Pitfalls in clinical diagnosis of anti-NMDA receptor encephalitis

To report pitfalls in the clinical diagnosis of anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis.

We retrospectively reviewed the clinical information of 221 patients with clinically suspected autoimmune neurological disorders who underwent testing for autoantibodies against neuronal cell-surface antigens between January 1, 2007 and September 10, 2017. Forty-one patients met the diagnostic criteria for probable anti-NMDAR encephalitis (probable criteria), but one was excluded because neither serum nor CSF was examined at the active stage. Thus, in 220 patients, sensitivity and specificity of the probable criteria were assessed.

NMDAR-antibodies were detected in 34 of 40 patients (85%) with the probable criteria; however, 2 of the 6 antibody-negative patients had ovarian teratoma. The median age at onset was higher in antibody-negative patients than those with antibodies (49 vs. 27 years, p = 0.015). The age at onset was associated with the probability of antibody detection (p = 0.014); the probability was less than 50% in patients aged 50 years or older. NMDAR-antibodies were also detected in 5 of 180 patients who did not fulfill the probable criteria; these patients presented with isolated epileptic syndrome (n = 2), atypical demyelinating syndrome (n = 2; one with aquaporin 4 antibodies), and autoimmune post-herpes simplex encephalitis (post-HSE) (n = 1). Sensitivity and specificity of the probable criteria was 87.2 and 96.7%, respectively.

The probable criteria are valid, but the diversity of clinical phenotype should be taken into account in diagnosing anti-NMDAR encephalitis particularly in patients aged 50 years or older, or with isolated epileptic syndrome, atypical demyelinating syndrome, or post-HSE.