Meta-analysis of the association between N-methyl-d-aspartate receptor antibodies and schizophrenia, schizoaffective disorder, bipolar disorder, and major depressive disorder

BACKGROUND N-methyl-d-aspartate receptor (NMDAR) antibodies have been documented in the serum of individuals with primary psychiatric disorders from several independent cohorts, but these findings have not been systematically assessed in aggregate or in relation to methodological covariates. METHODS We searched MEDLINE, EMBASE, and PsycINFO for studies in any language that provided data on NMDAR antibody seropositivity or absolute serum titers in schizophrenia or schizoaffective, bipolar, or major depressive disorders. We used a random effects model to pool estimates across studies. RESULTS Nine studies met the eligibility criteria. Five studies (3387 participants) provided data on NMDAR antibody seropositivity in psychiatric versus control groups based on high-specificity seropositivity thresholds (cell-based assays [CBAs]: 1:320 dilution, 1:200 dilution, visual score>1; enzyme-linked immunosorbent assay [ELISA]: 90(th) percentile of control titers). Meta-analysis showed significantly higher odds of NMDAR antibody seropositivity among those with schizophrenia or schizoaffective, bipolar, or major depressive disorders compared with healthy controls (odds ratio [OR], 3.10; 95% confidence interval [CI], 1.04-9.27; P=.043; I(2)=68%). Four studies (3194 participants) provided outcome data for these groups based on low-specificity seropositivity thresholds (CBAs 1:10 dilution; ELISA: 75(th) percentile of control titers). Meta-analysis showed greater heterogeneity and no significant between-group difference (OR, 2.31; 95% CI, 0.55-9.73; P=.25; I(2)=90%). Seropositive participants in psychiatric groups had various combinations of IgG, IgM, and IgA class antibodies against NR1, NR1/NR2B, and NR2A/NR2B subunits. Subgroup analysis revealed significantly higher odds of seropositivity among all participants based on 1:10 versus 1:320 dilution seropositivity thresholds (OR, 4.56; 95% CI, 2.41-8.62; P<.001; I(2)=0%; studies=2, n=2920), but no apparent difference between first-episode and chronic schizophrenia or schizoaffective disorder (OR, 1.15; 95% CI, 0.19-7.24; P=.88, I(2)=43%, studies=2, n=1108). Average NR2A/NR2B antibody titers determined by ELISA were significantly higher among participants with first-episode schizophrenia (P<.0001) and acute mania (P<.01) compared with healthy controls. Levels decreased by 58% at 8weeks in first-episode schizophrenia, and by about 13% at 4days in acute mania. CONCLUSIONS Individuals with schizophrenia or schizoaffective, bipolar, or major depressive disorders are collectively about three times more likely to have elevated NMDAR antibody titers compared with healthy controls based on high-specificity, but not low-specificity, seropositivity thresholds, though considerable methodological and statistical heterogeneity exists. Evidence concerning the effect of disease state and time of serum acquisition is varied and consistent, respectively. Adequately powered longitudinal studies employing standardized assay methods and seropositivity threshold definitions, and quantifying NMDAR antibodies in both sera and cerebrospinal fluid are needed to further elucidate the clinical and pathophysiological implications of this association.