IntegratedAssessment of Internal Metal Mixtures andPlasma Neurodegeneration Biomarkers in Relation to Mild CognitiveImpairment among Aluminum Workers

Abstract Workers in aluminum smelting and processing, besides aluminum, are also chronically exposed to other complex metal mixtures in their working settings, yet integrated evidence linking internal dose, blood neurodegeneration biomarkers, and mild cognitive impairment (MCI) remains limited for occupational surveillance. We measured plasma concentrations of 10 metals (Al, Zn, Li, Cr, Mn, Fe, Co, Cu, Se, Pb) and plasma neurodegeneration-related biomarkers (Aβ42/Aβ40, GFAP, NfL, Total-tau, P-tau181, P-tau396) in 382 aluminum workers in Shanxi, China, and evaluated associations with MCI. Multivariable logistic regression with restricted cubic splines and Bayesian kernel machine regression were applied to characterize exposure–response relationships and mixture effects, while structural equation modeling explored biomarker pathways. Higher Al and lower Zn were consistently associated with greater odds of MCI, and GFAP, NfL, and P-tau396 were positively associated with MCI. Pathway analyses suggested partial mediation of Al- and Zn-related associations through NfL and P-tau396. Integrating metals with biomarkers improved discrimination compared with single indicators (best AUC = 0.924). Model-averaged benchmark dose analysis (BMR = 10% extra risk) yielded an aluminum BMDL of ∼31 μg/L, within the observed exposure range, providing a candidate biomonitoring reference point to inform derivation of biological exposure limits.