Clinical Characteristics and Management of 50 Patients with Anti-GAD Ataxia: Gluten-Free Diet Has a Major Impact

Glutamic acid decarboxylase (GAD) is the rate-limiting enzyme in the synthesis of the inhibitory neurotransmitter gamma-aminobutyric acid (GABA). GAD is found in both the central and peripheral nervous systems (including the enteric nervous system) as well as in pancreatic beta cells [1]. Antibodies against pancreatic islet cell proteins were first detected in children with insulin-dependent diabetes mellitus (IDDM) and were subsequently characterised as GAD antibodies [2‐4]. Two major types of GAD enzyme exist, GAD65 and GAD67. These catalyse the formation of GABA at different cell locations and different time periods of development. The GAD67 enzyme is widespread in the central nervous system, whilst GAD65 is confined to nerve terminals. GABA is synthesised by GAD67 for neuronal activity unrelated to neurotransmission and synaptogenesis. On the other hand, GAD65 produces GABA for neurotransmission and is required at synapse [5].

The first neurological disease to be linked with GAD65 antibodies was stiff-person syndrome (SPS) [6]. GAD antibodies were subsequently shown to be present in the sera of up to 60% of patients with SPS, 80% of patients with IDDM (at a much lower titre than SPS), in patients with polyendocrine autoimmune syndromes and in some cases of sporadic, otherwise idiopathic ataxia [7‐11].

The first case series of patients with so-called anti-GAD ataxia was published in 2001 [12]. It has since been common practice to include anti-GAD antibody testing in the diagnostic workup of all patients with progressive idiopathic ataxia. The presence of high titre of anti-GAD antibodies should alert the possibility of an immune-mediated ataxia and the need to consider immunosuppressive treatment.

We have previously demonstrated significant overlap between anti-GAD-associated neurological diseases and gluten sensitivity [13]. Here, we present our 25-year experience of managing 50 patients with anti-GAD ataxia at the Sheffield Ataxia Centre.

This report is based on a retrospective observational case series of patients regularly attending the Sheffield Ataxia Centre. The South Yorkshire Research Ethics Committee has confirmed that no ethical approval is indicated given that all investigations/interventions were clinically indicated and did not form part of a research study.

We performed a retrospective review of all patients with progressive ataxia and high serological titres of GAD antibodies (defined as > 2000 U/ml, normal < 5). Estimation of anti-GAD was made using a commercial assay (RSR Limited) according to the manufacturers’ instructions. Briefly, the wells are coated with GAD 65, and the samples are added and incubated. GAD65-biotin is added to the wells and incubated. Streptavidin-peroxidase is added to the wells. TMB substrate is added and incubated, and the plate is read at 405 nm (if low values, it is also read at 450 nm). Plate also includes calibrators 2000, 250, 120, 35, 18 and 5 U/mL and positive and negative control samples. The wells are washed in between each stage.

All patients have been seen, and most are still under active follow-up at the Sheffield Ataxia Centre, Sheffield, UK. All of these patients have been extensively investigated for other causes of ataxia including extensive genetic testing using next-generation sequencing [14]. The review included detailed examination of the clinical records and MR imaging including MR spectroscopy (MRS) of the cerebellum. This technique is under regular clinical use at our centre as a monitoring tool for all patients with progressive ataxia, particularly those undergoing therapeutic interventions [15]. In these series of 50 patients, we have only included patients presenting with cerebellar ataxia and excluded patients with SPS, the other group of patients seen in neurology clinics that often have high titres of anti-GAD antibodies.

This report describes our series of patients with anti-GAD ataxia. Unlike previous smaller reports on the subject, all of the patients described here were clinically assessed and managed at the same centre, by the same team, being diagnosed after attending the Sheffield Ataxia Centre [23, 24]. A novel inference is the significant overlap of anti-GAD and GA, as 70% of patients with anti-GAD ataxia had serological evidence of gluten sensitivity and the majority responded to GFD alone without necessitating the use of immunosuppression.

There was a high prevalence of additional autoimmune diseases (90% of patients) and a family history of autoimmune diseases in first-degree relatives in 52%. This suggests that anti-GAD antibodies are a marker of multiple autoimmunity of which the cerebellar ataxia may be one of many autoimmune manifestations.

The association between anti-GAD antibodies, multiple autoimmunity and gluten sensitivity merits close consideration. Italian researchers have noted that the prevalence of additional autoimmune diseases in children with coeliac disease (CD) is significantly lower than in those patients with CD diagnosed in adulthood [25]. They concluded that GFD may reduce the risk of developing additional autoimmune diseases later on in life. This observation echoes our observed reduction in anti-GAD antibodies in patients with anti-GAD-related diseases and gluten sensitivity who go on a strict GFD [13].

Our results also highlight the preferential involvement of the cerebellar vermis, something that is commonly seen in immune-mediated ataxias [26]. The severity of anti-GAD ataxia by comparison to some degenerative (MSA-C) or genetic ataxias (SCA6) and some other immune-mediated ataxias (paraneoplastic cerebellar degeneration) seems, by comparison, mild. Indeed the majority (58%) had mild ataxia at presentation with only 18% having severe ataxia. The mild ataxia was also reflected in the MR spectroscopy measurements and the degree of cerebellar atrophy. However, if patients with anti-GAD ataxia remain untreated, their ataxia is generally progressive, with accumulation of significant disability over time. Treatment benefits most patients, but treatment should be considered at an early stage in an attempt to try and prevent the accrual of irreversible disability. The illustrative case reported here highlights this issue as the patient was not treated for 17 years, partly because the diagnosis of anti-GAD ataxia was made late. Her mild, slowly progressive ataxia seemed insufficient to justifying the use of immunosuppression for many years. This emphasises close clinical and MR spectroscopic monitoring, with a low threshold for treatment which is essential in preventing permanent disability. Of note is the slowly progressive course for years but with sudden and often severe irreversible exacerbations (as illustrated by the case report). This is another justification for early therapeutic intervention.

The pathological role of anti-GAD antibodies in the genesis of this immune ataxia is unclear. Since GAD65 is intracellularly located and is associated with a range of neurological conditions (ataxia, SPS, epilepsy) as well as IDDM, some have argued that anti-GAD65 antibodies have no pathogenic role to play. On the other hand, recent physiological studies in vitro and in vivo have demonstrated that binding of GAD65 by anti-GAD65 antibodies induces loss of GAD65 functions relating to GABA release, with an epitope dependence, leading to the development of cerebellar ataxia [27]. Given these observations, the question still remains as to why is GFD beneficial in those patients with gluten sensitivity and anti-GAD ataxia. The significant overlap between anti-GAD ataxia and gluten sensitivity also raises important considerations for anti-GAD pathogenicity. The response of anti-GAD ataxia to GFD in those patients who are gluten sensitive suggests that gluten sensitivity is part of the underlying pathogenesis. Furthermore, previous work from our group has highlighted a link between neurological manifestations of gluten sensitivity and anti-GAD antibodies, demonstrating reduction of the titre of anti-GAD antibodies after the introduction of GFD [13]. Indeed in 2 of the patients with anti-GAD ataxia and gluten sensitivity reported here, the level of anti-GAD dropped to below 2000 U/ml following GFD (unpublished observation). It is possible that such a drop may have been evident in other patients but the lack of routine quantification of high titres (> 2000 U/ml) by our clinical immunology laboratory does not allow such confirmation. In gluten ataxia, elimination of gliadin and transglutaminase (TG2 and TG6-related antibodies), all of which cross-react with cerebellar tissue, has a beneficial effect [28]. This would also likely be true in the context of patients with anti-GAD ataxia who are gluten sensitive. One possibility is that in these patients there is double insult causing cerebellar damage, one driven by gluten and the other by anti-GAD. Eliminating the first by strict GFD might help the ataxia (as observed here), but perhaps the combined use of both the GFD and immunosuppression may result in better overall outcome. In practice we have reserved the use of immunosuppression for those patients where GFD alone did not improve the ataxia. We have observed that patients with anti-GAD antibodies and gluten sensitivity proportionally are more likely to require additional immunosuppression than those patients with gluten ataxia alone even when both groups are on strict GFD. It is possible that a more effective approach would be to use both GFD and immunosuppression together at the time of diagnosis. The benefits of strict GFD can take up to a year to manifest and require complete elimination of all gluten-related antibodies that can only be achieved by strict adherence to GFD. Re-exposure to gluten causes reactivation of the immune process that results in cerebellar damage, and such dietary indiscretions may set the patient back for several months. These issues are challenging for many patients.

Although detailed neurophysiological assessments have not been done in all patients, there was electrophysiological evidence of cortical myoclonus in only three, two of which had serological evidence of gluten sensitivity and one had IgA deficiency (also suspected as having gluten sensitivity but negative serology due to the IgA deficiency). Gluten sensitivity is a well-recognised cause of cerebellar ataxia with cortical myoclonus and in this instance could have been a synergistic factor leading to the development of cortical myoclonus [18]. We have described a similar phenomenon in a cohort of patients with long-term exposure to lithium who were also gluten sensitive and developed cortical myoclonus [29]. In addition, our patient with the IgA deficiency was found to have evidence of exaggerated physiological acoustic startle responses. This tendency has been previously described in great detail in patients with SPS [30].

The presence of high titre of anti-GAD antibodies should not always preclude the investigation for other causes of ataxia. We have encountered 3 patients (not included in these series) with progressive ataxia, positive anti-GAD antibodies (> 2000 U/ml) but with alternative aetiology for their ataxia. The clinical suspicion of an alternative aetiology was as follows: One patient progressed rapidly and developed autonomic dysfunction fulfilling the criteria for probable MSA-C, a second patient in his 20s had severe global cerebellar atrophy (unusual for anti-GAD ataxia) which lead to genetic testing confirming a pathogenic mutation for ANO10 and a third patient had mild spastic paraparesis and waddling gait which lead to a genetic diagnosis of SPG7.

Whilst it is extremely rare to find such high levels of anti-GAD in genetic ataxias (in our genetic ataxia cohort, the figure for positive anti-GAD was 0.004%) in clinical scenarios where an alternative diagnosis is suspected examination of the CSF may prove helpful in demonstrating the intrathecal production of GAD antibodies thus further supporting the diagnosis of anti-GAD ataxia [12]. Indeed one may argue that the absence of a systematic study of the CSF is a limitation of this report. However, in everyday clinical practice, we found that high serological levels of anti-GAD in the context of an idiopathic ataxia were sufficient to diagnose and treat as anti-GAD ataxia.

Finally, we did not include in these series patients with SPS. In our experience, patients with SPS almost always have a degree of cerebellar ataxia, but their primary presentation is distinct and characterised by axial rigidity and spasms. SPS is also rare and in our experience less common than anti-GAD ataxia. Over the same period of time (25 years), we have encountered 18 patients with classic SPS (another area of our interest) as compared to 50 patients with anti-GAD ataxia. However, there may well be some ascertainment bias given that the patients with ataxia were seen at the Sheffield Ataxia Centre, one of only 2 National Ataxia Centres in the UK.

To conclude, we present here clinical data of a large series of patients with anti-GAD ataxia and highlight a significant overlap between anti-GAD and GA. This has significant therapeutic implications. We share our experience of treating these patients with GFD and/or immunosuppression and propose that early treatment intervention is more likely to be associated with better outcomes and limit permanent disability even if the initial presentation suggests a mild slowly progressive disease.