Differential, not uniform, elevation of plasma neurofilament light and glial fibrillary acidic protein across acute psychiatric presentations: implications for biomarker interpretation in psychiatry

Objective To investigate plasma neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP) levels across a range of acute psychiatric disorder presentations, and to determine whether elevations are specific to certain diagnoses. Methods We analysed biobanked lithium heparin plasma samples from 121 patients with acute psychiatric presentations (including mania, psychosis, alcohol use disorder, anorexia nervosa, depression, and adjustment disorder), and 59 healthy controls. NfL and GFAP were measured using Simoa assays. Group differences were compared via bootstrapped general linear models adjusted for age, sex, and weight. Results Among 180 participants, adjusted NfL was significantly elevated compared with controls in alcohol use disorder ({beta}=1.32, p=0.002), anorexia nervosa ({beta}=0.86, p<0.001), mania ({beta}=0.44, p=0.015), and psychosis ({beta}=0.34, p=0.034), but not in depression or adjustment disorder. Unadjusted NfL was highest in alcohol use disorder (mean 43 pg/mL) and anorexia nervosa (20 pg/mL) versus controls (9 pg/mL). Adjusted GFAP was elevated only in alcohol use disorder ({beta}=1.01, p=0.024). Conclusions Plasma NfL and GFAP are not uniformly elevated across psychiatric disorders, but show a disorder-specific pattern, with the largest elevations in alcohol use disorder and anorexia nervosa and more modest elevations in mania and psychosis. These findings support nuanced, biologically grounded, diagnosis- and stage-specific interpretation of these biomarkers in psychiatry, and caution against interpreting a single elevated value as evidence of neurodegeneration. Because some acute psychiatric presentations show elevations approaching the neurodegenerative range, control-derived reference ranges may be insufficient when distinguishing neurodegeneration from primary psychiatric disorders, supporting the need for diagnosis- and state-specific reference data.