By Ulrike W. Kaunzner, MD
Assistant Professor of Clinical Neurology, Weill Cornell Medical Center
A recent multicenter, retrospective review of 393 adult patients given a diagnosis of autoimmune encephalitis noted that 107 patients (27%) were misdiagnosed and ultimately had other diagnoses confirmed, but after a considerable delay.
Dalmau J, Graus F. Autoimmune encephalitis—misdiagnosis, misconceptions, and how to avoid them. JAMA Neurol 2022; Nov 8. doi: 10.1001/jamaneurol.2022.4154. [Online ahead of print].
Flanagan EP, Geschwind MD, Lopez-Chiriboga AS, et al. Autoimmune encephalitis misdiagnosis in adults. JAMA Neurol 2022; Nov 28:e224251. doi:10.1001/jamaneurol.2022.4251. [Online ahead of print].
Autoimmune encephalitis (AE) is a rare neurological disease with a prevalence of 13.7/100,000 people. The disease can occur as early as during the teenage years, and women are affected more often than men. AE presents with a variety of clinical manifestations, including behavioral and psychiatric symptoms, autonomic changes, movement disorders, and seizure-like activity. Symptoms can progress to loss of consciousness.
AE is considered if the workup for other neurological diseases, infections, and psychiatric and toxicological etiologies is negative. Antibodies against neuronal or glial proteins can be found, and there now are more than 20 AEs described. An excellent algorithm for the diagnosis of AE was published in 2016 by Graus et al, which proposes criteria for the diagnosis of possible AE: 1) subacute onset of memory deficits, altered mental status, or psychiatric symptoms; 2) at least one of the following: new focal central nervous system (CNS) deficits, unexplained seizures, cerebrospinal fluid (CSF) pleocytosis, or magnetic resonance imaging (MRI) findings of encephalitis; and 3) exclusion of alternative diagnoses. These AE criteria are prerequisite for the subsequent AE diagnostic algorithm.1
The retrospective study of Flanagan et al was conducted in six academic subspeciality centers over a period of seven years. A total of 393 adult patients with an initial diagnosis of AE were evaluated. Within this cohort, 107 patients (27%) had a misdiagnosis of AE. Out of this misdiagnosed cohort, 77 patients did not fulfill the previously mentioned first and second criteria of possible AE diagnosis. The remaining 30 patients fulfilled these two criteria, but eventually were diagnosed with an alternative diagnosis, and therefore did not fulfill the third requirement.
The reasons for misdiagnosis included insidious onset of symptoms, comorbidities that caused cognitive deficits, functional neurologic disorder, mitochondrial disease, normal MRI, progressive atrophy without signal abnormalities or enhancement, expanding lesions despite immunotherapy, and noninflammatory CSF.
However, the largest contributor of AE misdiagnosis was the overinterpretation of detected antibodies. Auto-antibody tests often had no clinical relevance, or the detected antibody did not correspond to the clinical situation. For example, the diagnostic accuracy for neuronal antibodies varies for serum vs. CSF testing and for the respective assay methodology. Other sources for antibody-related misdiagnosis include nonspecific glutamic acid decarboxylase (GAD), thyroid peroxidase (TPO), and voltage-gated potassium channel (VGKC) antibodies, or anti-N-methyl-D-aspartate (NMDA) antibodies that do not react with brain tissue.
In all 107 cases, the alternative diagnoses eventually were confirmed with biopsy, autopsy, genetic testing, infectious testing, laboratory testing, and imaging. These misdiagnosed patients ultimately had functional, neurodegenerative, psychiatric diagnoses; neoplasms; or cognitive deficits from comorbidities. Many of these patients had a delay to their correct diagnosis for more than a year, and 20% experienced side effects related to unnecessary immunotherapy.
In their editorial review, Dalmau and Graus argued that their 2016 clinical algorithm was designed to assist in the differential diagnosis of AEs and to initiate fast treatment without waiting for antibody test results. They reinforce that almost all 107 misdiagnosed cases could have been prevented if all three requirements had been fulfilled. Furthermore, they emphasize that serum antibody testing often is favored over CSF antibody testing, even at the expense of errors, particularly in a disease type that mostly affects the CNS.
Dalmau and Graus strongly urge to use the algorithm and to carefully apply the third requirement — the consideration of differential diagnoses. They urge physicians to be familiar with the different presentations of AE, to favor clinical reasoning over antibody results, to preferably use CSF testing, and to be aware that a positive test result does not confirm a clinically relevant antibody.
COMMENTARY
This is an important study, since AE misdiagnosis occurred in almost one-third (27%) of patients with AE, often leading to wrong or unnecessary treatment. In recent years, there has been a focus on the fast diagnosis of AE, particularly in young patients, to ensure prompt treatment and favorable outcome. However, the complex combination of clinical presentation, imaging findings, CSF results, and conflicting antibody findings warrants a careful assessment. If these pieces do not create a logical picture, as proposed in the AE algorithm, the diagnosis should be re-considered. Moreover, both author groups aim to raise awareness for the overinterpretation of antibody results and the misleading reliance on serum testing. Antibody testing often is the sought-after diagnostic tool, particularly in rare diseases. However, different specificities and sensitivities for antibody tests complicate their interpretation. Misdiagnosis occurs in many settings, including highly specialized centers, and we should be cautious about anchoring a patient to a diagnosis of AE, which might lead to an escalation of immune treatments and possible harm to the patient.
REFERENCE
- Graus F, Titulaer MJ, Balu R, et al. A clinical approach to diagnosis of autoimmune encephalitis. Lancet Neurol 2016;15:391-404.
