Subscribe to eTOC

Hashimoto's Encephalopathy Diagnostic Criteria Don't Predict Treatment Response
Does the Disorder Actually Exist?

Article In Brief

New data challenge the concept of the existence of Hashimoto's encephalopathy, based on findings that show that criteria for the disorder are not specific and do not assist in predicting whether patients will respond to steroids.


Drs. Josep Dalmau (left) and Francesc Graus were principal investigators on the study.

Does Hashimoto's encephalopathy (HE) exist? Maybe not, according to the results of a recent study that asked whether the diagnostic criteria for the presumed autoimmune disorder predict response to steroid treatment.

“The important message arising from this study is that we don't know if there is such an entity as Hashimoto's encephalopathy, and if there is, what its definition and the limits of the syndrome would be,” commented Robert Lisak, MD, the Parker Webber Chair in Neurology at Wayne State University School of Medicine in Detroit, who was not involved in the study.

HE was first described in 1966 in a patient previously diagnosed with Hashimoto's disease, an autoimmune thyroid disease. In the years since, criteria for the diagnosis of HE have centered on a subacute onset of cognitive impairment, psychiatric symptoms, or seizures; normal thyroid status or mild hypothyroidism; serum thyroid peroxidase antibodies (TPOAb) greater than 200 IU/mL; and no other etiologies. Importantly, there should be no neural antibodies detectable in the cerebrospinal fluid.

The discovery of new neuronal or glial autoimmune diseases is one reason experts have begun to question whether many past cases of suspected HE might instead have been one of these diseases.

A final criterion for diagnosis, and one which experts also find problematic, is response to treatment with steroids. Indeed, the first HE patient had little consistent response to prednisone, and the symptoms, which waxed and waned over the course of a year, resolved spontaneously.

“The problem here is that you make the diagnosis of Hashimoto's encephalopathy in retrospect, when the patient has already improved with steroids,” said Francesc Graus, MD, PhD, co-principal author on the new study, along with Josep Dalmau, MD, PhD, both from the August Pi i Sunyer Biomedical Research Institute in Barcelona, Spain, where Dr. Graus is a researcher in the neuroimmunology program.

Dr. Dalmau's laboratory has been a world leader in identifying new neural antibody-mediated diseases, and that led the group to want to look more closely at HE, Dr. Graus said. “We were concerned that quite a number of samples we received from referring physicians had an initial diagnosis of HE, and it was apparent this diagnosis was being used quite frequently,” and often inappropriately, as many patients in fact had neural antibodies that indicated they had a different disease.


Frequency of high serum TPOAb levels (>200 IU/mL) was similar in patients with autoimmune encephalitis without neuronal antibodies and control groups.

Study Design, Findings

To investigate further, they retrospectively selected 117 patient samples from their database, collected between 2011 and 2017, who were negative for neural antibodies. In 43 patients, the referring neurologist considered HE the most probable diagnosis. Dr. Graus and colleagues randomly chose 74 additional patients who met criteria for possible autoimmune encephalitis.

Of these 117 patients, 24 had serum TPOAb>200IU/mL and otherwise met criteria for HE before steroid treatment.

“We wanted to know if these patients responded to steroids,” Dr. Graus said, and whether clinical or laboratory findings predicted that response. They were subdivided into four groups, based on predominant clinical syndrome: subacute cognitive impairment, psychiatric symptoms, limbic encephalitis, and new-onset status epilepticus.

As controls, they chose 205 patients with a variety of autoimmune neurologic diseases, including neuromyelitis optica, multiple sclerosis, and neuronal surface antibody diseases such as NMDA receptor and GABAa receptor antibodies.

The first noteworthy finding was that the frequency of high TPO antibodies was similar between the suspected HE group and each of the control groups. “This indicates that TPO antibodies are not specific for this disease,” Dr. Graus said.

The second finding was that among those with suspected HE, only 31 percent responded to steroids, and neither the antibody level nor the clinical subtype predicted response. “In the rest, despite having criteria compatible with HE, treatment was not effective,” he said.

Dr. Graus noted that a 2006 study from a group at Mayo Clinic showed that one-third of patients with suspected HE who did not respond to steroids had some form of neurodegenerative disease, which was only revealed at autopsy.

“That is telling us that the criteria we are using for the potential diagnosis of HE are not very specific. The message for the neurological community is that first, patients should receive a full panel of neural antibodies—this has to be done, and the results must be negative,” Dr. Graus said. “Second, if you do a trial of steroids, because you have no alternative diagnoses, it should be short. We know from the literature that patients who are true responders have a very quick response to steroids. If you do not see a response in a few days, it is not worth it to keep trying for a long time.”

Expert Commentary

The study “raises the question of whether there is such a thing as HE,” commented Kenneth L. Tyler, MD, FAAN, Louise Baum Endowed Professor and chair of neurology at the University of Colorado School of Medicine in Aurora. “There aren't any validated diagnostic criteria, and the breadth of clinical presentations and lab criteria across patients labeled with it is pretty broad.”

Furthermore, he said, regarding TPO antibodies, “The one marker that seems to be a feature of this condition is actually one of the least specific autoimmune markers that we have. I think it leaves us in this messy spot. I am not sure this syndrome actually exists.”


“You cant define a disease by the presence of an auto-antibody if it doesnt make sense for the pathogenesis. The idea of there being a single entity to explain these cases is what the study authors are questioning, and I think they are right.”—DR. ROBERT LISAK

“We are dealing with a condition the pathogenesis of which is not clear,” added Dr. Lisak of Wayne State, and for which there is no mechanistic link established between circulating TPO antibodies and brain disease.

“In general, when you are defining a condition in which you don't understand the pathogenesis, and the laboratory test that you are actually measuring doesn't clearly relate to what you are describing in the syndrome,” the link is quite tenuous, he said. “You can't define a disease by the presence of an auto-antibody if it doesn't make sense for the pathogenesis. The idea of there being a single entity to explain these cases is what the study authors are questioning, and I think they are right.”

“This is one of the absolutely outstanding groups in terms of what it has contributed to the field of autoimmune encephalopathies,” said Dr. Tyler. “They have been a leader in identifying and characterizing these syndromes, so one always has to pay attention to their work.”

So, does Hashimoto's encephalopathy exist? “I don't know,” Dr. Graus said. “Perhaps it doesn't, but if it does, it certainly can't be defined by high TPO antibodies alone.” However, he added, “we can't forget there is a group of patients who do respond beautifully to steroids. But we have not found a good way to identify them.”


Dr. Dalmau receives royalties from Athena Diagnostics for the use of Ma2 as an autoantibody test and from Euroimmun for the use of NMDA, GABAB receptor, GABAA receptor, DPPX, and IgLON5 as autoantibody tests. Dr. Graus received a licensing fee from Euroimmun for the use of IgLON5 as an autoantibody test and honoraria from MedLink Neurology.

Link Up for More Information

• Mattozzi S, Sabater L, Escudero D, et al. Hashimoto encephalopathy in the 21st century Neurology 2020;94(2):e217–e224.
• Castillo P, Woodruff B, Caselli R, et al. Steroid-responsive encephalopathy associated with autoimmune thyroiditis Arch Neurol 2006;63(2):197–202