In this patient, the culprit lesion of truncal ataxia was not determined until 7 years after onset of the diseases. During the 7 intervening years, neither clinical nor neurophysiological examinations had demonstrated any findings indicative of peripheral nerve disorder, and conventional MRI had also provided no distinct finding of cerebellar atrophy. Seven years after onset of the disease, SPECT revealed slight reduction in blood flow in the cerebellum, which was left hemisphere dominant, and the images of MRI voxel-based morphometry demonstrated atrophy of the cerebellar cortex in the same area. Although the diagnosis was difficult to establish, cortical cerebellar atrophy was suspected. In addition, autoimmune cerebellar ataxia was suspected based on positive gliadin and thyroid antibodies and therefore the treatment with IVIg and steroid therapy was performed, leading to amelioration of truncal ataxia. Eight years after disease onset, the patient died and an autopsy was performed. The autopsy showed that the number of Purkinje cells in the cerebellum was moderately decreased in a selective manner. Consequently, the cerebellum was found to be the location of the culprit lesion causing truncal ataxia in the patient, leading to a definitive diagnosis of cortical cerebellar atrophy.
Moreover, prior to death, the patient had positive anti-SS-A and anti-SS-B antibodies without drying symptoms and was not thus diagnosed as having Sjögren syndrome. However, infiltration of lymphocytes as well as a decrease in the number of acini in the submandibular gland was found at autopsy. This led to the diagnosis of Sjögren syndrome.
In this patient, autoimmune cerebellar ataxia was suspected, given the findings of positive IgG and IgA antibodies to gliadin, positive anti-SS-A and anti-SS-B antibodies, and positive thyroid antibody, as well as the previous history of Basedow disease. Consequently, the treatment with intravenous immunogloblin and steroid therapy was performed, resulting in amelioration of truncal ataxia. A large number of reports on autoimmune cerebellar ataxia have been accumulated, including gluten ataxia , anti-GAD-antibody-positive cerebellar ataxia [2–4], Hashimoto encephalopathy [5, 6], Sjögren syndrome , and neuro-Behçet disease . This patient was diagnosed as having autoimmune cerebellar ataxia close to gluten ataxia because the patient had simple cortical cerebellar atrophy. This diagnosis was also appropriate given the lack of autopsy findings indicative of hereditary spinocerebellar degeneration such as SCA3 and SCA6 and multiple system atrophy, and because several types of autoantibodies  were found that have been reported previously in association with autoimmune cerebellar ataxia, including anti-gliadin antibody.
Gluten ataxia patients are known to be positive for anti-gliadin antibody. Hadjibassiliou et al. reported that anti-gliadin antibody cross-reacts with epitopes on Purkinje cells from human and rat cerebellum, and in doing so strengthens the impact of circulating antibodies against cerebellar Purkinje cells . Hadjibassiliou et al. also reported post-mortem findings from two patients with gluten ataxia. Examination of the central nervous system revealed patchy loss of Purkinje cells throughout the cerebellar cortex. The cerebellar white matter showed astrocytic gliosis, and a diffuse infiltrate of mainly T-lymphocytes. Substantial perivascular cuffing with inflammatory cells, mainly T lymphocytes with smaller numbers of B lymphocytes and macrophages, was present within the cerebellar white matter and the posterior columns of the spinal cord . In addition, Mittelbronn et al. reported that examination of the cerebellum of a 68-year-old male with gluten ataxia showed an absence of B- or plasma cells but multiple CD8+, suggesting that there are prominent cytotoxic effects in the neuropathogenesis of gluten ataxia . The pathological findings of gluten ataxia in these reports differ from our case in that they show infiltration of inflammatory cells in the cerebellum. As described above, the pathological findings reported previously with gluten ataxia include infiltration of inflammatory cells in the cerebellum, which was not observed in our patient.
Along with gluten ataxia, paraneoplastic cerebellar ataxia and anti-GAD-antibody-positive cerebellar ataxia cause autoimmune damage to Purkinje cells in the cerebellum. A large number of autopsied cases of paraneoplastic cerebellar ataxia have been reported, but often without findings of inflammatory reactions. In the pathological investigation of three patients complicated with paraneoplastic cerebellar ataxia and LEMS, Fukuda et al. reported that a loss of Purkinje cells in the cerebellum was evident but HE staining and immunostaining using CD3, CD4, CD19, and CD20 did not demonstrate infiltration of lymphocytes in the cerebellar cortex . Ishida et al. have also reported in an autopsied case of anti-GAD-antibody-positive cerebellar ataxia that near-complete depression of the Purkinje cells was observed without infiltration of inflammatory cells .
As stated above, only a few autopsies of gluten ataxia have been previously reported, however in both cases, immune therapy was not confirmed to be effective. We considered the diagnosis of autoimmune cerebellar ataxia possible in this case for the following reasons: immune therapy was effective; anti-gliadin antibody, anti-thyroid antibody and anti-SS-A antibody that are suggested to be related to autoimmune cerebellar ataxia were present; the patient's history of Basedow's disease; and the definite diagnosis of Sjögren's syndrome on autopsy. In the event that immune cerebellar neural disorders are preceded by humoral immunity, the pathological finding of poor infiltration of inflammatory cells, as found in this case and in some paraneoplastic syndromes can be explained. In this case, the administering of immune therapies including 2 instances of high-dose immunoglobulin therapy and steroid therapy might have affected the pathological findings.
Although antemortem diagnosis was difficult in the patient's life time, the loss of Purkinje cells that was found on autopsy led to the diagnosis of autoimmune cerebellar atrophy. Where this case differs from others that have been previously reported is in the atypical pathological finding of an absence of an inflammatory response. In the future we can expect an accumulation of autopsy cases and clarification of the corresponding pathological conditions, and this should provide more information on cases of autoimmune cerebellar ataxia, including gluten ataxia and anti-GAD-antibody-positive cerebellar ataxia.