Abstract
BACKGROUND: Alzheimer's disease (AD) is a multifactorial neurodegenerative disease. Identifying early diagnostic biomarkers are crucial for improving the accuracy and timeliness of the disease diagnosis and prognostication. Emerging evidence links metabolic dysregulation to AD pathogenesis. We evaluated serum metabolites and their association with regional brain hypometabolism measured using fluorodeoxyglucose positron emission tomography (FDG‑PET). METHODS: We studied 892 participants using the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort, including 286 cognitively normal (CN) individuals, 468 patients with mild cognitive impairment (MCI), and 138 with AD who had a complete baseline FDG‑PET and serum metabolomics data. After outlier removal and adjustment for age, sex, and years of education, we tested linear associations between individual metabolite levels and glucose standardized uptake value ratio (SUVR) as a measure of brain metabolism within predefined regions of interest (ROIs). RESULTS: In the AD group, higher levels of hydroxyproline (β = 0.24, P = 0.005) and aspartate (β = 0.17, P = 0.047) were associated with greater FDG SUVR prior to adjustment, although the association with aspartate did not remain significant after FDR correction. In the MCI group, higher levels of putrescine (β = -0.12, P = 0.010) and glutamine (β = -0.11, P = 0.019) were related to lower FDG SUVR prior to adjustment but lost significance after correction. In CN participants, higher levels of ornithine were associated with greater FDG SUVR (β = 0.13, P = 0.030) prior to adjustment. CONCLUSION: Our findings suggest that specific serum metabolites show differential associations with regional brain glucose hypometabolism in patients across the AD continuum. While these associations did not remain significant after correction for multiple testing, preliminary signals linking amino acid-related metabolites (e.g., hydroxyproline, aspartate, putrescine, glutamine, and ornithine) to brain metabolism highlight the potential role of metabolic pathways in AD pathophysiology. These results support the utility of metabolomics as a complementary approach for identifying early biomarkers of neurodegeneration.