Abstract
Specific metabolites have been reported to be potentially associated with Alzheimer's disease (AD) risk. However, the comprehensive understanding of roles of metabolite biomarkers in AD etiology remains elusive. We performed a large AD metabolome-wide association study (MWAS) by developing blood metabolite genetic prediction models. We evaluated associations between genetically predicted levels of metabolites and AD risk in 39,106 clinically diagnosed AD cases, 46,828 proxy AD and related dementia (proxy-ADD) cases, and 401,577 controls. We further conducted analyses to determine microbiome features associated with the detected metabolites and characterize associations between predicted microbiome feature levels and AD risk. We identified fourteen metabolites showing an association with AD risk. Five microbiome features were further identified to be potentially related to associations of five of the metabolites. Our study provides new insights into the etiology of AD that involves blood metabolites and gut microbiome, which warrants further investigation.