Patient 1
This was a 72-year-old man with a past medical history of hypertension and immune mediated hypothyroidism. He first presented to secondary care with a history of weight loss. His biochemistry profile revealed iron deficiency anaemia, low vitamin D, low folate and normal vitamin B12. CD was suspected and serological testing showed positive tissue transglutaminase antibodies (TTG), gliadin antibodies and endomysium antibodies (EMA). Gastroscopy and duodenal biopsy confirmed villous atrophy, crypt hyperplasia and increased intra-epithelial lymphocytes consistent with CD. He commenced gluten-free diet. Six months later, he presented with worsening mobility and further weight loss and at that point he was noted to have marked ataxia and right facial myoclonus that was speech sensitive. He admitted that he was not adhering to a strict gluten-free diet.
Neurophysiological assessment was undertaken with a Natus Quantum amplifier (Optima Medical Ltd., Guildford Surry, UK) at a sampling rate of 2048 Hz. The recording included a multichannel electroencephalography (EEG) and surface EMG polygraphy (analogue bandwidth 0.01–680 Hz). Data were exported for quantitative EEG/EMG analysis in Spike 2 (version 8.12) software (CED Ltd., Cambridge, UK). Somatosensory evoked potential (SEP) recordings were also performed and the possibility of C-reflexes (cortical reflexes looking for cortical reflex myoclonus) was assessed with surface EMG electrodes from arms and legs based on previously published methods [12, 13]. The recording showed clear evidence on jerk-locked averaging of cortically driven myoclonus (Fig. 1). There was no abnormality in the SEP studies, specifically no evidence of “Giant” SEPs or any suggestion of C-reflexes.
Further serological testing revealed positive GlyR-Abs (using a cell based assay) and negative GAD antibodies. Paraneoplastic antibodies were also negative. MRI spectroscopy was markedly abnormal with N-Acetyl-Aspartate to Creatinine (NAA/Cr) ratio of 0.67 from the superior vermis and 0.61 from the cerebellar hemisphere (normal value > 1). Reduced NAA/Cr is indicative of cerebellar neuronal dysfunction. PET-CT imaging revealed no malignancy. His neurological symptoms continued to worsen. Further immunohistochemistry on the original duodenal biopsy showed clonal changes supporting a diagnosis of refractory CD type 2 but without any evidence of enteropathy associated lymphoma. The patient was started on budesonide and mycophenolate. Levetiracetam was tried for the myoclonus. He continued to deteriorate clinically primarily due to worsening ataxia and eventually required PEG feeding as well as becoming bed bound. A repeat duodenal biopsy showed enteropathy-associated T-cell lymphoma (EATL). At that point he was deemed unfit for chemotherapy and passed away in hospital shortly after from pneumonia. This was 1 year after his initial presentation.
Patient 2
This was a 20-year old man who presented with 8-year history of progressive painful leg spasms and extreme fatigue. He experienced difficulty with running which he used to enjoy. He felt that his muscles were becoming increasingly stiff every time he went for a run. His family also noted cognitive slowing which interfered with his academic performance. Initial investigations by his General Practitioner showed him to be iron deficient. As a result, screening for CD revealed an extremely high level of TTG antibodies, EMA and gliadin antibodies. Duodenal biopsy confirmed CD and he started a gluten-free diet. He was referred to the Gluten sensitivity/neurology clinic in Sheffield UK. Examination revealed ataxic gait, increased tone and stiffness in both lower limbs with hyper-reflexia. He had exaggerated startle response. Imaging of the neural axis did not show any abnormality apart from reduced NAA/Cr ratio of 0.81 from the vermis. The ratio from the right cerebellar hemisphere was 0.9 (normal > 1).
EEG/EMG polygraphy undertaken as described in the previous patient did not show evidence of cortically driven myoclonus from the right thigh on cross correlation analysis. There was no abnormality in the SEP study. Exaggerated startle response was noted following presentation of an unanticipated auditory stimuli. Blink reflex excitability studies (implemented as a semiquantitative assessment of brainstem excitability) while using a short inter-stimulus interval between a conditioning and a test stimulus after 160 ms were performed. Single square pulse electrical stimulation of the supraorbital nerve was delivered at 20–25 mA and a 0.2 s pulse width. The blink reflex study was noted to be abnormal with persistence of a well formed R2 component recorded following the test stimulus. This electrophysiological assessment is used as a semiquiantitative assessment of brainstem excitability [14].
GlyR-Abs were positive and anti-GAD negative. After adopting a strict gluten-free diet he noticed gradual improvement in his symptoms. He was now able to run again and resume his academic career. After 2 years on strict gluten-free diet and normalisation of the gluten sensitivity-related antibodies his glycine receptor antibodies became negative. Repeat electrophysiology showed normalisation of the blink reflex study. Repeat brain image showed improvement of the NAA/Cr ratio from the vermis (from 0.81 to 0.88).