LAUSR

OBJECTIVES To address practical issues in measuring autoantibodies to neuronal cell-surface antigens (NSAs) in various autoimmune neurological disorders (ANDs). METHODS We retrospectively reviewed the clinical information of 221 patients with clinically suspected ANDs who underwent antibody testing for NSAs between January 2007 and September 2017. 31 were excluded. In 190 patients, antibody-detection rate (ADR) and antibody-phenotype association were assessed. RESULTS Fifty-four patients had NSA-antibodies: NMDA receptor (NMDAR) (n = 39), AMPA receptor (n = 3), leucine-rich glioma inactivated 1 (LGI1) (n = 3), glycine receptor (GlyR) (n = 3), GABA(A) receptor (n = 2), GABA(B) receptor (n = 1), metabotrophic glutamate receptor 5 (n = 1), or unknown (n = 6); 3 had multiple NSA-antibodies. ADR in patients with diagnostic criteria for "possible autoimmune encephalitis (AE)", "probable anti-NMDAR encephalitis", "definite autoimmune limbic encephalitis (ALE)", and "stiff-person spectrum disorder (SPSD)", was 34% (46/134), 85% (34/40), 46% (11/24), and 22% (4/18), respectively, but NSA-antibodies were not identified in 11 patients with systemic autoimmune disorders (SADs). Among 134 patients with "possible AE" criteria, NMDAR-antibodies were more frequently identified in patients with typical anti-NMDAR encephalitis than those without (34/40 [85%] vs. 4/94 [4%], p < 0.0001). LGI1-antibodies were identified in patients with ALE but not in the others (3/24 [13%] vs. 0/110 [0%], p = 0.005). GlyR-antibodies were identified in those with stiff-person syndrome plus (2/8, 25%) or stiff-limb syndrome (1/6, 17%). CONCLUSIONS NSA-antibodies were most frequently identified in "probable anti-NMDAR encephalitis", followed by "definite ALE", "possible AE", and "SPSD", but not identified in SADs. NMDAR, LGI1 and GlyR were associated with clinical phenotype. Cell-surface antigens should be determined based on individual phenotype.