Longitudinal characteristics of plasma biomarkers in Chinese older adults with Alzheimer’s disease

While cross-sectional associations between plasma biomarkers and Alzheimer’s disease (AD) pathology and cognition are well established, their longitudinal relationships remain less elucidated. We aimed to explore the utility of plasma β-amyloid (Aβ) 42/Aβ40 and glial fibrillary acidic protein (GFAP) in detecting and predicting AD progression, particularly considering neuroinflammation as a key contributor. Utilizing data from the Sino Longitudinal Study on Cognitive Decline (SILCODE), 455 older Chinese participants with available plasma Aβ42/Aβ40 and GFAP data were enrolled. Based on cohort-derived thresholds of plasma Aβ42/Aβ40 (A) and GFAP (As), participants were divided into four groups: A + As+, A + As-, A-As+, and A-As-. We employed generalized linear models to examine the cross-sectional and longitudinal associations between plasma Aβ42/Aβ40, GFAP, and their combination with plasma phosphorylated tau (p-tau), residual hippocampal volume (rHCV), and cognitive function. Among 455 participants (mean [SD] age at baseline, 66.11 [6.73] years; 282 [61.98%] female), the A + As + group consistently demonstrated the highest baseline levels and fastest longitudinal increases in plasma p-tau, alongside the lowest baseline levels and fastest longitudinal deterioration of rHCV and cognition. Lower baseline plasma Aβ42/Aβ40 was associated with worse memory and language cognitive function and higher baseline and accelerated increases in plasma p-tau181 and p-tau217. Elevated baseline plasma GFAP was significantly associated with higher baseline plasma p-tau181 and p-tau217, reduced rHCV and cognition, and predicted more rapid increases in plasma p-tau181 and p-tau217, and faster declines in rHCV and cognition. Longitudinally, plasma GFAP slope was positively related to the slopes of plasma p-tau181 and p-tau217 and rapid decreases in global, memory, and executive cognitive function. In contrast, the plasma Aβ42/Aβ40 slope showed a negative correlation only with the p-tau181 slope. These findings demonstrate the superior predictive utility of plasma GFAP in identifying and predicting plasma p-tau increase, neurodegeneration, and cognitive decline than plasma Aβ42/Aβ40 alone. Critically, the combination of both biomarkers yielded the most optimal results, highlighting their synergistic value for early detection and tracking of AD progression.