A German network sets out how to read antibody results in psychiatric patients
- The paper is a set of expert recommendations from GENERATE, the German Network for Research on Autoimmune Encephalitis, written by psychiatrists and neurologists for psychiatric care settings. It was developed through iterative discussion within the network; no formal Delphi procedure, systematic review or GRADE assessment was performed, and it does not define new disease entities or diagnostic criteria.
- It states that in German cerebrospinal fluid-based psychiatric cohorts, well-characterized neuronal cell-surface IgG antibodies are mostly below 2%. The often-quoted 77% of patients with anti-NMDA-R encephalitis first seen by psychiatrists comes from the first case series, and the authors write that such figures describe selected, severely ill cohorts, not the share of psychiatric patients who have encephalitis.
- On antibodies: serum neuronal antibody results should not be used to label autoimmune encephalitis or to justify immunotherapy in isolation, and paired serum and CSF testing is advised. Cell-based assays were positive in about 1% of control sera for NMDA-R, CASPR2 and LGI1 antibodies, and low-to-intermediate titer GAD65 antibodies are frequent in type 1 diabetes and in people without neurological disease.
- Other tests are described with their limits: FDG-PET abnormalities in 87–93% of patients with autoimmune encephalitis in meta-analyses, but FDG-PET is not recommended for routine work-up, and MRI abnormalities in fewer than 50% of anti-NMDA-R encephalitis cases. The recommendations rest on expert consensus and are to be revisited as evidence emerges.
A serum antibody result reaches a psychiatric client's chart far more easily than a lumbar puncture does. A German network of psychiatrists and neurologists has now written down how to read that result when the client came to psychiatry first.
The document and how it was made
The paper appeared in Molecular Psychiatry on 25 September 2026 as an open-access review. It proposes a stepwise work-up: clinical assessment, MRI, EEG, blood tests, CSF analysis, and antibody testing in serum and CSF. The authors state that no Delphi procedure, systematic review or GRADE assessment was performed, and that the text introduces no new diagnostic criteria.
Two stages, paired samples, three tiers
Stage 1 antibody testing follows the lumbar puncture, covers IgG only, and uses paired serum and CSF. Stage 2 is reserved for persistent suspicion or supportive findings. Results are sorted into high, intermediate and low confidence. The text names broad panels in patients without strongly suggestive signs as a source of incidental findings that do not improve accuracy. For German CSF cohorts it gives mostly below 2% for well-characterized surface IgG antibodies, and it says positive cell-based assays turn up in about 1% of control sera.
What the client hears
"They found antibodies" sounds like a diagnosis, and for many clients it comes as relief: the illness has a physical cause. In this document's own reading, a low-titer, serum-only result in a psychiatric patient is often incidental.
Before ordering, decide what you will say about a positive result and about a negative one. If a result returns from serum alone, tell the client it is a finding that needs a second sample, not a verdict. Promise no immunotherapy on that basis. Keep the door open for the features the paper treats as reasons to look further – catatonia, subacute onset, rapid cognitive decline, seizures – and send those clients to neurology.
A low-titer serum antibody is a reason to ask a second question, and the client should hear it that way.
The recommendations were developed by iterative discussion within the network, with no formal Delphi procedure, systematic review or GRADE assessment. The authors add that detection rates in psychiatric cohorts vary with cohort selection and antibody definition, that the true prevalence of autoimmune psychiatric syndromes is incompletely understood, and that the recommendations will be revisited as evidence emerges.