Abstract
EBioMedicine. 2026 Jul 11;130:106370. doi: 10.1016/j.ebiom.2026.106370. Online ahead of print.
ABSTRACT
BACKGROUND: Nearly all individuals with Down syndrome (DS) develop Alzheimer's disease (AD) dementia, primarily due to overexpression of the APP gene. Although specific cerebrospinal fluid (CSF) and plasma tau biomarkers have been investigated in DS-AD, how different tau species change in the DS-AD continuum in comparison to sporadic AD remains uncertain.
METHODS: In this cross-sectional study, we analysed CSF and plasma tau biomarkers in 461 samples from the DABNI and SPIN cohorts, including individuals with DS, cognitively normal euploid participants, and patients with sporadic AD. Biomarker differences were assessed using linear regression with Tukey post hoc comparisons. LOESS modelling was applied to estimate the age at which tau biomarkers became abnormal.
FINDINGS: We analysed 461 participants from the DABNI and SPIN cohorts. Both CSF and plasma tau biomarkers increased during the asymptomatic stages of DS and in euploid controls, coinciding with Aβ positivity; across the DS clinical spectrum the largest increases were observed for CSF NTA-tau (fold-change [fc] = 6.46-6.94), CSF p-tau217 (fc = 6.43-6.74) and plasma p-tau217 (fc = 4.63-6.54) (linear regression adjusted for age, sex and APOE-ε4 with Tukey post-hoc tests; all p < 0.001). During the dementia stages, CSF tau biomarkers showed only modest further increases (no CSF biomarker differed between pDS and dDS; all p ≥ 0.268), whereas plasma tau biomarkers retained a broader dynamic range across symptomatic phases (pDS vs dDS: plasma p-tau217 p = 0.001, p-tau181 p = 0.002, p-tau231 p = 0.004). Plasma p-tau217 showed the highest diagnostic accuracy, with areas under the curve (AUC) of 0.91-0.97 for biological categorisations and numerically higher values than CSF in symptomatic stages (pDS vs dDS: plasma p-tau217 AUC = 0.69 [95% CI 0.58-0.80] vs CSF p-tau217 AUC = 0.53 [95% CI 0.41-0.65]; DeLong test p = 0.019). In LOESS analyses, tau biomarkers diverged from age-matched controls in the late 30s to early 40s (e.g., plasma p-tau217 ≈ 37.3 years, CSF p-tau181 ≈ 38.1 years) and reached abnormality (+2 SD) over an approximately 20-year span between the fourth and sixth decades, outlining differential but temporally compressed increases. Finally, during symptomatic stages, tau biomarker levels remained stable in DS-AD, in contrast to sporadic AD, where levels declined with advancing age.
INTERPRETATION: These findings highlight the complementary roles of CSF and plasma tau biomarkers in tracking disease progression: CSF biomarkers capture early pathological changes, whereas plasma biomarkers more effectively reflect disease progression within symptomatic stages. Furthermore, tau biomarkers might support disease staging and monitor clinical progression in DS-AD, but with the need to adapt biomarker frameworks to this specific population.
FUNDING: La Caixa Foundation, Instituto de Salud Carlos III, Generalitat de Catalunya, National Institute on Ageing, Wellcome Trust, Jérôme Lejeune Foundation, Medical Research Council, Alzheimer's Association, National Institute for Health Research, EU Joint Programme-Neurodegenerative Disease Research, Alzheimer's Society.
PMID:42435584 | DOI:10.1016/j.ebiom.2026.106370
UK DRI Authors
