BackgroundEarly detection and staging of Alzheimer's disease (AD) remain challenging in low-resource settings where cerebrospinal fluid analysis and neuroimaging are not routinely accessible. Blood-based biomarkers such as phosphorylated tau at threonine 181 (p-tau181), glial fibrillary acidic protein (GFAP), and amyloid-tau ratios have improved diagnostic performance; however, their integration with systemic metabolic markers remains insufficiently investigated.ObjectiveTo characterize metabolic-neuroglial plasma signatures across the cognitive continuum and identify biomarkers capable of discriminating mild cognitive impairment (MCI) from mild AD and cognitively unimpaired (CU) individuals.MethodsSeventy adults classified as CU, MCI, or mild AD underwent plasma quantification of Aβ40, Aβ42, p-tau181, and GFAP alongside metabolic markers, including glycated hemoglobin, calcium, vitamin D, homocysteine, thyroid-stimulating hormone, C-reactive protein, and platelet count. Group comparisons were conducted using non-parametric tests. Multivariate approaches, principal component analysis, partial least squares discriminant analysis, and Random Forest (RF), were applied to detect discriminative biomarker patterns, incorporating automatically generated ratios.ResultsMCI participants showed higher plasma homocysteine and Aβ40 than CU. Mild AD was characterized by elevated p-tau181 and GFAP and a reduced amyloid-tau balance captured by Aβ40-based ratios. CU was distinguishable from both clinical groups using univariate markers and ratios, whereas separation between MCI and mild AD required multivariate integration. A six-biomarker RF signature robustly discriminated MCI from mild AD (AUC = 0.854; permutation p = 0.0044).ConclusionsIntegrated plasma panels combining neurodegenerative and metabolic markers improve staging across the AD continuum and may support clinical decision-making where advanced diagnostics are limited.
Santos et al. (Thu,) studied this question.