Value of plasma biomarkers to predict memory change in cognitively unimpaired individuals

Abstract

AbstractBackgroundPlasma levels of amyloid‐β1‐42/1‐40, phosphorylated‐tau (P‐tau) and glial fibrillary acidic protein (GFAP) can be used to predict amyloid‐β pathology in elderly cognitively unimpaired individuals, but their prognostic value in early Alzheimer’s Disease (AD) stages in the normal aging population needs further study. We aimed to investigate the value of these plasma biomarkers for predicting memory change in cognitively unimpaired individuals in a population‐based cohort.MethodFrom the EMIF‐AD PreclinAD study we selected 188 individuals with normal cognition, who had plasma samples, amyloid‐β status and at least one cognitive follow‐up available (total follow‐up years (mean(SD) = 4.3(.42)) (Table 1). Amyloid‐β status was defined using visual read of dynamic [18F]flutemetamol‐PET images or CSF amyloid‐β1‐42/1‐40<0.066 (Euroimmun). Simoa assays were used to measure plasma amyloid‐β1‐42/1‐40 (AMYBLOOD), P‐tau181 (Eli Lilly) and GFAP (Quanterix) levels. Linear mixed models, adjusted for age, sex and years of education, were applied to test associations of baseline plasma biomarker levels with subsequent change in memory (plasma*time), and its interaction with amyloid‐β status (plasma*time*amyloid‐β status).ResultFrom the total cohort tested, 31 subjects (16.5%) were amyloid‐positive. These individuals were older, showed worse baseline memory performance, lower plasma amyloid‐β1‐42/1‐40, and higher plasma P‐tau181 and GFAP levels compared to amyloid‐negative individuals (Table 1). Higher plasma GFAP levels at baseline were significantly associated with worse memory performance over time (GFAP*time: β=‐.03, p=.01). Interestingly, this relation was significantly stronger in amyloid‐positive compared to amyloid‐negative individuals (GFAP*time*amyloid‐β status: β=‐.07, p=.001) (Table 2, Figure 1). No significant associations were observed for baseline plasma amyloid‐β1‐42/1‐40 and P‐tau181 levels with change in memory over time (P‐tau181*time: β=‐.02, p=.07; amyloid‐β1‐42/1‐40 *time: β=.02, p=.12) (Table 2).ConclusionOur data suggest plasma GFAP can be used as a predictor of future memory decline in cognitively unimpaired, amyloid‐positive, individuals in the normal aging population. This supports the use of plasma GFAP as a prognostic marker for disease and treatment‐effect monitoring in the early stages of AD development.