Introduction: Metabolomics provide a promising tool to understand the pathogenesis and to identify novel biomarkers of dementia. This study aimed to determine circulating metabolites associated with incident dementia in a Chinese cohort, and whether a selected metabolite panel could predict dementia.
Methods: Thirty-eight metabolites in baseline serum were profiled by nuclear magnetic resonance in 1440 dementia-free participants followed 5 years in the Shanghai Aging Study.
Results: Higher serum levels of glutamine and O-acetyl-glycoproteins were associated with increased risk of dementia, whereas glutamate, tyrosine, acetate, glycine, and phenylalanine were negatively related to incident dementia. A panel of five metabolites selected by least absolute shrinkage and selection operator within cross-validation regression analysis could predict incident dementia with an area under the receiver-operating characteristic curve of 0.72.
Discussion: We identified seven candidate serum metabolic biomarkers for dementia. These findings and the underlying biological mechanisms need to be further replicated and elucidated in future studies.
Keywords: Alzheimer's disease; dementia; metabolites; metabolomics; nuclear magnetic resonance.
© 2020 the Alzheimer's Association.