Clinical manifestations

Hashimoto encephalopathy may occur in all ages, from young infants to patients older than 80 years. There is a female predominance with a female:male ratio of 4:1.

Presentation and course

A wide variety of clinical presentations for Hashimoto encephalopathy have been described in the literature, making it difficult to present a truly typical case (27; 35). Most patients have a subacute course (days to weeks) of progressive cognitive symptoms and then slow recovery over months. A relapsing-remitting course is not unusual. The hallmark presenting symptom is impairment in level of consciousness or cognitive decline (encephalopathy) typically associated with seizures, psychiatric symptoms (personality changes, mood disorders, hallucinations, delusions, mania, or catatonia), and movement disorders (tremors, ataxia) or focal stroke-like deficits. Seizures may occur at any point in the disease course, and they are a common form of presentation in children (23). Abnormal movements, ataxia, and cranial neuropathies are variably present (27; 35). Some have identified a subgroup of patients presenting with ataxia as the main complaint, with or without encephalopathy, proposing a cerebellar subtype of Hashimoto encephalopathy (19; 10).

Brain MRI, standard CSF analysis, and EEG may show abnormalities, but these are generally not specific. The disease typically rises higher on the differential diagnosis when high-titer serum thyroid peroxidase or thyroglobulin antibodies are found and as infections and other precisely defined causes of paraneoplastic and autoimmune encephalitis are excluded. Thyroid function should be normal or show minor abnormalities, and screening for tumors (as part of an evaluation for paraneoplastic disorders) should be negative.

Patients typically begin to improve within days to a few weeks after starting corticosteroids or other immunosuppressive therapy, but maximal improvement may take weeks to months. Patients may relapse during corticosteroid taper or much later after completing a course of immunosuppressive therapy. Relapses generally involve symptoms similar to the initial attack.

Three large case series provide a perspective on the typical symptoms and findings.

Chong and colleagues reported 85 patients with encephalopathy, abnormal EEG, and high titer antithyroid antibodies (06). Symptoms included stroke-like signs, seizures, myoclonus, psychosis, or a combination of those. Most had proteinorachia with normal CSF white blood cell count, though patients with marked CSF pleocytosis were also reported. Abnormal EEG was nearly universal (present in 98% of patients), with diffuse slowing as the most commonly reported electroencephalographic abnormality. About half had abnormal brain MRI, including focal and diffuse white matter abnormalities. Thyroid function was variable, but the most common finding was subclinical hypothyroidism. As this study was performed in 2003, these patients were not tested for brain-specific autoantibodies such as NMDAR or LGI1. No associated tumors were reported, although systematic screening was not done. Of those treated with corticosteroids, almost all improved, but a relapsing course was reported in most (60%) of patients.

Laurent and colleagues reviewed 251 published cases (22). In this series, the predominant symptoms included seizures, psychiatric symptoms, confusion, speech impairment, myoclonus, gait disorder, and memory impairment. Ten percent had isolated psychiatric disorders. A minority of patients (15%) progressed to coma. Similar to the above study, most (about 80%) of patients had proteinorachia and normal CSF white blood cell count. Eighty percent had abnormal EEG (most had diffuse slowing and about 10% had epileptic activity) and about half of patients had nonspecific white matter hyperintensities on MRI. Of the patients tested for anti-Hu and anti-Yo, all were negative. No other autoantibody testing for known forms of autoimmune encephalitis were reported. Almost all patients (90%) responded to immunosuppressive therapy within 3 months, at least partially. A subset of patients experienced relapses.

Mattozzi and colleagues reported 24 patients with suspected Hashimoto encephalopathy, serum anti-TPO antibodies over 200 UI/mL, and adequate exclusion of neuronal or glial antibodies in serum and CSF (26). An important distinction between this series compared to previous work is that the authors adapted the first diagnostic criteria to a “subacute onset of cognitive impairment, psychiatric symptoms, or seizures”, allowing for the inclusion of patients with psychiatric or epileptic symptoms alone, without encephalopathy, as long as they had high levels (more than 200 UI/mL) of anti-TPO antibodies. In contrast, original diagnostic criteria simply called for “positive” antithyroid antibodies, without any set cut-off value, and only included patients who had some degree of encephalopathy. These modified criteria lead to the inclusion of patients with distinct clinical presentations: (1) psychiatric symptoms; (2) possible autoimmune encephalitis; (3) new-onset refractory status epilepticus-like; and (4) limbic encephalitis. Interestingly, 29% had psychiatric symptoms alone and 25% had epileptic symptoms alone whereas these patients would not have been included earlier studies. Only one patient had positive antibodies against the amino(NH2)-terminal domain of alpha enolase, 27% had mild clinical or subclinical hypothyroidism, and median duration of symptoms was 19 days. Only about 32% had a complete clinical response to corticosteroids. Importantly, no pretreatment factors were identified to predict response to treatment. Finally, the authors used a control group comprised of 205 patients with various neurologic conditions, including multiple sclerosis, neuromyelitis optica spectrum disorder, and cryptogenic new onset refractory status epilepticus, in which they found the same prevalence of anti-TPO antibodies (8%) compared to patients with suspected Hashimoto encephalopathy, calling into question the specificity of these antibodies. They also found a low frequency of antibodies against the amino(NH2)-terminal domain of alpha enolase in both patients with suspected Hashimoto encephalopathy and controls (one in 24 and one in 13, respectively).

Clinical vignette

A previously well 26-year-old man with a prior history of Addison disease and Hashimoto thyroiditis presented with a series of three unprovoked nocturnal generalized convulsions over 5 days for which he was started on levetiracetam. Over a 2-month period he then developed progressive cognitive impairment with disorientation, working memory issues, and fatigue. After a cluster of three additional generalized convulsions, he was admitted to the hospital.

Brain MRI showed increased T2 signal in the medial temporal lobes. Malignancy screening with testicular ultrasound and with CT of the chest, abdomen, and pelvis was negative. EEG showed intermittent diffuse slowing. Anti-TG were mildly elevated and anti-TPO were highly elevated. Thyroid hormone levels and TSH were normal. CSF analysis showed normal cell counts and normal protein but had 17 oligoclonal bands. Infectious studies and studies for antibodies associated with autoimmune encephalitis (NMDAR, LGI1, Caspr2, AMPAR, GABA-B receptor, GABA-A receptor, GAD65, Ma2, etc.) were negative.

Treatment with intravenous methylprednisolone (1 g daily during 5 days) followed by a slow taper of oral prednisone was initiated. His mental status began to improve within days of starting treatment. Over the following 8 months, his seizures were gradually controlled, and his cognition and energy returned to normal except for a mild deficit in learning new information. He returned to work. During corticosteroid taper, he had two periods with an increase in seizures, which responded to an increase in prednisone followed by a slower taper. Follow-up brain MRI studies showed resolution of the abnormal changes.

Biological basis

Etiology and pathogenesis

The cause of Hashimoto encephalopathy remains unknown, but it is thought to be autoimmune in etiology. This is based on the observations that the disease has a female predominance, CSF analysis often shows indirect signs of inflammation (elevated cell counts, proteins oligoclonal bands, and IgG index), a relapsing and remitting course is often observed, it frequently occurs in the context of Hashimoto thyroiditis and other autoimmune disorders, and generally responds to immunosuppressants (19). Proposed mechanisms include cerebral vasculitis and antibody-mediated encephalitis.

Immunology. Antithyroid antibodies are not thought to directly interact with brain tissues. Indeed, antithyroid antibodies are rather nonspecific, as they are found in 13% of the general population (17), they are detected in much lower levels in CSF than in serum in patients with and without autoimmune thyroid disease (18), and their titers do not correlate with the severity of neurologic disease (15). These factors make a direct effect of the antithyroid antibodies on brain tissue unlikely. Nonetheless, antithyroid antibodies could represent a bystander for pathological antibodies that have yet to be discovered. Mane-Damas and colleagues described a case of pediatric Hashimoto encephalopathy where the patient’s serum applied on rat brain resulted in a hippocampal staining pattern (24). This pattern persisted after the washout of anti-TPO antibodies, suggesting the presence of other unspecified antibodies targeting a neuronal antigen. Blanchin and colleagues, on the other hand, observed direct binding of anti-TPO antibodies to human astrocytes (03), suggesting a pathogenic role for antithyroid antibodies in the disease. More recently, Benvenga and his colleagues found 27 and 47 CNS-expressed proteins that share homology with Tg and TPO, respectively, further supporting this hypothesis (02).

The discovery of defined CNS autoantibodies to NMDAR, AMPAR, LGI1, and other targets has provided convincing explanations for additional patients who might previously have been diagnosed with Hashimoto encephalopathy. Additional specific brain autoantibodies have been discovered since then (21). Other cell-based autoimmune mechanisms may also exist. See Dalmau and Graus’ article on autoimmune encephalitis for discussion of these disorders (07).

Autoantibodies to the amino terminus of alpha-enolase have been reported by some authors to be present in the sera of patients with Hashimoto encephalopathy but not in healthy controls (12; 34; 20; 25). A minority of patients with Hashimoto thyroiditis also had these antibodies. However, Mattozzi and colleagues argued that antibodies to alpha-enolase are not specific to Hashimoto encephalopathy because they have also been detected in patients with limbic encephalitis (due to known neuronal antibodies and in cases of Creutzfeld-Jakob disease) (14; 20; 26).

Furthermore, using flow cytometry in a cohort of 22 patients with Hashimoto encephalopathy, Pfeuffer and colleagues found a persistent increase in CD4+ T cell activation and CD4/CD8 T cell ratio in the patients with a relapsing disease course, but not in those with a monophasic illness (29). They suggested that persistent intrathecal CD4 T cell expression could be considered as a potential biomarker to identify patients at risk for refractory disease.

Pathology. Though the literature on the pathology of Hashimoto encephalopathy is scant, the few available pathology reports point toward a vasculitic process. Indeed, a “mild lymphocytic infiltrate around venules and arterioles” and a “perivascular cuffs of lymphocytic cells” were observed in two patients with suspected Hashimoto encephalopathy (09; 06). Diffuse gliosis and perivascular lymphocytic infiltration were found in the brain biopsy of a tumor-like lesion in a patient with suspected Hashimoto encephalopathy and on autopsy of a pediatric case of graft-versus-host-disease with suspected Hashimoto encephalopathy (31; 33). In our opinion, the majority of these cases did not meet all diagnostic criteria for Hashimoto encephalopathy and, therefore, their true correlation with the disease remains uncertain.

Epidemiology

In the largest series, most (70% to 80%) patients were female (06; 22). There is a broad age range, from young children to the elderly, with a median age in the 40s to 50s. Coexisting thyroid disease is very common. For example, Chong and colleagues found 35% of patients with subclinical hypothyroidism, 22% with euthyroid state, and 20% with overt hypothyroidism. The prevalence of other comorbid autoimmune disorders is also common and has been estimated at 30% in one series (05). The prevalence of the disease was calculated in the last decade to be 2.1 per 100,000 population (11). In an epidemiological study, where presumably false-positive cases of Hashimoto encephalopathy were correctly classified by the detection of neuronal antibodies such as anti-NMDAR, the prevalence decreased to 0.6 per 100,000 population (08).

Prevention

Hashimoto encephalopathy is extremely rare, and there are no known forms of primary prevention. Patients who have had Hashimoto encephalopathy may be at risk for relapses. The duration and nature of immunosuppressive therapy that may decrease the risk of relapse is not known. Patients with repeated attacks of Hashimoto encephalopathy may be treated with longer-lasting or stronger immunosuppressive therapies.

Differential diagnosis

Associated or underlying disorders

Hashimoto encephalopathy is a diagnosis of exclusion and should only be considered after completing a thorough diagnostic workup. Owing to its wide spectrum of subacute clinical manifestations, any toxic, metabolic, infectious, neoplastic, inflammatory, and psychiatric condition should be considered first:

The diagnosis of autoimmune encephalitis should be considered in the presence of specific elements, as previously described by Graus and colleagues (13). These include a subacute onset of working memory deficits, altered mental status or neuropsychiatric symptoms, AND at least one of the following: new focal CNS findings, seizures not explained by a previously known seizure disorder, CSF pleocytosis, oligoclonal bands or elevated IgG index, or MRI features suggestive of encephalitis. Then, specific etiologies of autoimmune encephalitis should be ruled out by using accepted sets of diagnostic criteria and testing serum and CSF for onconeuronal and neuronal cell-surface antibodies (eg, anti-NMDA receptor encephalitis, anti-LGI1 encephalitis, Bickerstaff brainstem encephalitis, acute disseminated encephalomyelitis, etc.). In a series of 144 cases referred to the Mayo Clinic between 2006 and 2019 with suspected Hashimoto encephalopathy, 73% had an alternative nonimmune disorder, and a majority did not have evidence of an underlying neurologic condition (32).

Diagnostic workup

Diagnostic criteria. Graus and colleagues have proposed diagnostic criteria for Hashimoto encephalopathy that were based on those previously proposed by Castillo and colleagues (05; 13):

We believe, like many others, that euthyroid patients should also be included because antithyroid antibodies can be present throughout all phases of autoimmune thyroid disorders (05; 26).

Although response to corticosteroids has been considered important in some prior papers, it is not included in the current guidelines. Mattozzi and colleagues found that only a minority of patients who fulfilled their inclusion criteria for the diagnosis of Hashimoto encephalopathy fully responded to corticosteroids (26).

Workup. Hashimoto encephalopathy is a diagnosis of exclusion. The differential diagnosis includes a number of autoimmune and nonautoimmune causes of encephalopathy, and, as such, appropriate diagnostic workup for these conditions should be performed in the appropriate clinical settings. Complete blood counts, metabolic workup, and renal and liver function tests should be performed. Infectious workup such as blood cultures and urinalysis should be considered. If there is a concern for malnutrition, vitamin B12 and thiamine levels should be checked and empiric repletion should be considered. Thyroid function should be also be assessed and severe hypo- or hyperthyroidism excluded before making a diagnosis of Hashimoto encephalopathy.

We recommend for all patients a lumbar puncture with oligoclonal bands and IgG index, a brain MRI with gadolinium, a baseline EEG, as well as serum antibodies to thyroglobulin and thyroperoxidase. Cancer screening should be considered on a case by case basis, taking into account patient age and sex, clinical presentation, and other risk factors for malignancy.

Lumbar puncture. If a marked pleocytosis is present, other diagnoses should be entertained, such as an infectious etiology. Testing for known causes of autoimmune encephalitis (eg, antibodies to NMDAR, AMPAR, LGI1, GABA-B) should be performed, particularly if the clinical presentation is consistent with limbic encephalitis or shows characteristic features of those disorders. For some of these entities, such as anti-NMDAR encephalitis, CSF testing is preferred as this is more sensitive and specific than serum testing.

Brain MRI with gadolinium. If diffuse white matter changes or gadolinium enhancing lesions are present, other CNS inflammatory or infectious conditions should be considered.

EEG. Especially in the setting of seizures, as some electrographic patterns have been linked to specific types of autoimmune encephalitis (eg, extreme delta brush in anti-NMDA receptor encephalitis). Continuous EEG monitoring could also be considered in order to rule out known seizure syndromes and psychogenic nonepileptic seizures, depending on the presentation.

Serum antibodies to thyroglobulin or thyroperoxidase are required to make the diagnosis. Antibodies are detected by various methods, including radioimmunoassay and chemiluminescent assay. Importantly, administration of IVIG may cause false-positive results (30).

Management

There is no class 1 evidence for the treatment of Hashimoto encephalopathy. Steroid responsiveness has sometimes been considered a key element of Hashimoto encephalopathy, although this may lead to a bias in how cases have been identified for case series. Accordingly, the most common treatment protocols involve corticosteroids given as intravenous pulses of methylprednisolone or prolonged oral prednisone tapers over weeks to months (06; 28; 22; 26). In their case series, Mattozzi and colleagues found that only 32% of patients who fulfilled pretreatment criteria for Hashimoto encephalopathy completely responded to corticosteroids (26). Relapses, especially if corticosteroids are tapered rapidly, are common. IVIG have also been used, but there is less information on the response rates. Plasmapheresis has been employed in some cases, but poses practical problems if prolonged or repeated treatments are needed. The use of steroid-sparing agents, such as cyclosporine and azathioprine, has also been reported. In patients with severe encephalitis, stronger forms of immunosuppressive therapy, such as rituximab and cyclophosphamide, may be considered.

Outcomes

Reported response to treatment in patients with Hashimoto encephalopathy is highly variable across the literature. Because most published cases do not fully meet accepted diagnostic criteria for the disease and because there are no known pretreatment factors that predict response to immunosuppressants, the true prognosis of Hashimoto encephalopathy remains unknown (26). In their series, Mattozzi and colleagues found that most patients seemed to respond at least partly to corticosteroids; out of 19 patients treated with corticosteroids, 14 (74%) had a positive response, including nine (32%) who had a complete recovery. Some patients (33% in Mattozzi and colleagues’ series) tend to present a relapsing and remitting course and some patients require escalation of immunosuppressive therapy, including the use of steroid-sparing agents, IVIG, and plasmapheresis, with variable treatment response. Individual patient characteristics and side effects of corticosteroids and other immunosuppressive therapies, such as increased risk of infections and decreased bone density, should be taken into account when making treatment decisions.