Association of AD risk factors with plasma GFAP levels in a memory clinic population

Abstract

AbstractBackgroundPlasma glial fibrillary acidic protein (GFAP) is an emerging AD biomarker. We aimed to study how plasma GFAP correlated with age, clinical stage, sex, and the APOE4 status and whether amyloid pathology may mediate these associations.MethodsCohort of 330 individuals ranging from cognitively unimpaired to dementia due to AD from the Geneva Memory Center. Plasma GFAP was measured using Single molecule array (Simoa) technology. The effect of the variables of interest on plasma GFAP levels was assessed using linear regression models and partial eta squared (eta2 p). In a subset of 169 subject with amyloid PET, mediation analysis was performed to test the mediating effect of amyloid levels on the association between variable of interest and plasma GFAP.ResultsAge had the strongest effect on plasma GFAP levels (eta2 p = 0.39, p< 0.0001), followed by clinical stage (eta2 p = 0.10, p< 0.0001), sex (eta2 p = 0.06, p< 0.0001), and APOE4 (eta2 p = 0.02, p = 0.007). The effect of APOE4 status on plasma GFAP was fully mediated via the cortical amyloid uptake. The bootstrapped indirect effect was 0.09 (p< 0.0001) and the proportion of the effect of APOE4 status on GFAP that goes through the cortical amyloid uptake was 69%. No mediating effect by amyloid was found for age and sex.ConclusionThese results show that in a memory clinic population, i) AD risk factors and severity affected plasma GFAP levels and, ii) the effect of APOE4 on GFAP is amyloid‐dependent whereas that of age and sex is amyloid‐independent. GFAP cut‐offs for the early diagnosis of AD should take age and sex into account.