The Importance of the CSF Specimen for Antibody Determination in NMDA Receptor Encephalitis
Abstract & Commentary
By Bianca D. Santomasso, MD, PhD
Neuro-Oncology Fellow, Memorial Sloan-Kettering Cancer Center, New York, NY
Dr. Santomasso reports she is a stockholder in Novartis.
Synopsis: The data from this large cohort of patients confirm the importance of submitting cerebrospinal fluid for assessment when NMDA receptor encephalitis is suspected.
Source: Gresa-Arribas N, et al. Antibody titers at diagnosis and during follow-up of anti-NMDA receptor encephalitis: A retrospective study. Lancet Neurol 2014;13:167-177.
Since its discovery in 2007, anti-n-methyl-d-aspartate (NMDA) receptor encephalitis has entered the mainstream of neurology as an important and potentially treatable form of autoimmune encephalitis. The disorder predominantly affects young women and children, and can occur with or without tumor association (usually an ovarian teratoma in roughly half of patients). Neurologic improvement usually occurs with immunotherapy and resection of teratoma if one is present; however, relapses and refractory cases are also seen and incompletely understood.
Patients with anti-NMDA receptor encephalitis present with a subacute onset and stereotyped course characterized by psychosis, memory deficits, seizures, and language disintegration that progresses into a state of unresponsiveness, with catatonic features often associated with abnormal movements and autonomic and breathing instability. Establishing the diagnosis depends on detection of an IgG antibody targeting the GluN1 subunit of the NMDA receptor. The two principle techniques that have been used for antibody testing in this disorder are 1) immunohistochemistry of brain tissue (which produces a highly characteristic pattern of reactivity); and 2) a cell-based assay (CBA) of human embryonic kidney 293 cells expressing the GluN1 subunit of the NMDAR. It is typically thought that these two assays complement each other for laboratory quality assurance of the diagnosis. However, the authors of this study mention that reports by others have suggested that serum testing using only CBA is sufficient for the identification of NMDA receptor antibodies and the diagnosis of NMDA receptor encephalitis. Therefore, they set out to clarify the appropriate antibody testing by directly comparing the sensitivity and specificity of different NMDA receptor antibody techniques in paired serum and cerebrospinal fluid (CSF) samples.
In this study, the authors retrospectively examined a large cohort of patients and report the sensitivity and specificity of paired serum and CSF samples by immunohistochemistry-based and cell-based assays. Two hundred fifty patients with NMDA receptor encephalitis (established by antibody positivity in both immunohistochemistry and cell-based assays) were compared with 100 control patients with encephalopathy. The authors found that all patients with anti-NMDA receptor encephalitis have anti-GluN1 antibodies in the CSF (100% detectable with both techniques), but antibodies in serum are found less often: 91% with immunohistochemistry and 86% with the cell based assay, suggesting that false-negative cases can occur when only serum is used. Put another way, the diagnosis of anti-NMDA receptor encephalitis would be missed in 13% of patients if only serum and a cell-based assay were used. Using a live cell-based technique actually worsened the serum results. None of the 100 paired serum and CSF samples from control samples showed NMDA receptor antibodies with either of the techniques (specificity 100%).
The authors then went on to do quantitative studies with antibodies using antibody titers. From a smaller group of patients for whom serial clinical data (modified Rankin Score) and specimens were available, antibody titers were determined with brain immunohistochemistry and prognostic significance of these antibody titers are reported. Results from their multivariate analysis showed that patients with higher antibody titers in serum or CSF at diagnosis were more likely to have poor outcome or the presence of a teratoma or both. Additionally, patients with a tumor were more likely to have antibodies detectable in serum than those without a tumor. There was also a possible association between good clinical outcome and early decrease of CSF antibody titers during the first month of the disease, but this was not statistically significant. The change in titers in CSF correlated better with clinical relapses than that in serum, but an association between the change in antibody titers and symptoms along the course of disease in this study was not seen. By last follow-up, most patients had a decrease in their serum and CSF titers regardless of clinical outcome, with possible explanations being either a slow spontaneous fading of the immune response or a burned-out stage of the disease. For unclear reasons, after clinical recovery, 24 of 28 CSF samples and 17 of 23 serum samples from patients remained antibody positive. Finally, the authors found that the same epitope specificity (amino acid 369 of GluN1) is present in all tested patients, regardless of their clinical outcome or stage of disease.
The most important point to be taken from this study is that CSF antibody determination is crucial in the initial diagnostic testing for suspected cases of NMDA receptor encephalitis. Examination of serum alone is not sufficient and could lead to a delay in neoplasm diagnosis and immunotherapy initiation in up to 13% of patients. In practice, any patient presenting with suspected autoimmune encephalitis should have both serum and CSF submitted to a commercial laboratory to optimize detection of the more than 20 antibodies (including anti-GluN1) classified as pertinent to autoimmune encephalopathy, since some individual autoantibodies are more readily detected in one specimen type or by a specific assay method.
Assay methods are known to influence sensitivity and specificity. Studies that test serum with cell-based assay alone have identified anti-GNuN1 antibodies in patients with schizophrenia, Creutzfeldt-Jakob disease, Parkinson's disease, and in healthy individuals; these findings could not be reproduced in studies that used both cell-based assays and brain immunohistochemistry in serum and CSF samples. False positives resulting in misdiagnosis of a cognitive disorder as autoimmune might lead to erroneous use of potentially toxic immunotherapy.
The finding that patients with high antibody titers and little or no decrease of CSF antibodies in the first months of the disease are less likely to have a good outcome than those with low titers or a clear decrease in CSF titers is intriguing, but it needs confirmation with prospective studies. Further studies should also investigate whether the level of antibody that persists after clinical recovery predicts relapses and need for chronic immunotherapy. In the meantime, while we wait for prospective studies to confirm CSF antibody titers as a potential prognostic biomarker, a recent large study of 577 patients suggests that clinical assessment of the patient is most important for informing clinical decisions along the course of disease.1
- Titulaer MJ, et al. Treatment and prognostic factors for long-term outcome in patients with anti-NMDA receptor encephalitis: An observational cohort study. Lancet Neurol 2013; 12: