Prediction of longitudinal synaptic loss in Alzheimer's disease using tau PET and plasma biomarkers

Abstract INTRODUCTION We investigated the associations of longitudinal synaptic loss and cognitive decline with tau burden and plasma biomarkers in Alzheimer's disease (AD). METHODS Twenty cognitively impaired (CI) individuals and 16 healthy controls (HC) underwent cognitive and plasma biomarker assessments, amyloid positron emission tomography (PET), tau PET, and synaptic density PET; after 1 year, tau and synaptic density PET were repeated. The relationships among tau burden, plasma biomarkers, synaptic density, and cognition were investigated. RESULTS The CI group had more longitudinal synapse loss and tau deposition than HCs. Longitudinal synaptic loss was positively associated with longitudinal cognitive decline, negatively with longitudinal tau deposition. Plasma glial fibrillary acidic protein (GFAP) mediates the relationship between longitudinal tau deposition and longitudinal synaptic loss. Tau burden, plasma phosphorylated tau181, and GFAP could predict longitudinal synaptic loss and cognitive decline. CONCLUSIONS The CI group had more longitudinal synapse loss and tau burden increases than HCs. Tau pathology and plasma GFAP could predict longitudinal synapse loss and cognitive decline. Highlights Cognitively impaired individuals had more longitudinal synapse loss in the medial temporal lobe, and increased tau burden in the widespread neocortex than healthy controls. The longitudinal change of synaptic density was negatively associated with the longitudinal change of tau burden, and positively associated with longitudinal cognitive decline. Plasma glial fibrillary acidic protein (GFAP) mediates the relationship between longitudinal tau deposition and longitudinal synaptic loss. Tau burden, plasma phosphorylated tau181, and GFAP could predict longitudinal synaptic loss and cognitive decline.