently elicited stiff-person syndrome-like symptoms (36). When animals were immunized with human GAD65, no neurologic symptoms were observed (14). On the other hand, accumulated physiological evidence suggests a pathogenic role for anti-GAD65 antibodies in the development of cerebellar ataxias, which is explained below.

Epitope-specific actions. Previous studies using slice-tissue preparations showed that the addition of CSF IgGs from different patients with anti-GAD ataxia to the medium resulted in presynaptic inhibition of GABAergic synapses between basket cells and PCs (66; 101; 130). Furthermore, in in vivo preparations, application of CSF IgGs resulted in impairment of cerebellar-mediated modulations, as confirmed by various types of tasks, including motor cortex, gait, behavioral tasks, and blink reflexes (90; 89; 85). Importantly, these pathogenic actions of IgGs from patients with anti-GAD ataxia were elicited by the binding of GA65 with anti-GAD65 antibodies itself; these actions were abolished after the absorption of anti-GAD65 antibodies using recombinant GAD65 (65), and anti-GAD65 antibodies elicited no actions in slices from GAD65 knockout mice where inhibitory transmissions were mediated compensatorily by GAD67 (85). Studies using monoclonal antibodies showed that these actions were epitope-dependent (89; 85). The b78 monoclonal antibody with epitope-specificity in cerebellar ataxias and stiff-person syndrome has pathogenic actions, whereas another monoclonal antibody has no such actions in type 1 diabetes mellitus (89; 85). Furthermore, the actions of CSF IgGs were different even between cerebellar ataxias and stiff-person syndrome, including impairments of exocytosis in the former and a decrease in GABA synthesis in the latter (89). Because the identification of disease-specific anti-GAD65 Abs is complicated by the conformational nature of many of these epitopes, epitope mapping was analyzed by competition assay using human monoclonal Ab, not peptides and deletion mutants (97). Consistent with physiological data, the recognition of b78-defined epitope was different among anti-GAD65 antibodies in cerebellar ataxias, stiff-person syndrome, and epilepsy (89). Notably, low titer anti-GAD antibodies has no pathogenic actions (97).

Internalization and dissociation of GAD65 with vesicles. Anti-GAD65 antibody is internalized, presumably during exocytosis or endocytosis (59; 85). Because GAD65 is attached on the cytosolic face of vesicles, GAD65 might be temporally exposed during exocytosis, providing a chance for binding with antibodies (113; 16). However, the exact mechanisms are still not clear. Anti-GAD65 antibody is reported to impair the association of GAD65 with vesicles, resulting in deficits in GABA packaging into vesicles and shuttling of vesicles to the release sites (85).

From functional disorders to cell death. A decrease in GABA release attenuates the spill-over GABA-induced presynaptic inhibition on glutamate release from neighboring parallel fibers, resulting in a major imbalance between GABA and glutamate and excitotoxicity (95). Consistently, one autopsy study showed the complete loss of Purkinje cells (67).

Clinical correlation. Previous studies showed that some patients respond well to immunotherapies and that the clinical improvement in cerebellar ataxia correlates well with the fall in Ab titers (93; 100). These findings indicate that Ab titers better reflect functional disorders rather than cell death. The peak cerebellar ataxia might be influenced by secondary cell-mediated mechanisms.

Taken together, further experimental studies are needed to determine the access route of antibodies to GAD and to identify if the epitope of antibodies intrathecally synthetized have distinct syndrome-specificity (36). Furthermore, these studies do not exclude an involvement of cell-mediated mechanisms. Infiltrates of CD8+ cells are observed in stiff-person syndrome (68). GAD-specific T cells have been observed in the blood and CSF of patients with stiff-person syndrome and cerebellar ataxias (20). It should also be considered that patients may react differently to immune aggressions.

The aim of any induction therapy is to minimize cerebellar ataxias, which includes various immunotherapies, ranging from intravenous immunoglobulins, glucocorticosteroids, immunosuppressants, plasmapheresis, and rituximab, either alone or in combinations (93; 98; 36). Furthermore, maintenance therapies (eg, oral prednisolone, intravenous immunoglobulins, immunosuppressants, or rituximab, alone or in combination) are also recommended to prevent relapse (93; 98). Because anti-GAD65 ataxia exhibits a chronic time course, the efficacy and therapeutic dose of prednisolone should be monitored carefully to avoid side effects (98). In this regard, it seems there are no significant differences in the response to each type of the above immunotherapies (93; 98). A decrease in anti-GAD65 antibody level during the treatment can be used as therapeutic index (98). Evidence suggests that prognosis is better in the subacute type than in the chronic type of anti-GAD ataxia (02).

Interestingly, 35 of the 50 patients with anti-GAD antibody (70%) had evidence of gluten sensitivity. Half of them showed therapeutic response to a gluten-free diet, suggesting considerable overlapping between gluten ataxia and anti-GAD ataxia (56).

Primary autoimmune cerebellar ataxia (PACA). Despite their autoimmune nature (eg, subacute onset, autoimmune disease history, predominant gait ataxia, association of autoantibodies, and good benefits from immunotherapies), some cerebellar ataxias do not fit into the above categories. In these cases, the autoimmune condition is considered within the spectrum of primary autoimmune cerebellar ataxia (PACA) (92; 39), which was originally proposed by Hadjivassiliou and colleagues (42). Notably, patients with primary autoimmune cerebellar ataxia (PACA) predominantly show HLA type DQ2, which predominates in certain autoimmune diseases (celiac disease, gluten ataxia, type1 diabetes mellitus, stiff-person syndrome, autoimmune thyroid disease, and autoimmune polyendocrine syndromes) (52). In primary autoimmune cerebellar ataxia (PACA), no definite antigens are known to trigger the autoimmune insult in the cerebellum. Various types of less well characterized autoantibodies are associated in some patients (Table 1), and immunohistochemical studies show a variety of staining patterns (60% of the patients) (42), suggesting heterogeneous etiology in primary autoimmune cerebellar ataxia. It is not certain whether the presence of these antibodies is related to a specific etiology or such ataxia remains to be under the umbrella of primary autoimmune cerebellar ataxia (52).

Patients diagnosed with primary autoimmune cerebellar ataxia usually develop clinically evident cerebellar ataxia in their early 50s (52). The cerebellar ataxia, mainly gait ataxia, exhibits a slowly progressive course, which is not as slow as in degenerative cerebellar ataxias. Although some patients who show subacute clinical course have been misdiagnosed with postinfectious cerebellitis, they do not show self-limiting improvement, which characterizes postinfectious cerebellitis. MRI at presentation can be normal or show vermian atrophy, whereas MR spectroscopy can show preferential and sometimes exclusive involvement of the vermis at early stages (45). The proposed diagnostic criteria stress three stages: (1) exclusion of immune-mediated cerebellar ataxias with a known trigger factor (eg, paraneoplastic cerebellar degeneration, gluten ataxia, percutaneous coronary intervention) and exclusion of immune-mediated cerebellar ataxias with characterized pathogenic autoantibodies such as anti-mGluR1 and dipeptidyl-peptidase-like protein 6 antibodies; (2) satisfying at least two of the following conditions: pleocytosis and/or oligoclonal bands in CSF, history and/or family history of autoimmune diseases, and presence of autoantibodies showing autoimmunity (these can be against any organ, eg, thyroid); and (3) exclusion of alternative causes for ataxia (45).

Due to the heterogeneous nature of this entity, there has been no reliable single diagnostic marker. One study shows the Hu-like immunoreactivity using a commercial indirect immunofluorescence assay, originally developed for the detection of well-characterized paraneoplastic antibodies (monkey cerebellum slides and anti-human IgG FITC conjugated antiserum: Inova Diagnostics) (57). This commercially available kit might provide additional diagnostic aid for primary autoimmune cerebellar ataxia.

A small number of studies described more beneficial effects of immunotherapies (intravenous immunoglobulins, prednisolone, plasmapheresis, or rituximab) in patients at the subacute phase (four of six patients) than in patients at the chronic stage (nine of nine patients) (129). Another retrospective study based on 118 patients with immune-mediated cerebellar ataxias (55 patients with nonparaneoplastic immune-mediated cerebellar ataxias) showed that 55 of these patients responded well to immunotherapies, and the improvement was more prominent in nonparaneoplastic patients (72). The progression to wheelchair dependence was faster in patients with neuronal nuclear or cytoplasmic antibody than those positive for plasma membrane protein antibody. MR spectroscopy of the cerebellum seems to be an important tool in monitoring the response to immunotherapies (52). Notably, a systematic study on 22 patients with primary autoimmune cerebellar ataxia revealed that mycophenolate had therapeutic benefits in SARA score, which was associated with an improvement in MR spectroscopy (an increase in NAA/Cr ratio of the cerebellar vermis) (49).

Primary autoimmune cerebellar ataxia likely accounts for patients with so-called Hashimoto encephalopathy. In the original classification, Hashimoto encephalopathy was classified as an independent entity (82; 88; 01). However, the association with other autoantibodies (eg, low titer anti-GAD antibodies or anti-gliadin antibodies) and heterogeneous physiological actions in the CSF and the absence of a plausible pathogenic mechanism by the thyroid antibodies make a single entity unlikely (102). These patients, however, show a good response to corticosteroids (82 88; 01).

Latent autoimmune cerebellar ataxia. The concept of latent autoimmune cerebellar ataxia (LACA) was introduced, drawing parallels with latent autoimmune diabetes in adults (LADA) (140).Type 1 diabetes is defined by the immune-mediated destruction of pancreatic Langerhans' β-cells by an immune mechanism, leading to insulin secretion failure. However, a subset of this population exhibits a gradual decline in insulin secretion and initially ambiguous autoimmune characteristics, complicating the diagnosis of type 1 diabetes. This subset has been classified as latent autoimmune diabetes in adults.

In a similar vein, among latent autoimmune cerebellar ataxias, a subset has been empirically observed to follow a slow progression or present ambiguous autoimmune manifestations, such as well-characterized autoantibodies. Some patients exhibit only mild instability and eye movement disturbances, without overt cerebellar ataxias, and with poorly-characterized autoantibodies. Consequently, these patients do not meet the diagnostic criteria primary autoimmune cerebellar ataxia. However, their symptoms improved with immunotherapies, leading to a final diagnosis of immune-mediated cerebellar ataxias.

We, therefore, proposed a clinical spectrum latent autoimmune cerebellar ataxia and suggested the following diagnostic criteria (84):

By definition, latent autoimmune cerebellar ataxia is likely to follow a course of slow progression. Ultimately, the autoimmune mechanisms will affect the cerebellum, resulting in clinical cerebellar ataxias and eventually marked cerebellar atrophy.

Thus, “the notion of LACA is introduced to encourage clinicians to carefully examine the possibility of slow-evolving IMCA, as well as to stress the importance of the early intervention of immunotherapies during a period when there is cerebellar reserve” (84).

Rare etiologies in nonparaneoplastic immune-mediated cerebellar ataxias

Miller Fisher syndrome. In 1932, Collier described a patient with ophthalmoplegia, ataxia, and areflexia and suggested it was a variant of Guillain-Barré syndrome (18). In 1956, Miller Fisher proposed this condition (acute onset of ophthalmoplegia, ataxias, and areflexia syndrome) to represent one distinct entity, based on examination of three patients (28). At present, it is widely accepted that Miller Fisher syndrome is much less common than Guillain-Barré syndrome, though there is a general agreement that it is part of the same spectrum of infection-triggered autoimmunity (52). A large-scale study involving 50 consecutive patients concluded that viral infection (usually respiratory infection) or bacterial infection (usually Campylobacter jejuni) predated the appearance of neurologic manifestations, with a median interval of 8 days (103).

The main disabling clinical features of Miller Fisher syndrome are ophthalmoplegia and ataxia. At the onset, patients generally exhibit diplopia, ptosis, and gait ataxia with only minor sensory symptoms (52). The ophthalmoplegia develops initially as a symmetrical failure of upgaze, followed by lateral gaze, and finally failure of downgaze. The gait ataxia is often prominent. Nerve conduction is generally normal, despite the pathological evidence of axonal or demyelinating sensory neuropathy. CSF examination often shows high protein level. Up to 90% of patients with Miller Fisher syndrome have high serum titers of anti-GQ1b Ab. Because this is not found in patients with Guillain-Barré syndrome and chronic inflammatory demyelinating polyneuropathy, a positive anti-GQ1b Ab is considered a specific marker of Miller Fisher syndrome (92). Furthermore, anti-GQ1b Ab cross-reacts with epitopes present in the liposaccharide of Miller Fisher syndrome associated Campylobacter jejuni strains, suggesting the possibility of molecular mimicry (58).

The pathophysiology of Miller Fisher syndrome-related ataxia has been a matter of debate: whether it is of peripheral or central origin. In support of the peripheral origin, Miller Fisher proposed the involvement of Ia afferents (28). Furthermore, a disparity between proprioceptive information from muscle spindles and kinesthetic information from joints was suggested (81). GQ1b is expressed in large-diameter dorsal ganglion neurons (92). Mapping of the lesions in a group of patients with Miller Fisher syndrome and overlapping Guillain-Barré syndrome using MRI showed enhanced lesions in the spinocerebellar tracts at the level of the lower medulla (135). On the other hand, the involvement of the cerebellum has been confirmed in a number of studies. One study involving 10 patients who were examined with FDG-PET showed hypermetabolism in the cerebellum and brainstem (75). Furthermore, immunocytochemical staining of the cerebellum using sera of patients with Miller Fisher syndrome confirmed the presence of autoantibodies to antigens in the molecular layer (06; 77). Another study that employed MR spectroscopy showed the involvement of the cerebellum during the period of illness followed by complete resolution of the MR spectroscopy changes once the patient fully recovered (118).

In the large-scale study mentioned above, all patients showed full recovery within 6 months (103). In general, Miller Fisher syndrome is a mild and self-limiting disease, thus, requiring no immunotherapy (52). Although some reports described the benefits of corticosteroids, intravenous immunoglobulins, and plasmapheresis, there are no significant differences in the speed of recovery and the final outcome between patients who receive immunotherapies, including intravenous immunoglobulins and plasmapheresis, and those treated conservatively (103).

Anti-VGCC ataxia. The P/Q-type voltage-gated Ca2+ channel (VGCC) is involved in the release of neurotransmitters at cerebellar neurons and motor nerve terminals. Anti-P/Q type VGCC Ab (anti-VGCC Ab) is found in paraneoplastic cerebellar degeneration. A proportion of patients with anti-VGCC Ab and paraneoplastic cerebellar degeneration were reported to develop Lambert-Eaton myasthenic syndrome (35). Anti-VGCC Ab can be present in paraneoplastic ataxia associated with small cell lung cancer but has also been identified in patients with no cancer. One study reported the presence of anti-VGCC Ab in eight of 67 patients who showed chronic cerebellar degeneration (11). However, the VGCC Ab titers were not high, casting doubts on its clinical relevance (138). Good prognosis was reported in these patients with nonparaneoplastic conditions. In both paraneoplastic and nonparaneoplastic conditions, immunotherapies were used, including IVIg, prednisone, and mycophenolate mofetil.

Anti-DPPX ataxia. Encephalitis associated with anti-DPPX is characterized by agitation, mild confusion, hyperekplexia, myoclonus, trunk stiffness, and cerebellar ataxias (08; 04; 132). These clinical features suggest hyperexcitability of the CNS (04), which could be attributed to dysfunction of the potassium channel. Patients often complain of digestive symptoms (ie, pain, diarrhea). Because cerebellar ataxias with myoclonus can be the sole manifestation in this entity, the combination of these manifestations should be examined by measurement of anti-DPPX Ab (52). The response to immunotherapy is generally good, but long-term and aggressive treatment may be required in some patients (132).

Anti-Caspr2 ataxia. In a cohort of 38 patients with anti-Caspr2 Ab-associated encephalitis, 77% of the patients showed three or more core manifestations, including encephalic signs, cerebellar ataxias, peripheral nerve hyperexcitability, dysautonomia, neuropathic pain, insomnia, and weight loss (136). Neoplastic lesions are also detected in some patients (136). CSF examination is usually normal, although cell proliferation or high protein level is sometimes observed (136). MRI is generally normal (136). Half of the patients show full or good recovery after immunotherapy, suggestive of good prognosis (136). On the other hand, another retrospective study identified stereotypes episodes of paroxysmal cerebellar ataxias in five of 20 patients (73). The ataxic episodes, including gait imbalance, limb ataxia, and slurred speech, usually last a few minutes to a few days and usually improve following immunotherapy. Interestingly, the permanent cerebellar ataxias and episodic ataxias are not associated with neuromyotonia or Morvan syndrome but are associated with limbic encephalitis (73; 105). Anti-Caspr2 is classified into these two groups—limbic predominant group and peripheral nerve hyperexcitability-predominant group without overlapping (73; 105).

Anti-LGI1 ataxia. Leucine-rich glioma-inactivated 1 (LGI1) is a secreted neuronal protein known to form a trans-synaptic complex comprising the presynaptic disintegrin and metalloproteinase domain-containing protein 23 (ADAM23), which interacts with VGKC Kv1, and postsynaptic ADAM22, which interacts with AMPA receptors (64). LGl1 is highly expressed in the neocortex and the hippocampus but moderately expressed in the cerebellum (60). LGI1 is a major antigen in autoimmune limbic encephalitis but rarely in immune-mediated cerebellar ataxias. One large-scale study showed that three of 55 patients presented with cognitive impairments, such as amnesia, confusion/disorientation, seizures, mood disorders, and sleep disorders, with a subacute time course (64). None of the patients had malignant tumors. They showed significant improvement in the modified Rankin scores following immunotherapy with IVIg, intravenous/oral corticosteroids, or their combinations. Although another subsequent study identified motor disorders, including cerebellar ataxia in seven of 34 patients, it did not include detailed clinical information (106). One case report described a young adult with predominant gait ataxia associated with disinhibited behaviors and visual hallucinations (128).

Anti-IgLON5 ataxia. IgLON5 is an adhesion molecule widely distributed in the CNS (32). One systematic study that examined the clinical manifestations of anti-IgLON5 disease identified four phenotypes: (1) sleep disorder with parasomnia and sleep breathing difficulty in eight (36%) patients; (2) bulbar syndrome, including dysphagia, sialorrhea, stridor, or acute respiratory insufficiency in six (27%) patients; (3) syndrome resembling progressive supranuclear palsy (PSP-like) in five (23%) patients; and (4) cognitive decline with or without chorea in three (14%) patients (32). The condition was characterized by the association of HLA-DRB1*10:01 and HLA-DQB1*05:01 alleles. The main clinical feature of anti-IgLON5 ataxia is gait instability. Although disequilibrium was documented as the reason for the instability, the exact mechanism remains unclear. The gait instability was attributed to cerebellar dysfunction in some patients (31). Notably, postmortem examination showed a novel neuronal tauopathy predominantly involving the hypothalamus and brainstem tegmentum (115). The prognosis was poor. No association with malignancy was documented. Only 10% of the patients demonstrated mild and transient improvement following immunotherapy with corticosteroids and IVIg (32).

Anti-NMDA ataxia. The association of cerebellar ataxia with anti-NMDA antibody is very rare. An anti-NMDA antibody-positive 3-year-old boy with chronic cerebellar ataxias and atonic seizure was reported (91). Another young adult patient with teratoma-related opsoclonus-myoclonus syndrome was subsequently reported (134). The presence of these antibodies may simply reflect multiple neuronal autoimmunity rather than being responsible for the ataxia.

Anti-Gluk2 ataxia. Antibodies against the glutamate kainate receptor subunit 2 (GluK2) have been described in eight patients with predominant cerebellar manifestations. The neurologic symptoms had a subacute progression (<  6 weeks). Four patients with a median age of 19 years presented with prominent clinical manifestations compatible with cerebellitis, including opsoclonus in one and two who developed obstructive hydrocephalus (79). Four older patients developed a more diffuse encephalitis, with limb or gait ataxia in two. MRI studies available in seven patients showed multifocal T2-fluid-attenuated inversion recovery (FLAIR) abnormalities in the cerebellum in four. All patients had CSF pleocytosis. Three of seven patients who received immunotherapy had partial or full recovery. An active tumor (one relapsing Hodgkin lymphoma and one ovarian teratoma) was diagnosed in two patients.

Anti-mGluR1 ataxia. The association of anti-mGluR1 Ab with cerebral ataxias was reported initially in two patients with Hodgkin lymphoma (124) and one patient with prostate adenocarcinoma (63). On the other hand, the association of anti-mGluR Ab with cerebellar ataxias was also described in nonparaneoplastic conditions (127). One large-scale study showed the association of neoplasm in 11 of the patients (127). The main neurologic manifestations were subacute cerebellar gait and limb ataxias, sometimes associated with extra-cerebellar symptoms, such as behavioral changes (irritability, apathy, mood, personality change, psychosis with hallucinations, and catatonia), cognitive changes (memory problems, executive functions, and spatial orientation deficits), or dysgeusia. Seizures were uncommon. MRI at onset mostly showed normal or abnormal findings such as T2/FLAIR hyperintensities or leptomeningeal gadolinium enhancement in a few patients. However, as the disease progresses, the patients showed cerebellar atrophy on MRI. Notably, 40% of the patients showed significant improvements or complete resolution of symptoms to immunotherapies including IVIg, steroids, mycophenolate mofetil, cyclophosphamide, and rituximab alone or in combinations. In consistence with this good prognosis, CSF from the patients decreased mGluR1 clusters in cultured neurons, suggesting that anti-mGluR Ab elicits synaptic dysfunctions.

Anti-glycine R ataxia. Glycine receptors are mainly distributed in the spinal cord, brainstem, and cerebellum. Autoantibodies toward GlyR were first reported in 2008 in a single patient with progressive encephalomyelitis, rigidity, and myoclonus (PERM) (62). The clinical features of PERM in this patient included stiff-person syndrome (characterized by stiffness of the axial and lower limb muscles), brainstem signs, hyperekplexia (brainstem myoclonus or excessive startle), and other neurologic defects. Cerebellar ataxia is one of these diverse clinical features. In a subsequent systematic study of 45 patients with anti-glycine R Ab, limb and gait ataxias were identified in 13% of the patients (12). Furthermore, 80% of the patients had nonparaneoplastic conditions. Patients with anti-glycine R Ab-associated disease generally show good response to immunotherapy, including different combinations of intravenous methylprednisolone, oral prednisone, IVIg, and plasma exchange.

Anti-MAG ataxia. Myelin-associated glycoprotein is a glycoprotein specific to Schwann cells in the peripheral nervous system; it is important in the maintenance of the myelin sheath. The association of anti-MAG Ab with distal demyelinating neuropathy is well established in patients with monoclonal gammopathy of unknown significance (141). On the other hand, in the CNS, myelin-associated glycoprotein also plays a role in the maintenance of myelin integrity and inhibition of regeneration of cerebellar neurons (23). Cerebellar ataxia associated with anti-MAG Ab was described in five patients (141). All patients had IgM gammopathy, and four of the five showed clinically evident neuropathy. One patient showed chronic cerebellar ataxias alone (141). Anti-MAG ataxia responds well to rituximab, and such response is associated with improvement in MR spectroscopy (141).

Autoimmune GFAP astrocytopathy. Glial fibrillary acidic protein (GFAP) is the main intermediate filament in mature astrocytes and a component of their cytoskeleton (108). Autoimmune GFAP astrocytopathy is often associated with other autoimmune diseases, such as type 1 diabetes, thyroiditis, or even other types of autoimmune encephalitis (NMDA encephalitis) (27; 78). Paraneoplastic conditions were identified in a third of the patients (139), and a history of upper respiratory tract infection was reported in 40% of the patients (78). Autoimmune GFAP astrocytopathy is characterized by fever, headache, convulsions, delirium, meningism, loss of visual acuity, and ataxia (27; 78). Cerebellar ataxia was described as accompanying meningoencephalomyelitis (40%). Atypical patients with progressive cerebellar ataxia, proximal myoclonus, and bulbar symptomatology have also been reported (108). CSF studies showed inflammatory changes (monocytic pleocytosis), high protein, and low glucose levels (27; 78). Brain and spinal cord MRI studies showed nonspecific findings. Intravenous methylprednisolone is usually effective in acute treatment (78), although some patients also require additional treatment, such as plasma exchange or IVIg. Maintenance therapy using mycophenolate mofetil, azathioprine, or rituximab is necessary in 20% to 50% of the patients in order to prevent relapse (78).

CLIPPERS. Chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids (CLIPPERS) is characterized by marked perivascular T cell inflammation, mainly in the pons with compatible perivascular gadolinium enhancement on MRI, which responds to corticosteroids (112; 26). The lymphocytes, mainly CD4-dominant lymphocytes, infiltrate the perivascular area and are more diffuse in the white matter. Such infiltration is mostly observed in the pons and adjacent rhombencephalic structures, such as the cerebellar peduncles, cerebellum, medulla, and the midbrain (112; 26).

A review of 56 reported cases showed considerable differences in clinical manifestations (26). The condition mainly affects males (67% of the patients), aged 13 to 86 years (mean age at onset: 52.4 years) (26). The patients show subacute onset of varying initial features related to the brainstem pathology, frequently including pancerebellar ataxias, dysarthria, dysphagia, dysgeusia, oculomotor abnormalities, altered facial sensation, facial nerve palsy, vertigo, pyramidal signs, and sensory disorders. In contrast, fever, loss of consciousness, or meningism are rare. The clinical course seems to be relapsing-remitting in nature. MRI shows characteristic changes that reflect perivascular lymphocyte infiltration in the pons and peripontine lesions. The hallmark feature is multiple "punctate" or "curvilinear" gadolinium-enhancing lesions, resulting in "peppering" of the pons with or without peripontine lesions. CSF examination shows either normal or mild-to-moderate rise in protein level with mild increase of white cells.

There are no specific serum or CSF biomarkers for CLIPPERS. MRI compatible CLIPPERS could also result from various pathologies. Thus, it is uncertain whether CLIPPERS is one etiology or a syndrome of heterogeneous etiologies.

Early intervention with corticosteroids improves outcome (26). It is recommended that the initial treatment with intravenous methylprednisolone is followed by maintenance immunotherapy using the combination of oral prednisolone and corticosteroid-sparing immunosuppressants. Because withdrawal of corticosteroids results in disease exacerbation, long-term maintenance therapy is required (26). Steroid-sparing agents, such as cladribine, need to be assessed in a specific trial (131).

IgG4 disease-related ataxia. IgG4-related diseases are characterized by the infiltration of lymphocytes and plasma cells that secrete IgG4, leading to tissue fibrosis and destruction. These diseases manifest in various phenotypes, including autoimmune pancreatitis, sclerosing cholangitis, sialadenitis, retroperitoneal fibrosis, interstitial nephritis, and arterial inflammation, which can result in aortitis and large vessel vasculitis. Pachymeningitis is recognized as a neurologic manifestation. It can present with headaches or cranial nerve symptoms. Additionally, the disorder may involve the pituitary gland. Generally, corticosteroids have been found to provide therapeutic benefits.

A patient with ataxia also presenting with frontotemporal lobe dementia was reported (41). The patient did not experience headache but had complications due to inflammation of the kidneys and large vessels. A biopsy from the perinephric tissue confirmed IgG4 disease. MRS revealed a significantly low NAA/Cr area ratio in the vermis, suggesting that the cerebellum could be a target of IgG4-related diseases (41). Treatment with corticosteroids and mycophenolate improved the ataxias, but the patient subsequently developed pericardiac effusion and died from complications of this.

Myoclonus and cerebellar ataxias associated with COVID-19. Most frequently, severe acute respiratory syndrome virus 2 (SARS-CoV-2) causes fever and respiratory symptoms. However, distinct neurologic manifestations associated with SARS-CoV-2 infection have been recognized as potential complications since the beginning of the COVID-19 pandemic. Fatigue, headache, impairments of smell and taste, dizziness, impaired consciousness, anosmia/ageusia, Guillain-Barré syndrome, and cerebrovascular disease are most frequent. Ataxia is a less frequent complication.

One systematic review examined the association of myoclonus and cerebellar ataxia (13). The authors identified 51 patients from surveys on reports published from November 1, 2019 to December 6, 2020. The mean age was 59.6 years, ranging from 26 to 88 years, and 21.6% were female. The median latency between COVID-19 symptoms and myoclonus/cerebellar ataxias was 13 days. Myoclonus and cerebellar ataxias had an acute onset, usually within one month of COVID-19 symptoms. Among these 51 patients, 23.5% (12 out of 51) of cases had myoclonus and cerebellar ataxias, 35.3% (18 out of 51) of cases had myoclonus without cerebellar ataxias, and 41.2% (21 out of 51) of cases had cerebellar ataxias. Myoclonus was multifocal or generalized, activated by action in 56.7% (17 out of 30) cases and by sensory stimuli in 46.7% (14 out of 30) of cases. Myoclonus and cerebellar ataxias were concurrently associated with other neurologic symptoms, including cognitive changes (45.5%) or a Miller Fisher syndrome variant (21.2%).

MRI studies were generally unremarkable. One case showed fluid attenuated inversion recovery (FLAIR) hyperintensities in the cerebellar hemispheres and vermis and cerebellar leptomeningeal enhancement, and one case showed FLAIR and diffusion-weighted imaging (DWI) hyperintensities involving the white matter of the middle cerebellar peduncles. Anti-neuronal Ab testing was mostly negative. One case showed the association of Abs toward Purkinje, striatal, and hippocampus cells. Eighty percent (24 out of 30) of cases reported improvement or resolution of these neurologic symptoms within 2 months, either spontaneously or with antiepileptic therapies and immunotherapies, including methylprednisolone and IVIg.

Myoclonus and cerebellar ataxias are attributed not to the direct invasion of SARS-CoV-2, but immune-mediated mechanisms (along the lines of post-infectious cerebellitis), due to the absence of SARS-CoV-2 RNA or intrathecal SARS-CoV-2-specific antibody production in most of the patients (117; 137), and the benefits of methylprednisolone. In addition, hyperalbuminorrachia and increased astroglial protein S100B levels were suggestive of blood-brain barrier (BBB) dysfunction (109).

Cerebellar ataxia induced by immune check point inhibitors. Immune-checkpoint inhibitors, such as a PD-1 inhibitor, are currently considered effective for treating several cancers, such as melanoma, lung, urogenital, and gastrointestinal cancers (34; 122). However, discrete neurologic complications have been reported in up to 12% of the patients, including those with myasthenia, Guillain-Barré syndrome, and encephalitis (133). Notably, paraneoplastic cerebellar degenerations has also been reported. Thus, the use of immune-checkpoint inhibitors appears to increase the risk of paraneoplastic cerebellar degenerations, especially in patients with the types of cancer frequently associated with paraneoplastic cerebellar degenerations (for example, small cell lung cancer) (34; 24).

LTDpathies. In the cerebellar circuitry, the combined activation of parallel fibers (PFs) and climbing fiber results in long-term depression (LTD) of the excitatory synapses between PF-Purkinje cells (PC) (94). LTD represents a form of experience-dependent plasticity at the synaptic level, and ataxic symptoms in cerebellar patients could be due to a dysregulation of PF-PC LTD. VGCC, mGluR1, and GluR delta were critical molecules in the induction of LTD. As discussed earlier, these autoantibodies are preferentially detected in immune-mediated cerebellar ataxias. Notably, these three subtypes show common clinical features, including good prognosis with no or mild cerebellar atrophy in nonparaneoplastic syndrome, suggesting functional disorders of the cerebellar cortex. Thus, autoantibodies are implicated in blocking the induction of PF-PC LTD. The authors propose the novel concept of LTDpathies involving characteristic PF-PC LTD-related elementary dysfunction (94).

Summary of differential diagnosis

In contrast to paraneoplastic immune-mediated cerebellar ataxias, non-paraneoplastic immune-mediated cerebellar ataxias respond well to immunotherapy. In addition, the cerebellum has capacity for compensation and restoration. Clinicians should not lose this therapeutic window and should provide early treatment during the period when cerebellar reserve exists. When cerebellar ataxias are the sole clinical manifestation, differential diagnosis from degenerative and genetic cerebellar ataxias is important. If the presence of cerebellar ataxia is accompanied by extracerebellar manifestations, such as limbic encephalitis, long tract signs, brainstem-related clinical features, and neuropathy, the autoimmune processes that can damage multiple targets should be recalled (eg, Miller Fisher syndrome, cerebellar ataxias associated with antibodies toward ion channel or related proteins, cerebellar ataxias associated with antibodies toward synaptic adhesion molecules, cerebellar ataxias associated toward transmitter receptors, anti-MAG ataxia, autoimmune GFAP astrocytopathy, and CLIPPERS). Clinicians should keep in mind the motto of "Time is Cerebellum" (100).

Confusing conditions

Nonparaneoplastic immune-mediated cerebellar ataxia is usually characterized by an acute/subacute time course. Thus, clinicians should differentiate from cerebellar stroke, Wernicke encephalopathy, multiple sclerosis, and paraneoplastic cerebellar degeneration.

Although cerebellar stroke typically presents with a sudden onset, some patients do not report an acute onset and visit the hospital only a few days after symptom onset, complaining mainly of dizziness, vertigo, vomiting, or mild instability. Thus, the differential diagnosis of a gradual worsening of symptoms over a period of a few days should also include cerebellar stroke (diagnosed on MR imaging).

Paraneoplastic cerebellar degeneration is sometimes preceded by prodromal clinical symptoms such as nausea, vomiting, and dizziness, resembling viral infection-related disease. Subsequently, patients show gait ataxia, which is followed by pancerebellar involvement. A definite diagnosis is based on (1) confirmation of the presence of cancer, which develops within 5 years of diagnosis of cerebellar ataxia (in most patients the cancer will be detected within the first 2 years), or (2) appearance of well-characterized onconeural antibodies. The use of whole-body PET scan is essential in distinguishing between paraneoplastic and nonparaneoplastic immune ataxias.

On the other hand, when nonparaneoplastic immune-mediated cerebellar ataxia shows chronic or insidious time course, ethanol-induced cerebellar ataxia, toxic-induced cerebellar syndrome (TOICS), and degenerative cerebellar ataxia should be considered.

After checking for an exposure to certain toxic agents, such as ethanol, organic mercury, organic solvent (toluene, thinner), certain medications (phenytoin, lithium, metronidazole), and excluding hypothyroidism, imaging studies will exclude conditions such as Chiari malformation, a tumor, or multiple sclerosis, or they will simply demonstrate cerebellar atrophy. In general, the presence of cerebellar atrophy is not specific to a particular etiology and can be seen in all of the ataxias. Further genetic analysis should be conducted for the differential diagnosis of autosomal dominant cerebellar ataxias (ADCAs) and autosomal recessive cerebellar ataxias (ARCAs). In case of pure cerebellar atrophy associated with the hot-cross ban sign as well as brainstem atrophy, cerebellar variant of multiple systemic atrophy (MSA-C) should be considered. It should be noted that MSA-C can progress very rapidly and mimic immune ataxias.

Associated or underlying disorders

The association with other autoimmune diseases, such as thyroiditis, type 1 diabetes mellitus, pernicious anemia, and Sjogren syndrome is common. The HLA type DQ2 is detected in the majority of patients with gluten ataxia or primary autoimmune cerebellar ataxia.

Diagnostic workup

Clue to diagnosis. Although the etiology is diverse, non-paraneoplastic immune-mediated cerebellar ataxias share common clinical manifestations. The onset of cerebellar ataxias is usually acute/subacute, and it can be chronic or insidious. The presence of other autoimmune disorders in the patient or in first-degree relatives may be relevant. The main ataxic feature is gait ataxia, which hinders standing and steady walking. Other manifestations include variable degrees of limb incoordination, dysarthria, and nystagmus. MRI can be normal or show atrophy mainly in the vermis, depending on the duration of the disease. Pleocytosis or oligoclonal bands are sometimes observed in the CSF (98). Studies have highlighted the potential usefulness of MR spectroscopy as a sensitive biological marker of the response to treatment in patients with cerebellar ataxia (45; 49). The ratio N-acetylaspartate/creatine area is decreased in patients with immune-mediated cerebellar ataxias.

Autoantibodies. Analysis of autoantibodies can be a good tool for the diagnosis of immune-mediated cerebellar ataxias (38; 93; 92; 97; 98; 48; 52) (Table 2).

Table 2. Classification of Autoantibodies in Immune-Mediated Cerebellar Ataxias

Immunohistochemistry, if available, can add further confirmation of the diagnosis. Autoantibodies can be divided into two categories: (1) specific autoantibodies suggestive of specific etiologies, and (2) nonspecific autoantibodies found in other neurologic conditions, including cerebellar ataxias, which provide only possible autoimmune pathomechanisms. The former types of autoantibodies include anti-TG 2, six antibodies for gluten ataxia, anti-Yo, Hu, CV2, Ri, and Ma2 antibodies for primary autoimmune cerebellar ataxia (93; 92; 97; 98; 48; 52). On the other hand, the nonspecific autoantibodies are also subdivided into two subcategories: autoantibodies that are assumed to have pathogenic roles and the nonpathogenic autoantibodies (ie, a diagnostic marker). Autoantibodies to ion channels or ion channel-related proteins (Ca or K channel), glutamate receptors, or GABA synthesis enzyme have been considered to be pathogenic in the development of cerebellar ataxias, some of which have been confirmed in both in vitro and in vivo preparations (17; 30; 80; 69; 70; 71; 111; 110; 104; 116). It should be noted that autoantibodies that impair neuronal excitability or synaptic transmission can elicit other neurologic features, such as limbic encephalitis. We proposed previously a nomenclature based on the pathogenic role of the autoantibody, ie, the pathogenic antibody to be coined to the name of the clinical entity (eg, anti-GAD ataxia) (52).

Some autoantibodies are reported to be associated with cerebellar ataxias only in a few patients. The significance of these autoantibodies remains to be determined.

Management

When autoimmunity is triggered by another condition, priority should be given to the treatment of the underlying condition (for example, gluten-free diet in gluten ataxia and surgical excision of the neoplasm in paraneoplastic cerebellar degeneration) (38; 93; 92; 97; 98; 48; 52). In such cases, subsequent immunotherapies are only necessary if there is evidence of progression despite removal of the trigger. On the other hand, when autoimmunity is not triggered by any underlying condition, immediate immunotherapies are recommended at an early stage.

Induction immunotherapy is provided first to stabilize cerebellar ataxias, followed by maintenance immunotherapy to prevent relapse (93; 98). Various immunotherapies have been used, ranging from intravenous immunoglobulins, glucocorticosteroids, immunosuppressants (in particular, mycophenolate), plasmapheresis, and rituximab, either alone or in different combinations, and the selection is often based on the etiology. To date, however, there are virtually no large-scale randomized studies involving therapeutic strategies (93; 98).

Studies have highlighted the potential usefulness of MR spectroscopy as a sensitive physiological biomarker of the response to treatment in patients with cerebellar ataxias (48; 52). The ratio NAA/Cr area is decreased in patients with cerebellar ataxias, relative to controls, and increases in those who respond to immunotherapy and exhibit improvement of the autoimmune process.

Outcomes

The therapeutic threshold is best illustrated in gluten-free diet therapy in gluten ataxia (93). In a long-term observational study, we examined the outcome of gluten-free diet in 371 patients with gluten ataxia, including 74% with mild ataxia (ability to walk unaided), 16% with moderate ataxia (inability to walk without support), and 10% with severe ataxia (wheelchair use) (93). Strict gluten-free diet was applied for 1 year, which was confirmed by various serological tests that confirmed elimination of gluten sensitivity. Importantly, clinical improvement correlated significantly with the severity of cerebellar atrophy and was evident in patients with mild ataxia. Gluten-free diet halted and stabilized cerebellar ataxias but did not improve clinical symptoms in some patients with severe ataxia.

The pathological process of anti-GAD-ataxia includes transition from functional disorders to cell death (99; 86). Such shift could be applied as basis for restorable and nonrestorable stages. Division of the clinical course into stages leads to the notion of cerebellar reserve, which is defined as the capacity for compensation and restoration from certain damage. The physiological mechanisms of cerebellar reserve rest on specific features of the cerebellum: the presence of various forms of synaptic plasticity and divergent projections of mossy fibers to microzones (ie, redundant afferents in one microzone). Physiologically, predictive controls, a specific function of the cerebellum, is maintained in the early stage of immune-mediated cerebellar ataxias (96). Thus, the aim of any immunotherapy is to halt the etiology progression during the time when cerebellar reserve is preserved (96; 100; 98).