Cerebral metabolic covariance in delirium: pattern response to symptomatic changes.

BACKGROUND: Delirium is an acute neuropsychiatric condition linked to increased dementia risk, yet its mechanisms remain unclear. Previous studies reported cerebral hypometabolism. Spatial covariance of 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) was applied to acutely unwell inpatients with and without delirium (Delirium, Conunwell) to derive a delirium-specific metabolic pattern (DP). DP expression was evaluated in healthy older adults (Conhealthy) and patients with delirium superimposed on dementia (DSD) and tracked longitudinally with symptom resolution. METHODS: Seventy participants were included (30 Conhealthy, 10 Conunwell, 13 Delirium, 17 DSD). Voxel principal components (PCs) identified intercorrelated metabolic patterns. RESULTS: The DP distinguished Conunwell from Delirium, with relative hypometabolism in default, frontoparietal, visual, and frontostriatal networks, and relative hypermetabolism in sensorimotor and limbic hubs. Delirium and DSD demonstrated higher DP expression than controls. In participants with follow-up, recovery was paralleled by reduced DP expression. DISCUSSION: Delirium exhibited a metabolic profile of network dysfunction that may be modifiable and clinically responsive. HIGHLIGHTS: Derivation of a DP, primarily attributable to delirium itself, rather than acute illness or dementia. The DP revealed metabolic dysfunction spanning large-scale networks, consistent with the proposed model of delirium as global brain failure. DP expression was elevated in delirium, both with and without dementia, compared to healthy and acutely unwell controls. In participants with follow-up, clinical recovery was paralleled by a reduction in DP expression. The delirium pattern reflected clinically responsive and modifiable network dysfunction.