Anti-NMDAR encephalitis most commonly presents to psychiatric services, so early identification of this disorder is essential. We aim to validate the two screening criteria (Scott et al. and Herken and Pruss) which have been proposed to identify first episode psychosis patients who should have anti-NMDAR antibody testing. The performance of the screening criteria were assessed using anti-NMDAR encephalitis cases published in the literature, and antibody positive and negative cases from a state-wide cohort (Queensland, Australia). Sensitivity, specificity and area under receiver operator characteristic curve analysis was performed. There were 258 anti-NMDAR encephalitis cases and 103 control cases, which demonstrated high performance of both Scott et al. "screening recommended" criteria (sensitivity 97.3%, specificity 85.4%, AUC 0.914) and Herken and Pruss "yellow flags" criteria (sensitivity 91.5%, specificity 83.5%, AUC 0.875). These criteria remained accurate when neurological variables were excluded, and in cases without psychosis. The Scott et al. "screening not recommended", and Herken and Pruss "red flags" criteria did not demonstrate clinical utility for first episode psychosis case screening. The screening criteria with good performance identify an atypical picture of psychiatric presentation with increased risk of anti-NMDAR positivity prior to overt neurological symptoms or investigations and may be beneficial to include in the routine psychiatric assessment process.
Keywords: Anti-NMDAR; Autoimmune; Encephalitis; Psychosis; Screening.
Copyright © 2020 Elsevier Ltd. All rights reserved.