Brain dysfunction and thyroid antibodies: autoimmune diagnosis and misdiagnosis

Hashimoto encephalopathy, also known as steroid-responsive encephalopathy associated with autoimmune thyroiditis, has been defined by sub-acute onset encephalopathy, with elevated thyroid antibodies, and immunotherapy responsiveness, in the absence of specific neural autoantibodies. We aimed to retrospectively review 144 cases referred with suspected Hashimoto encephalopathy over a 13-year period, and to determine the clinical utility of thyroid antibodies in the course of evaluation of those patients. One hundred and forty-four patients (all thyroid antibody positive) were included; 72% were women. Median age of symptom onset was 44.5 years (range, 10-87). After evaluation of Mayo Clinic, 39 patients (27%) were diagnosed with an autoimmune CNS disorder [autoimmune encephalopathy (36), dementia (2) or epilepsy (1)]. Three of those 39 patients had neural-IgGs detected (high glutamic acid decarboxylase-65, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor-receptor and neural-restricted unclassified antibody), and 36 were seronegative. Diagnoses among the remaining 105 patients (73%) were functional neurological disorder (n = 20), neurodegenerative disorder (n = 18), subjective cognitive complaints (n = 14), chronic pain syndrome (n = 12), primary psychiatric (n = 11), sleep disorder (n = 10), genetic/developmental (n = 8), non-autoimmune seizure disorders (n = 2) and other (n = 10). More patients with autoimmune CNS disorders presented with sub-acute symptom onset (P < 0.001), seizures (P = 0.008), stroke-like episodes (P = 0.007), aphasia (P = 0.04) and ataxia (P = 0.02), and had a prior autoimmune history (P = 0.04). Abnormal brain MRI (P = 0.003), abnormal EEG (P = 0.007) and CSF inflammatory findings (P = 0.002) were also more frequent in the autoimmune CNS patients. Patients with an alternative diagnosis had more depressive symptoms (P = 0.008), anxiety (P = 0.003) and chronic pain (P = 0.002). Thyoperoxidase antibody titre was not different between the groups (median, 312.7 versus 259.4 IU/ml; P = 0.44; normal range, <9 IU/ml). None of the non-autoimmune group and all but three of the CNS autoimmune group (two with insidious dementia presentation, one with seizures only) fulfilled the autoimmune encephalopathy criteria proposed by Graus et al. (A clinical approach to diagnosis of autoimmune encephalitis. Lancet Neurol 2016; 15: 391-404.) (sensitivity, 92%; specificity, 100%). Among patients who received an immunotherapy trial at our institution and had objective post-treatment evaluations, the 16 responders with autoimmune CNS disorders more frequently had inflammatory CSF, compared to 12 non-responders, all eventually given an alternative diagnosis (P = 0.02). In total, 73% of the patients referred with suspected Hashimoto encephalopathy had an alternative non-immune-mediated diagnosis, and more than half had no evidence of a primary neurological disorder. Thyroid antibody prevalence is high in the general population, and does not support a diagnosis of autoimmune encephalopathy in the absence of objective neurological and CNS-specific immunological abnormalities. Thyroid antibody testing is of little value in the contemporary evaluation and diagnosis of autoimmune encephalopathies.