Article Text


Hashimoto’s encephalopathy mimicking Creutzfeldt-Jakob disease: brain biopsy findings
  1. C P Doherty1,
  2. M Schlossmacher1,
  3. N Torres1,
  4. E Bromfield1,
  5. M A Samuels1,
  6. R Folkerth2
  1. 1Department of Neurology, Partners Neurology, Brigham and Women’s Hospital, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02115, USA
  2. 2Department of Pathology (Neuropathology)
  1. Correspondence to:
 Dr C P Doherty;

Statistics from

A previous report in this journal described seven cases of Hashimoto’s encephalopathy (HE) clinically resembling Creutzfeldt-Jakob disease (CJD).1 Brain biopsies in such cases are rare and have suggested “vasculitis”.2 We contribute a report of rapidly progressive dementia in a patient undergoing brain biopsy before the diagnosis of HE was established, showing features suggesting early spongiform change but with inflammation.

A 57 year old woman was taken to a local hospital following a generalised seizure. She was discharged that night after negative cranial computed tomography and cerebrospinal fluid (CSF) analysis. Within a few days she was noted by family members to be acting strangely and hallucinating. Her doctor found her to be somnolent and rigid without focal neurological findings. Magnetic resonance imaging of the brain showed a questionable increase in gadolinium contrast uptake in a 7 mm area of the left medial frontal cortex. An electroencephalogram (EEG) showed bihemispheric slowing without epileptiform activity. Despite an extensive inpatient evaluation (including biochemical, haematological, endocrine, infectious, autoimmune, and toxic analyses), no cause for the encephalopathy could be found. The patient was then referred to our institution for brain biopsy and further care.

Samples of left frontal cortex showed light microscopic evidence of rare vacuoles abutting neurones, suggesting early spongiform change (fig 1). Glial fibrillary acid protein staining showed moderate gliosis. A few perivascular lymphoid cells and macrophages were present, with microglia scattered throughout the parenchyma; there was no evidence of true vasculitis (inflammation and fibrinoid necrosis of arterial vessels) or of microglial nodules. Because prion disease could not be excluded histologically and because the clinical suspicion of CJD remained high, a tissue block was sent to the National Prion Disease Reference Laboratory (NPDRL).

Subsequently the patient’s condition deteriorated. She became wheelchair bound, akinetic, and mute, with startle myoclonus and prominent frontal release signs. CSF analysis showed increased protein (890 mg/l, reference range < 450 mg/l]). A sample was sent to be tested for the 14–3–3 protein, which is a CNS protein detected in spinal fluid (named after the chemical formula for one of its seven distinct isoforms and two phosphorylated forms) and has been shown recently to have a high sensitivity and specificity (96% and 99%, respectively) for the diagnosis of CJD, which exceeds that of the EEG. It has recently been added to the clinical criteria for the “probable” sporadic form of the prion disease.3 Magnetic resonance imaging repeated with gadolinium and diffusion weighed sequences showed only postsurgical changes. Repeat EEG showed continued bihemispheric slowing with the addition of triphasic waves. The patient had a family history of hypothyroidism but at the time of admission was euthyroid (thyroid stimulating hormone 2.89 mU/l, reference range 0.5–5 mU/l; thyroxine 65.79 pmol/l, reference range 64.5–141.9 pmol/l; triiodothyronine 1.386 nmol/l, reference range 1.078–2.464 nmol/l). Nevertheless, in the absence of a clear cut cause and in view of prior reports of the clinical overlap between CJD and HE, we measured thyroid peroxidase antibodies, which were increased (79 IU/ml, reference range 0–20 IU/ml). An ultrasound of the thyroid showed several hypoechoic areas, fine needle biopsies of which were non-diagnostic. The fact that many euthyroid middle aged women can be shown to have mildly increased thyroid peroxidase antibodies inevitably gave rise to speculation that the finding was incidental. However, in the absence of another obvious cause for the rapidly evolving dementia, and giving due weight to the findings in previous studies that the majority of reported cases were euthyroid, we began steroid treatment for a presumptive diagnosis of HE.1

Within 24 hours the patient had visible neurological improvement and within one week she was mobile and conversant, with reversal of most of her cognitive and motor deficits. Shortly thereafter, the NPDRL reported that the tissue did not show evidence of abnormal prion protein. The CSF did not contain the 14–3–3 protein. The patient was discharged on a slow taper of steroids. Six weeks later she was seen in the outpatient clinic and was considered fully recovered by her husband and her doctors.

Rapidly progressive encephalopathy with myoclonus, extrapyramidal rigidity, and akinetic mutism suggest the diagnosis of prion disease such as CJD. However, the differential diagnosis is wide and includes infections of the CNS, toxins, vitamin B12 deficiency, hypothyroidism, autoimmune or idiopathic cerebrovasculitis, paraneoplastic encephalitis, epileptic pseudodementia, and other rapidly progressive dementias, especially familial forms of Alzheimer disease. HE, a disorder of presumed autoimmune origin in which patients present with altered levels of consciousness, seizures, myoclonus, and extrapyramidal rigidity, has recently been added to this list.1,4 Patients with HE are usually euthyroid, with increased antithyroid antibodies often discovered only after the diagnosis of encephalopathy.1,4

Although presentation with a seizure is unusual in CJD, our patient’s other clinical features of rapid cognitive dissolution, myoclonus, and extrapyramidal rigidity were highly suggestive of prion disease.5 The initial laboratory evaluation, the lack of convincing imaging findings, and the EEG were also consistent with early prion disease.5 In fact, according to World Health Organization criteria, the clinical and EEG features of this case alone already indicated a diagnosis of “probable” CJD.3 Furthermore, the presence of vacuoles on the brain biopsy was very suggestive of the spongiform change (due to swelling of neuritic processes) in early CJD, a diagnostic possibility additionally evoked by the moderate gliosis.3 Although the absence of pathognomonic amyloidal (“kuru”) plaques on a limited biopsy does not exclude a diagnosis of “definite” CJD in a suspected patient, we felt that the features of the case justified immunoassay for prion protein in the tissue and CSF.

The presence of perivascular mononuclear cells, though sparse, is not typical for CJD, and may, in the end, be the greatest supportive evidence of encephalitis as the true cause of the symptoms. The changes are reminiscent of those illustrated by Nolte et al,2 although we strongly disagree with their term “vasculitic”, instead preferring to reserve that term for true necrotising arteritis, which implies a different pathogenesis. Unlike their patient in whom the inflammation was confined to basal leptomeninges in the setting of prior subarachnoid haemorrhage, ours involves parenchymal vessels, with permeation of the cortex by microglia, suggesting a more credible substrate for the seizures and cognitive deficits.

In conclusion, we recommend that patients who, according to World Health Organization criteria, have clinical and EEG features of probable CJD3 be considered for empirical treatment with steroids for HE, pending laboratory results for antithyroid antibodies as well as for 14–3–3 protein. Furthermore, we present additional evidence that HE is an encephalitic process, likely autoimmune given the dramatic response to anti-inflammatory treatment.

Figure 1

Haematoxylin and eosin stained section of cortex (×400) shows vacuoles within or closely adjacent to neuronal cytoplasm (arrows), giving a spongiform appearance. Elsewhere, microglia were scattered throughout the brain parenchyma and a sparse mononuclear infiltrate was seen in perivascular spaces (not shown). Glial fibrillary acidic protein immunostain (not shown) highlighted established gliosis.


View Abstract


  • Competing interests: none declared

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.