Despite comprehensive serum and CSF antibody panels, a characteristic EEG pattern, and an excellent predictive clinical score (APE²), diagnosing antibody-mediated encephalitis remains difficult. This is due to the insidious disease course, resemblance to more common psychiatric and epileptic disorders, and a significant latency in lab results.

14 patients met criteria for antibody-mediated encephalitis, but five presented diagnostic conundra. This included three patients with exclusively psychiatric symptoms that were attributed to a long-standing psychiatric history. Two of these eventually had NORSE and non-neuronal antibodies, while the third was only diagnosed with anti-NMDA syndrome by antibody panel results. The fourth patient presented with NORSE after a viral prodrome. There was a slight elevation in serum anti-GAD65 in the second of three samples. The fifth patient exhibited epilepsy, Morvan syndrome, transient MRI signal changes and neutrophilic pleocytosis following a viral prodrome, but all other studies were negative. MRI was normal in the other four cases. EEG was abnormal in all five cases, including one with an extreme delta-brush.

Developments in immunofluorescence, pathology and symptomatology have provided the means to tackle the complexities of autoimmune encephalitides. Nonetheless, our case series indicates the need to refine diagnostic guidelines. For instance, chronic psychiatric illness may obscure the progressive neuropsychiatric changes crucial for diagnosis. And current lab tests may not reveal the directly pathological antibody. We suggest adding any EEG abnormalities to formal clinical scores. Even if nonspecific, these can provide relatively quick and inexpensive diagnostic clues.