Background: Delirium is common among hospitalised older adults yet frequently missed or misattributed to dementia and primary psychiatric disorders (PPD) due to overlapping features. Plasma neurofilament light chain (NfL) can be elevated in perioperative and intensive care patients with delirium, but its utility in general hospital inpatients with delirium remains unclear. We examined whether plasma NfL and glial fibrillary acidic protein (GFAP) can distinguish delirium from dementia and PPD in a population of general hospital inpatients.
Methods: In Phase 1 (Delirium Discovery Cohort), we measured NfL and GFAP in hospitalised patients with delirium and compared them with dementia and PPD groups to derive an optimal NfL threshold. Phase 2 tested this threshold's performance by pairing the Phase 1 delirium group with internal independent dementia and PPD groups.
Results: In Phase 1 (n = 168), inpatients with delirium showed higher levels of both NfL and GFAP than dementia or PPD groups. Receiver operating characteristic (ROC) curve analysis yielded an optimal NfL threshold of 24.1 pg/mL (AUC = 0.88, 96% sensitivity, 69% specificity). GFAP showed marginally lower accuracy (AUC = 0.86), and combining NfL with GFAP (AUC = 0.90) did not significantly improve performance. In Phase 2 (n = 70), the NfL threshold of > 24 pg/ml maintained strong performance (AUC = 0.95, 96% sensitivity, and 77% specificity).
Conclusions: Plasma NfL shows promise as a biomarker for delirium in hospitalised patients, distinguishing it from dementia and PPD presentations with high sensitivity. Larger multicentre studies are needed to refine cut-offs, evaluate temporal dynamics, and determine integration into acute hospital delirium management pathways. The availability of a blood marker to aid in the identification of delirium would have significant health, economic and social benefits.