CSF Biomarkers Reveal When Alzheimer's Brain Changes Outpace Symptoms

Alzheimer's disease (AD) is defined by two hallmark brain pathologies — amyloid plaques and tau tangles — that accumulate silently for years before symptoms emerge. Despite this, the precise longitudinal trajectories of CSF biomarkers across each clinical stage, and how those trajectories relate to cognitive decline, have remained poorly characterized. This study from the Alzheimer Center Amsterdam addresses that gap directly, providing one of the most granular stage-by-stage pictures of CSF biomarker dynamics to date.

The study enrolled 197 participants from three Amsterdam cohorts (Amsterdam Dementia Cohort, Subjective Cognitive Impairment Cohort, and the PreclinAD study) with at least two CSF samples collected between November 2003 and July 2019. The sample comprised 83 amyloid-negative cognitively normal controls (mean age 63 years), 31 amyloid-positive cognitively unimpaired individuals (mean age 67), 30 with amyloid-positive MCI (mean age 67), and 53 with amyloid-positive dementia (mean age 65). Maximum follow-up reached 19.5 years, with a median of 2 years. CSF was analyzed using the Lumipulse G600II platform for Aβ1-42/Aβ1-40 ratio, total tau (t-tau), and phosphorylated tau-181 (p-tau). Cognition was assessed via MMSE and the delayed recall component of the Rey Auditory Verbal Learning Test (RAVLT).

For amyloid, the Aβ1-42/Aβ1-40 ratio declined significantly in controls (β = −8.55×10⁻⁴ per year, p<0.001) and continued declining in amyloid-positive cognitively unimpaired individuals (β = −1.05×10⁻³ per year, p<0.001), while stabilizing at already-low levels in MCI and dementia groups. Critically, 10 controls (12%) converted to abnormal amyloid over a mean of 4.8 years, demonstrating active pathological progression even in apparently healthy older adults. This stabilization of amyloid in later stages suggests that amyloid accumulation plateaus once full pathological burden is reached.

Tau dynamics told a different story. Both t-tau and p-tau increased across all groups, but the rate of increase was significantly steeper in amyloid-positive cognitively unimpaired individuals (t-tau: β = 17.24 pg/mL/year, p<0.001; p-tau: β = 3.10 pg/mL/year, p<0.001) and amyloid-positive MCI (t-tau: β = 30.80 pg/mL/year, p<0.001; p-tau: β = 4.40 pg/mL/year, p<0.001) compared to controls (t-tau: β = 8.49 pg/mL/year, p=0.002; p-tau: β = 1.36 pg/mL/year, p=0.001). T-tau continued rising in the dementia group (β = 24.97 pg/mL/year, p=0.002). Among controls, 12% reached abnormal p-tau and 14.5% reached abnormal t-tau levels during follow-up, underscoring that tau pathology can emerge even in those initially classified as normal.

Cognitive findings reinforced the biomarker story. Delayed recall declined most sharply in amyloid-positive cognitively unimpaired individuals (β = −0.31 per year, p<0.001) and was significantly associated with CSF amyloid levels (β = 102.29, p=0.03), suggesting amyloid drives early subtle memory loss. MMSE scores declined most in MCI (β = −1.25 per year, p<0.001) and dementia (β = −1.89 per year, p<0.001) groups, where tau burden is highest. Together, these findings support the hypothesis that amyloid initiates the cascade and early memory changes, while tau drives broader cognitive deterioration.

The clinical implication is significant: the accelerated tau rise occurs in amyloid-positive but still cognitively intact individuals, making this the optimal window for secondary prevention trials. Intervening before tau accumulation accelerates — rather than waiting for MCI or dementia — may offer the greatest chance of preserving cognition. These data provide concrete benchmarks for expected biomarker change rates that can inform power calculations and endpoint selection in future AD trials.