UK Biobank analysis links higher plasma glucose to 0.32-year brain-age gap increase per SD

The study trained models on 4,333 participants' multimodal MRI phenotypes and validated a LASSO regressor with 3.26-year mean absolute error. In the larger cohort, nine metabolites correlated with brain-age gap; glucose showed the strongest signal. Two-sample Mendelian randomization using 392 genome-wide SNPs supported a directional effect of glucose on accelerated aging metrics. This extends prior observational work by quantifying everyday glucose variation rather than diabetes thresholds alone. The design cannot rule out residual confounding from unmeasured lifestyle or vascular factors, and the MR assumes no pleiotropy. Comparable patterns appear in cohorts tracking HbA1c trajectories and incident mild cognitive impairment, where each 1 mmol/L increment raised risk by 1.1-1.3 fold over 10 years. Public-health data already show rising prediabetes prevalence coinciding with earlier-onset dementia in some registries; integrating glucose monitoring into mid-life brain-health trials would test whether modest reductions alter imaging trajectories within five years.